A Small Pilot Study Shows Why a Reliable Cannabis Breathalyzer Remains Elusive
A roadside test that could distinguish recent cannabis use from residual THC remains a scientific challenge. In a 2023 pilot study, researchers from the National Institute of Standards and Technology (NIST) and the University of Colorado Boulder found that THC measurements in breath varied too widely for a single reading to reliably show whether someone had smoked cannabis within the previous hour.
The study, published in the Journal of Breath Research, was designed primarily to evaluate a method for collecting and measuring cannabis-related compounds in exhaled breath—not to validate a finished handheld breathalyzer. Unlike alcohol, which is readily exhaled as a vapor, THC has very low volatility and is thought to travel in tiny aerosol particles produced in the respiratory tract. Those particles can be difficult to collect consistently.
The researchers analyzed 42 breath samples from 18 participants in Colorado. Participants used the same legal-market cannabis flower, containing approximately 25% THC-A, and provided breath samples during an initial intake visit and again during an experimental session four weeks later. During the experimental session, samples were collected about 15 minutes before cannabis use and roughly one hour afterward. The samples were captured with an impaction-filter device and analyzed in the laboratory using liquid chromatography with tandem mass spectrometry.
Of the 14 participants who supplied paired breath samples during the experimental session, only eight showed the expected increase in THC after cannabis use. THC was not detected in three post-use breath extracts, while the remaining post-use samples were similar to or lower than the corresponding baseline samples. Across all samples, THC was quantified in 31% of intake-session breath extracts, 36% of pre-use experimental extracts and 80% of samples collected one hour after use.
The team also collected blood immediately after participants returned from using cannabis. Those measurements were used mainly to assess whether participants had followed the study protocol, rather than to establish blood THC as a dependable measure of impairment or a practical roadside alternative. Twelve participants showed the expected increase in plasma THC, but three had no detectable THC immediately after the reported use.
The results underscore two separate problems: detecting THC and interpreting what its presence means. Even if a breath test accurately detects THC, the result may indicate recent exposure without showing whether a person is impaired. Federal transportation-safety guidance notes that, unlike alcohol, blood THC concentrations do not have a consistent, established relationship with driving impairment. The National Highway Traffic Safety Administration’s drug-impaired-driving guidance describes this distinction and the limitations of current testing methods.
The pilot study was small and was not intended to produce definitive statistical conclusions. Cannabis use occurred in participants’ homes and was not directly observed, breath collection conditions varied, and the study included only a limited number of post-use time points. The authors said larger studies with verified abstinence, standardized breathing procedures and repeated measurements will be needed before researchers can determine whether breath testing can distinguish recent use from THC left over from earlier consumption.
For now, the findings do not rule out a future cannabis breath test. They do show that a single THC measurement should not yet be treated as a reliable indicator of recent use—and, by itself, cannot establish impairment. NIST’s broader cannabis-breathalyzer research program is focused on improving collection methods, understanding the behavior of cannabinoids in breath and developing the measurement standards that a dependable device would require.
