Originally published 2019-09-28.
Yes, a breath test for cannabis exists, at least as a working prototype. In 2019 chemists and engineers at the University of Pittsburgh built a handheld device that uses tiny carbon nanotubes to detect THC, the main psychoactive compound in cannabis, in a person's breath. In laboratory tests it picked up THC at levels comparable to, or better than, the mass spectrometry equipment used as the scientific gold standard. What it could not do, and what no device can yet do reliably, is tell how impaired a driver actually is. That gap is the real story behind the headlines about police soon carrying marijuana breathalyzers.

Courtesy of the University of Pittsburgh
This guide explains how the Pittsburgh device works, why measuring cannabis impairment is so much harder than measuring alcohol, how police test drivers today and what the research says about cannabis and road safety.
What the Pittsburgh researchers built
The work came from the lab of chemistry professor Alexander Star, together with Ervin Sejdic, then an associate professor of electrical and computer engineering at Pitt's Swanson School of Engineering. The lead author was chemistry doctoral student Sean Hwang. Their paper, "Tetrahydrocannabinol Detection Using Semiconductor-Enriched Single-Walled Carbon Nanotube Chemiresistors", appeared in the journal ACS Sensors in August 2019.
The team built a prototype that looks much like an alcohol breathalyzer: a person blows into it, and a sensor inside analyses the sample. Sejdic stressed that creating something usable outside the lab was a key step toward real-world use.
How the carbon nanotube sensor works
Carbon nanotubes are rolled sheets of carbon atoms, many thousands of times thinner than a human hair. The Pitt team used a type that behaves as a semiconductor, which only became available in sufficiently pure form a few years before the study. When molecules in a breath sample, including THC, stick to the surface of the nanotubes, they change how easily electric current flows through them.
The key signal is how quickly the current recovers afterwards. THC leaves a distinctive pattern. Because breath also contains water, carbon dioxide, alcohol, methanol and acetone, the researchers used machine learning to teach the system to recognise THC's signature even when those other substances were present.
The device was tested only in the laboratory, using prepared breath-like samples rather than people on the roadside.
Detecting THC is not the same as measuring impairment
With alcohol, decades of research link the concentration in breath or blood to predictable effects on driving, which is why many countries use a fixed legal limit. Cannabis does not work that way.
- THC lingers. It is stored in fat and can be detected in body fluids for days or even weeks, long after any effects have faded, especially in regular users.
- Tolerance varies. Two people with the same THC level can be affected very differently, depending on how often they use cannabis.
- Timing matters. THC levels in blood peak quickly after smoking and fall sharply, often while effects are still felt; edibles follow a slower curve.
Even the researchers acknowledged the problem. There is no agreed level of THC that defines impairment, and the scientific and legal debate over where to draw that line continues. A breath test may be useful in showing recent use, since THC is thought to stay in breath for a shorter time than in blood or urine, but recent use and dangerous impairment are not the same thing.
How police test drivers today
When the Pitt study appeared, 33 US states allowed cannabis in some form, medical or recreational. Officers suspecting cannabis-impaired driving typically rely on observation, standard field sobriety tests and, in many places, specially trained drug recognition experts. A blood test after arrest can confirm THC is present, but it cannot be done at the roadside and does not show impairment at the time of driving.
Some forces have trialled oral fluid tests, where a swab of saliva is analysed for recent drug use. Massachusetts State Police tested such swabs in 2018, and pilot programmes were reported in several other states, including Colorado, California, Kansas and Michigan. Critics, including some defence lawyers and civil liberties groups, argue that these tests share the same weakness as breath tests: they show recent use, not impairment, and may not hold up well in court.
That concern about fairness to legal users is legitimate. A person who used cannabis legally a day earlier should not face a drink-driving style charge simply because a trace remains in their body.
Does legal cannabis make roads safer or more dangerous?
Claims on both sides are often overstated, so it is worth looking at what studies actually found.
A 2016 Columbia University study in the American Journal of Public Health found that states with medical marijuana laws saw traffic deaths fall by about 11 percent on average after the laws took effect, with the clearest drop among drivers aged 25 to 44. The authors stressed that the study covered medical, not recreational, laws, that effects varied from state to state, and that factors such as reduced drinking and driving might explain part of the pattern.
Recreational sales tell a less comforting story. In 2018 the Highway Loss Data Institute estimated that collision insurance claims rose a combined 6 percent in Colorado, Nevada, Oregon and Washington after retail cannabis sales began, compared with neighbouring states. A separate analysis by the Insurance Institute for Highway Safety found police-reported crashes rose about 5 percent in Colorado, Oregon and Washington.
Single-year figures, such as a fall in Nevada's road deaths in the first year of legal sales, cannot show cause and effect. Road deaths swing from year to year for many reasons, including weather, fuel prices, enforcement and car safety.
What science says about cannabis and driving
According to the US National Institute on Drug Abuse, cannabis impairs judgment, coordination and reaction time, and higher THC levels are linked to greater impairment. Research reviewed by the institute suggests cannabis use roughly doubles the risk of a crash.
The same research shows why crash statistics are tricky. Cannabis is the illicit drug most often found in drivers involved in accidents, but because THC remains detectable long after use, a positive test does not prove the driver was impaired. Many of those drivers also had alcohol or other drugs in their system, and the combination of cannabis and alcohol appears more dangerous than either alone.
The practical takeaway is simple. If you have used cannabis, wait until you feel completely unaffected before driving, and longer after edibles, whose effects can last for many hours. Never combine it with alcohol behind the wheel. If you take cannabis for a medical condition and are unsure whether it affects your driving, talk to a doctor.
Where cannabis breath testing stands
The Pittsburgh prototype showed that detecting THC in breath is technically possible with a portable device. Several companies have since worked on similar products, but turning a lab result into a roadside tool that courts accept requires large field trials, independent validation and, above all, a scientifically sound definition of impairment.
Until that definition exists, any breath test will answer only a narrow question: has this person used cannabis recently? Deciding whether they were unsafe to drive will still depend on observation, behaviour and the judgment of trained officers, with all the strengths and flaws that brings.
Sources: University of Pittsburgh via ScienceDaily, 27 August 2019 (photo courtesy of the University of Pittsburgh); IIHS-HLDI, October 2018: crashes rise in first states with retail cannabis sales; Columbia Mailman School of Public Health, 2016: traffic fatalities and medical marijuana laws; National Institute on Drug Abuse: Does marijuana use affect driving?

COMMENTS