New test is able to detect THC in breath for the first time after eating or smoking cannabis

Federal scientists report the first detection of tetrahydrocannabinol (THC) in human breath after people ate cannabis edibles, marking a careful but notable step toward measuring recent use outside a lab.

There is a catch, though, and it is important. The signals vary widely from person to person and from one moment to the next, which means no one should rush to treat a single breath reading as proof of recent use or impairment.

Lead author Jennifer L. Berry is a research chemist at the National Institute of Standards and Technology (NIST). She worked alongside colleagues from the same institution and a partnering medical campus who contributed to the study’s design and analysis.

Why THC breath tests matter

Researchers have long wanted a simple roadside tool, but chemistry keeps getting in the way.

THC and related cannabinoids are large, sticky molecules that ride in airborne droplets rather than cleanly evaporating like alcohol, which complicates collection and measurement.

That is why teams now test both aerosol capture devices and exhaled breath condensate (EBC) approaches that chill the breath to trap what comes out.

Even with these methods, prior studies found THC hanging around before any observed use, a reminder that abstinence windows and device sensitivity can skew results.

What the study learned

The new study enrolled 29 adults, asked them to abstain ahead of time, and collected breath samples before consumption and again at about 47 minutes, 92 minutes, and 180 minutes after eating their own labeled edible product.

Two devices were used, one designed to catch aerosol particles and another that collects a cooled condensate.

The team then analyzed each sample for THC and 10 other cannabinoids using liquid chromatography tandem mass spectrometry.

Despite the abstinence requests, THC appeared in pre-use breath from 27 of 29 participants, ranging from trace levels to more than 0.4 ng per device.

After edible ingestion, 19 participants showed a significant increase at one or more post-use time points, 4 showed no change, and 6 showed a significant decrease.

“This proof-of-concept study shows that THC concentration in breath can increase after the ingestion of cannabis-infused edibles, but the uncertainty of breath measurements and a longer time window need to be further explored,” wrote Berry.

THC breath tests and baseline issues

“The finding of detectable and/or quantifiable THC before cannabis use shows how challenging it is to interpret a single THC concentration measurement in breath with no prior information about THC at baseline,” wrote Berry.

The most striking pattern, beyond the individual variability, is the high rate of detectable THC before anyone ate an edible.

That baseline problem is not unique to breath. A 2025 systematic review of studies connecting blood THC to driving performance found that most papers reported no linear relationship, even when driving metrics like lateral control and reaction time were measured with care.

Oral fluid tests are popular because they are easy to use, but their purpose matters.

An analysis of over 18,000 paired samples found that although THC in oral fluid can show whether THC is present in the blood, using it to identify drivers above standard blood limits would lead to many false positives. 

NIST’s own breath work in cannabis smokers supports taking repeated measurements rather than a one-off reading, since THC often falls sharply within hours and device methods differ in what they capture.

Put simply, the new edible study adds a missing piece, but it does not turn breath into an instant yes or no test. It suggests that trends over time, not single snapshots, are more promising for identifying recent use across routes of administration.

What this means

Traffic safety policy has leaned on numerical cutoffs for alcohol for decades. Cannabis does not play by the same rules, and federal safety agencies have said as much in reports to Congress that caution against equating blood THC numbers with impairment in the way we do for alcohol.

“If you have chronic users versus infrequent users, they have very different concentrations correlated to different effects,” said Frances Scott, a physical scientist at NIJ’s Office of Investigative and Forensic Sciences.

The team also tracked other cannabinoids. In many participants, CBN and CBG rose and fell in step with THC, while CBD sometimes did not, a hint that multi-analyte panels or ratios may carry more information than THC alone when the goal is to pinpoint recent use after edibles.

That pattern echoes earlier breath work showing some non-THC cannabinoids appear only after use and could act as markers when combined with THC kinetics. It is a direction that calls for standard methods, shared reference materials, and longer sampling windows.

Next steps for THC breath tests

This was a small, carefully designed study without standardized dosing, and participants brought their own edibles labeled THC that the researchers did not independently verify.

Sampling ended at 3 hours, which likely missed some peak values, and device differences could influence capture and recovery. None of the breath measures were tied directly to driving performance in this experiment.

A credible breath test will require validated collection protocols, calibration standards, and clear guidance on how to interpret rising or falling readings over time.

Those pieces are starting to come together in studies that map breath changes minute by minute and compare matrices side by side.

For now, the message is measured. Breath can carry THC after edibles, the kinetics can be tracked, and multi-timepoint strategies may be the fairest and most informative way to use that signal in real life.

The study is published in the Journal of Analytical Toxicology.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–