In a sparely furnished, exceptionally clutter-free office in Oakland, California, Mike Lynn is blowing into a black plastic box the size of a rat trap. It is making a loud, steady beep. An electronic blue bar on the side progresses. “It takes a certain amount of breath,” he says after around eight seconds, when the bar has advanced to about a quarter. “I can do it in one breath, some people take two.”
Lynn, whose career has taken him from emergency medicine doctor to venture capitalist, is now the co-founder and CEO of Hound Labs. The startup began in 2014 and has received $14m in funding to date including $8.1m from Benchmark Capital which also provided early stage funding for Uber, Twitter and eBay. Hound Labs aims to do what no one has yet done: produce a pot breathalyzer for use at the roadside. While alcohol breathalyzers are standard, so far there is no such equivalent for marijuana. And that’s because it is a tough problem: unlike alcohol, breath levels of THC – the psychoactive chemical in cannabis that makes people high – are much lower. Up until now it has taken large, costly instruments to detect. “The analogy is taking 20 Olympic swimming pools and finding one drop of water,” says Lynn. “But that is what we are so excited about – because we figured it out.”
The time is ripe: stoned drivers are an increasing problem as more US states, and even whole countries, move towards legalizing cannabis for medicinal or recreational use. California went from medicinal only to recreational use on 1 January. Canada is planning to go legal this summer. According to one study based on data from the state of Washington, the number of cases of impaired driving in which marijuana use was suspected rose from about 19% of the total in the years before full legalization to 25% the year afterwards. “You are going to get a lot more adults driving with higher THC levels in their blood,” says Wayne Hall, a professor of public health policy who studies cannabis at the University of Queensland. Recent cannabis use approximately doubles accident risk, he says.
The device Lynn is blowing into – and which takes a few minutes to deliver a yes or no reading – is a Hound Labs prototype. And no, Lynn isn’t going to blow positive – he isn’t a user and it would be illegal for him to be high anyway because he is also a sworn police officer.
Accurately assessing recent marijuana use is surprisingly difficult. Saliva, urine and blood tests do not distinguish marijuana use in the past few hours from marijuana used yesterday or last week. THC is stored in body fat and can remain in saliva for several days and in blood and urine for weeks. In blood, for example, THC levels increase sharply upon smoking but then fall rapidly as impairment increases because the molecules dissolve into the fat (alcohol, in contrast, is water soluble).
In most states in the US, including California, impairment needs to be shown to prosecute someone for driving under the influence of marijuana. Get pulled over on suspicion and field sobriety testing – think the walk-and-turn exercise and tracking a pointer with your gaze – will ensue, followed, if you fail, by a blood test at the station. Yet it can be hard to get a conviction, says Jonathan Feldman, the legislative advocate for the California Police Chiefs Association, because defense attorneys can successfully argue first that the sobriety testing isn’t objective and second that presence in blood isn’t a legitimate indicator of impairment. “Whether you agree or disagree with marijuana legalization,” Lynn says. “We need to rapidly find technology that is going to provide objective data to get people off the road who shouldn’t be driving.”
Lynn argues breathalyzers solve the problem because of a neat relationship. The two-to-three-hour window of impairment – which he points out is what has been cited by the US National Highway Traffic Safety Administration – aligns with how long THC can be detected in breath. It also works if you have been munching on cannabis in food, says Lynn. It just takes longer to show up and longer to dissipate.
Hound Labs isn’t the only one racing to develop marijuana breathalyzers for the roadside. Others include Canadian-based Cannabix Technologies and Colorado-based Lifeloc Technologies, which already makes alcohol breathalyzers. Washington State University has a project, too – but it is in hiatus after state funding ran out last autumn. The companies’ breathalyzers all work in different, proprietary ways. Hound Labs uses filtering and chemicals to isolate the THC. Cannabix and Washington State’s are based on creating miniaturized mass spectrometers. Lifeloc, meanwhile, has licensed a lab test from Sandia National Laboratories. Lynn envisages selling versions of his device to consumers and employers as well as law enforcement agencies.
Thus far it would appear Hound Labs is furthest along – it has even set up a track test at an old navy base to study the type of impairment that cannabis causes in drivers. The company has done many hundreds of tests on human subjects to ensure the breathalyzer is accurate and has worked with police agencies to field test the device’s design, says Lynn. A company-funded clinical trial of the breathalyzer is currently being done at the University of California, San Francisco, with results expected in a few months. Cannabix, meanwhile, says it is getting close to completing its own trials of its device and has clinical trials in the wings.
Toxicology experts agree that a roadside marijuana breathalyzer would help to solve a growing public health problem – but they aren’t holding their breath it can be done.
Marilyn Huestis is a veteran of the field and former chief of chemistry and drug metabolism at the National Institute on Drug Abuse. Her 2013 study of THC in breath has been key to the current wave of breathalyzer interest. She says it is an open question whether handheld devices will be sensitive and specific enough to detect THC in the small concentrations that exist in breath. “We don’t know enough about these devices yet. So far nobody has published any data,” says Huestis, who is listed as an adviser to Cannabix on its website.
Kim Wolff, professor of addiction science at King’s College London, questions the impairment window. “It’s not accurate to suggest that impairment only occurs within a three-hour window after cannabis smoking or that this is well correlated with driving-related impairment,” says Wolff, citing a 2016 Swiss study that recommends an eight-hour delay before driving after being high.
Feldman, who represents California’s police chiefs, has met Lynn and thinks the breathalyzer may be an option but is waiting to see the results of clinical trials.
Lynn says Hound Labs will get its data out there. He agrees it is possible that some smokers will still be impaired after the three hours. But, he says, that’s no different from the situation with alcohol where some drivers may blow below a 0.08% blood alcohol concentration (the drink driving limit in US states) but still be impaired – which is why an officer can still arrest a driver even when the breathalyzer reading falls short. The Hound Labs breathalyzer, he says, “will pick up the people who are by far the most impaired.”