Inside the Marijuana Breathalyzer Arms Race

In November, Massachusetts and California legalized marijuana consumption as Florida, Arkansas, and North Dakota became the 26th, 27th, and 28th states to legalize the consumption of medically prescribed marijuana. When the new legislation goes into effect, half of Americans will live in states where marijuana can be legally consumed and roughly 67 million Americans — some 21 percent of the population — will be able to smoke without any non-federal restraints beyond their own sense of responsibility. And if they want to drive stoned, they will be able to do so.

That poses a problem for law enforcement officials, who are poorly equipped — literally and figuratively — to penalize chemically altered drivers when the chemical in question is tetrahydrocannabinol. THC doesn’t intoxicate a person the same way alcohol does and doesn’t cycle through the liver and get flushed from a person’s system. The chemical hangs out in the human bloodstream for up to 30 days. This is problematic for stoners because blood tests regularly return positive for drivers who aren’t high but have smoked weed in the days or weeks leading up to an encounter with police. And whereas there is a nationally — and, perhaps more importantly, scientifically — agreed upon standard for impairment (0.8 percent blood alcohol content level), there is no one line set for what constitutes a problematic high. Variations are common between and even within states, a fact that is evidence of both problematic policies and core issue with enforcement: The lack of a means to measure intoxication.

Consumer research suggests that the global market for breathalyzers is currently worth slightly more than $500 million annually. It’s about to double — potentially triple — in size thanks to the rapid expansion of access to marijuana. Demand may be driven by law enforcement or employers, but what’s absolutely clear is that it will be massive and there is a great deal of money to be made.

What is also clear is that two companies stand to gain the most from a potential weed breathalyzer boom and that they are racing each other on a well-defined course. Hound Labs outside Berkeley, California, and Gainesville, Florida’s Cannabix Technologies both use local university chemistry departments to inform their research and collaborate with local law enforcement and volunteers for testing purposes. Both have medically trained drug scientists on staff. Both are operated by entrepreneurs embracing a second act. But the race isn’t a simple first-to-market affair.

Despite their similarities, Hound Labs and Cannabix are fundamentally different companies built on diverging visions of what marijuana legalization means for Americans. Hound Labs is attempting to create a combination alcohol and marijuana breathalyzer using technology they refuse to divulge; Cannabix hopes to focus solely on a tool that measures THC levels with the science of mass spectrometry and ion mobility. Both believe they have the more viable product and both believe their device is more accurate. Both can’t be right.

Mike Lynn got the idea for a breathalyzer one day as he rode his bike around Mt. Diablo just south of Oakland. An avid road cyclist, Lynn nearly got struck by hotboxing motorists who were attempting to pass him on narrow, unguarded mountain roads. This was shortly after Colorado’s 2012 marijuana legalization, and carloads of stoners were pulling on joints brought back to California from Boulder and Denver. Precarious mountain roads were becoming more precarious.

“He did not feel safe on the roads,” said Jenny Lynn, Mike’s wife and Hound Labs co-founder. “He stopped biking because he was concerned for his safety.”

Mike was in a unique position to investigate the idea of marijuana breathalyzers: He worked as an ER physician, was a deputy reserve sheriff, and had experienced the effects of impaired driving from his biking days. Mike reasoned — in 2012 — that there was no adequate breathalyzer that would be able to detect THC levels in the breath. He could develop one, but he didn’t think it was going to be marketable without combining a marijuana breathalyzer with an alcohol breathalyzer: Efficiency and simplicity of use were of course key, but also, the two were often taken together and worked together to impair a person.

By December 2015, after a few years of research with chemistry labs at the University of California, Berkeley, Hound Labs had produced a prototype for a patent-pending device that would both measure alcohol and marijuana intoxication. Jenny Lynn says the product is “unique, nothing like the science that Cannabix uses,” because her company’s ultimate focus was “accuracy of the lab, optimized, and miniaturized.”

“It’s like a docking station,” Lynn continues. “The dimensions are of a reasonable size, probably like a small laptop, an inch-and-a-half in height, not as thin but super portable. You take the handheld and you plug it in, you connect it and it initiates this automated process that does the breath sample analysis. The base station is programmed to perform a series of automated routines that chemically isolate the THC molecules from the breath, generation a composition for measurement. Once that’s created, the base station triggers a measurement mechanism, optically measuring THC levels.”

Lynn insists the science behind the breathalyzer has been checked with mass spectrometry throughout each stage of engineering — along with local law enforcement, who’ve controversially been handed off preliminary copies of the breathalyzer for real-life beta testing. After all, there’s a huge liability on the line: “We want this to work,” she said. “We want to make sure the wrong people aren’t wrongfully arrested.”

It’s a sentiment Lynn — and Cannabix’s Kal Malhi — repeats often, because it represents a key obstacle: There is no scientific way of measuring how impaired someone is by marijuana, since marijuana reactions vary greatly.

Experts — both in the breathalyzer development industry and the burgeoning legal cannabis field — consistently point to legal limits of alcohol in the blood as the guiding post for how the United States will tackle marijuana legalization. The inherent problem with this type of logic is that alcohol and marijuana are not at all alike, save for the fact that they can both impair you.

Measuring THC in the breath is significantly different — and more difficult — from measuring alcohol in the breath. Marijuana molecules appear at a rate of one per one trillion in the breath. Comparatively, alcohol molecules appear in the breath at a rate of one per one thousand.

It’s not just the concentration of molecules floating around in our breath that is different; it’s the way they affect us over time. To understand this, a person has to understand Henry’s Law, which was the foundational basis by which alcohol breathalyzers were developed after Prohibition. The concept dates to 1803, from chemist William Henry, who theorized that when a volatile substance — alcohol, in this case — was dissolved in a liquid — here, your blood — the concentration in the vapor directly floating above the liquid — the breath you exhale — will be proportional to the ratio of the chemical within the liquid. This law maintains the factual accuracy of the breathalyzer’s inherently accepted truth, that when an intoxicated human blows into a breathalyzer, their breath is a reflection of the proportion of alcohol swimming through their blood.

Henry’s Law, however, doesn’t translate to marijuana consumption and how it impairs a person. The most glaring issue is the fact that THC doesn’t act like alcohol molecules — they are less frequent and have the capability to remain lodged in a person’s system for a longer amount of time than alcohol. Livers do the dirty work of cycling blood, scrubbing the alcohol from your blood before re-releasing it into your system, punishing a heavy drinker with a hangover. Not so with marijuana. Those molecules can just float around a body for days on end, for up to 30 days.

Not only that, but smoke has a gaseous element to it, unlike alcohol’s liquid actions. That means marijuana taken by smoking floats through the lungs first, lingering for about a couple hours before dissipating within the bloodstream. Using the breath here is advantageous: If you’ve smoked in the past couple hours and let a huff of air out, chances are strong that the THC you’ve inhaled will blast through your lungs and out with your exhalation, signalling to anyone with the right tools that you’ve gotten high in the past 120 minutes. But — there’s a huge but here — THC would also be latently in your system for up to a month after an initial smoking incident.

For those set on developing a breathalyzer, this presents a huge problem: How can you guarantee that a person is truly intoxicated and hasn’t just lit up a joint in the past month? In 1993, the National Highway Traffic Safety Administration tried to figure out the correlation between marijuana and driving impairment. The 143-page report — one of the most exhaustive federally funded studies of marijuana to date — relied on small sample sizes of mostly men and used blood as its baseline way of testing for impairment, finding a frustrating lack of correlation between marijuana use and impairment:

“This still leaves the question open why performance appears to.be more affected by THC in laboratory than actual driving tests,” the NHTSA reports reads. “Many researchers defend the primacy of laboratory performance tests for measuring drug effects on skills. related to driving on the basis of superior experimental control. Certainly some control is always necessary to reduce the confounding influence of extraneous factors that would otherwise so increase measurement error as to totally obscure the drug’s effects. However, only some extraneous factors are truly sources of measurement error and others either attenuate or amplify drug effects in real driving and must be considered as relevant to a test’s predictive validity.”

Furthermore, there’s evidence from state traffic databases, including Colorado’s, suggesting that there are strong correlations between accidents and 11-nor-9-carboxy-THC, a chemical released when the body metabolizes marijuana, rather than with THC itself. That would suggest that people exposed to marijuana are more likely to crash than those smoking it — complicating how to detect marijuana and whether a positive identification of marijuana within a person’s system actually impairs judgment.

Brian Vicente, a lawyer and executive director of marijuana advocacy group Sensible Colorado, has worked extensively on the legal parameters of marijuana impairment. He says marijuana breathalyzers are “a timely topic.” The laws by which states differ in prosecuting for marijuana impairment currently vary greatly, with most using nanograms measured in blood (Hound Labs’ breathalyzers use picograms, which measures to the trillionth of a gram; nanograms measure to the billionth of a gram). States like Utah have zero tolerance laws, where the slightest trace of marijuana in the blood is grounds for arrest. Others waver: Colorado allows 5 nanograms per liter of blood, most are around 3 nanograms per liter, others fall in various fractions in between. (The two states share a very long border.)

“There’s no real scientific level of what defines impairment,” Vicente says. Blood is the most widely used method, after a field sobriety test; if a person is suspected to be impaired, they go to jail, get blood tested, and a determination of impairment is made either way. But don’t dismiss blood. “It’s actually pretty accurate,” Vicente adds. “It tests active THC versus inactive THC. Marijuana can stay in your psychoactive system for up to a month, but blood can differentiate between active and inactive, which means we [lawyers] usually have a good case.”

Which makes measuring alcohol and marijuana in the system especially attractive to law enforcement. “It would be a major law enforcement savings if that were the case,” Vicente said. Blood isn’t perfect, and it’s chock full of potential disturbances, particularly if the person has mixed alcohol and marijuana. It’s invasive, has to be done at a police station, and uses law enforcement resources, which is to say: It’s messy and not cheap, and breathing into a machine sure as hell is a lot more efficient.

That efficiency and delicate tightrope walk of accuracy and wrongful accusation is what drives Cannabix CEO Kal Malhi, a former real estate agent who lost his job in the Great Recession of 2008 and joined the Royal Canadian Mounted Police, where he saw firsthand how difficult it was to ascertain if a person was high — medical marijuana has been legal in Canada since 2001.

Malhi knew as a police officer that drug tests most often rely on body fluids — urine, blood, spit. Those weren’t going to work with marijuana, which could sit idle in the blood for days on end. “In Canada, police officers aren’t allowed to take blood,” Malhi points out, comparing that system to the way American jails tend to test the level of THC in the system — which, as Malhi points out, can be a messy, time-consuming process, given that the level of impairment from marijuana can last 2.5 to 3 hours.

He got wind of some work done at Sweden’s Karolinska Institutet, where researchers were beginning to consider testing marijuana levels via breathing, the same way roadside alcohol tests are done. Malhi borrowed the concept, launching Cannabix, combining the Swedish breath technology with the science of mass spectrometry, spearheaded by University of Florida chemistry professor Richard Yost. The patented technique detects and measures ions by using a Filed Asymmetric Ion Mobility Spectrometry device, or FAIMS. In 2015, Yost told The Gainesville Sun that developing FAIMS initially had nothing to do with pot — he was simply trying to figure out a way to measure speedy ions moving through space accurately, working with law enforcement to develop a way to detect bombs.

Malhi saw an opportunity, knowing that ions move uniquely within an individual who has smoked marijuana, poising Yost’s research in a key position to change the course of marijuana impairment testing. “Mass spectrometry isolates ions — when you breathe out, there are various metabolites that come out, five different compounds in your breath,” Malhi explained. “[The breathalyzer] can separate and just look for THC.”

And that THC, too, evolves as time trudges on. In the first couple hours of being intoxicated, a marijuana smoker’s levels of THC are mirrored in the ions they breathe out, which take on a particular metabolite shape that signal to the observer the fact that the person has imbibed marijuana in the past couple hours.

To be clear, the breathalyzer — which would flash a simple “yes” or “no” based on the user’s most recent use — does not indicate intoxication. “It confirms only that you have used it,” Malhi clarified. “The test would be only one part of the police officer’s investigations, based on evidence that they [the marijuana user] has been driving in a way that is not consistent with others. The interaction should indicate impairment with visual or verbal interaction; if those two stages are positive, the breathalyzer would come in as a third step. Our test could confirm you used it [marijuana].”

In other words, Cannabix’s breathalyzer differs from Hound Labs’ in that it doesn’t necessarily provide a surefire way for us to know if a person is actually intoxicated, just that the person has used. Is that enough to warrant the label of being under the influence? No, which makes Cannabix’s breathalyzer potentially problematic. Sure, you have THC floating around your breath. Does that make you impaired? Not necessarily.

Malhi says this is the first step towards breathalyzer technology that will make it possible for us to know for sure when a person is impaired by pot. Cannabix has differed in its testing from Hound Labs by working towards a clinical, medically approved device that is approved by the FDA, criticizing Hound Labs for “just handing out products to sheriffs to test.”

“We are testing with medicinal marijuana patients to get results that are only positive during the two-and-a-half hour window,” Malhi said. “We need to correlate each breath test with a blood test. Hound [Labs’] is a 2-in-1 — that’s not realistic, that’s not what the courts are going to approve. A police officer can’t do clinical trials out on the road.”

It’s important to take a step back and realize what Malhi is implying here — that Hound Labs is rushing to market with a product that may land innocent people who are not impaired in jail simply for having smoked weed at some juncture. He’s suggesting that doing real-life beta testing with law enforcement is an ineffectual, downright dangerous method of testing. The upside of that approach is that Jenny Lynn can realistically estimate that her company will have a product by the end of 2017 and then go through the FDA approval process. The downside is, as Malhi points out, that it puts an emphasis on getting to market that might incentivize compromise on technical issues.

“A police officer can’t do clinical trials,” Malhi says. “We think our technology is the ideal technology given where we are in the scientific process. We are doing scientific trials. An earlier market for us isn’t as important. We’re not in a rush to get this product out; we want an accurate device.” (Lynn responded, saying that Hound Labs is conducting clinical trials at the University of California at San Francisco. “We did prototype testing with law enforcement to get feedback on our device for the next version of our prototype,” she said.)

While Hound Labs is looking to get a usable product to law enforcement by the end of 2017, Cannabix is aiming for mid-2018, at which point more states may have legalized marijuana.

Cannabix and Hound Labs might differ in their scientific approaches and business ethos for creating a marijuana breathalyzer, but both sound eerily similar when it comes to why they are creating a breathalyzer. “Growing and selling marijuana is associated with being a gateway drug,” Malhi said. “Our device is helping to protect society.” (Note: Marijuana’s role as a “gateway drug” has been controversial, with some suggesting it’s more tied to increased alcohol use than escalation into harder drugs.)

Hound Labs has created a separate foundation designed to raise money to fund research on how marijuana affects the mind and body. For being a substance that has been used for thousands of years with legendary effects, the way THC works in changing the biochemistry of our brains and how we operate is poorly understood, primarily because of its Schedule 1 classification. For a country that is rapidly accepting marijuana as a drug that isn’t necessarily as harmful as harder drugs and is open to controlled usage of it within parameters that are increasingly reflecting that of alcohol, we know terrifyingly little about what it means to be impaired by marijuana.

That might change in 2017, but it might be too late — which makes Lynn’s and Malhi’s missions all the more urgent. “We know our breathalyzer won’t be on the market soon enough,” Lynn said. “We have to add stoned driving to the national conversation.”