The Cannabis DUI Debate: Why THC Blood Tests Fail and the Search for a Better Solution
Alcohol has a breathalyzer. A 0.08% blood alcohol concentration reliably indicates impairment in the vast majority of people. Law enforcement has a tool that is fast, non-invasive, field-deployable, and legally defensible. Cannabis has nothing comparable, and this gap represents one of the most vexing policy challenges in the legalization era.
As more states legalize cannabis, the question of how to identify and prosecute cannabis-impaired driving has become urgent. The approaches currently in use are scientifically flawed, legally contested, and potentially unjust. The search for something better is underway, but a reliable solution remains frustratingly out of reach.
Why THC Blood Levels Do Not Work
The fundamental problem is pharmacokinetic: THC does not behave like alcohol in the body. Alcohol distributes evenly through body water, rises predictably after consumption, and clears at a roughly constant rate. Blood alcohol concentration correlates reliably with impairment. THC does none of these things.
THC is lipophilic — it dissolves in fat, not water. After inhalation, THC blood levels spike within minutes to levels that may exceed 100 ng/mL, then plummet within 1-2 hours to single digits as THC redistributes from the blood into fatty tissues throughout the body. Peak impairment roughly coincides with peak blood levels after smoking, but the relationship is not as clean as with alcohol.
The bigger problem is chronic use. Regular cannabis consumers accumulate THC in their fat stores, which slowly releases back into the bloodstream over days and weeks. A daily cannabis user may have a resting blood THC level of 1-5 ng/mL without having consumed cannabis in 24 hours and without being impaired in any way. A study published in Clinical Chemistry found that 30% of daily cannabis users tested above Colorado’s 5 ng/mL per se limit a full 24 hours after their last use.
This means that per se THC limits — laws that make it illegal to drive with THC above a certain blood concentration — can convict sober people while failing to identify others who are genuinely impaired. It is a fundamental mismatch between the measurement and the thing being measured.
Per Se Limits by State
Despite the scientific problems, several states have enacted per se THC driving limits:
5 ng/mL THC in whole blood: Colorado, Washington, Montana. This is the most common threshold but also the most criticized. As noted above, frequent cannabis consumers can exceed this threshold while completely unimpaired.
2 ng/mL THC in whole blood: Ohio (for medical patients), with some states considering similar lower thresholds.
Zero tolerance: Arizona, Delaware, Georgia, Indiana, Iowa, Michigan, Oklahoma, Rhode Island, Utah, and Wisconsin have zero-tolerance laws where any detectable amount of THC constitutes a per se violation. These laws are the most problematic because THC metabolites can be detected in blood for days to weeks after last use.
No per se limit: California, New York, Oregon, and several other legal states have rejected per se THC limits, instead requiring prosecutors to prove actual impairment through observed behavior, field sobriety testing, and expert testimony. This approach is more scientifically sound but harder to prosecute consistently.
The inconsistency across states creates confusion for consumers, particularly those who travel between legal states and may not realize they could face DUI charges in one state for a THC blood level that would be legal in another.
Oral Fluid Testing: A Better Approach?
Oral fluid (saliva) testing has emerged as the most promising near-term alternative to blood testing. THC is present in oral fluid primarily through direct oral deposition during smoking or ingesting, and its detection window is much shorter than blood testing — typically 6-12 hours for occasional users and up to 24-48 hours for heavy users.
This shorter detection window is precisely the advantage. Oral fluid testing is more likely to identify recent cannabis use that correlates with potential impairment, rather than detecting residual THC from days or weeks ago.
Several jurisdictions have adopted or piloted oral fluid testing programs. The UK and Australia have used roadside oral fluid testing for cannabis for several years. In the United States, Michigan launched a pilot program in 2024, and Alabama authorized oral fluid testing in 2025.
The Draeger DrugTest 5000 and the SoToxa Mobile Test System are the two devices most commonly used in law enforcement. Both can produce a result in 5-10 minutes from a roadside oral fluid sample, detecting THC above a threshold of approximately 25 ng/mL in oral fluid.
However, oral fluid testing has its own limitations. The detection threshold can be exceeded by passive exposure to cannabis smoke in enclosed spaces, though this requires fairly extreme conditions. Eating, drinking, or using mouthwash shortly before testing can affect results. And the core problem remains: detection of THC presence is not the same as measurement of impairment.
The Cannabis Breathalyzer Race
Multiple companies are developing breathalyzer-like devices that detect THC in exhaled breath. The appeal is obvious — a device as fast and familiar as an alcohol breathalyzer that could detect very recent cannabis use.
Hound Labs has been the most visible company in this space, developing a device that detects THC in breath at the parts-per-trillion level. Their technology uses a proprietary chemical capture and detection method that can identify THC molecules in the tiny quantities present in exhaled breath.
The scientific premise is that THC is detectable in breath only for approximately 2-3 hours after inhalation, making breath detection a much tighter indicator of very recent use compared to blood or oral fluid testing. If you test positive on a THC breathalyzer, you almost certainly consumed cannabis within the last few hours.
However, even breath testing does not solve the fundamental impairment correlation problem. A person who consumed a small amount of cannabis two hours ago might test positive on a breath test while being less impaired than someone who consumed a large amount four hours ago and tests negative. And edible consumers — who never inhale THC — may not be detectable at all, despite being significantly impaired.
Behavioral Assessment: The Most Honest Approach
Many traffic safety experts argue that the field is chasing the wrong goal. Instead of seeking a chemical measurement that correlates with impairment, they advocate for standardized behavioral assessments that directly measure what matters: the driver’s ability to operate a vehicle safely.
Drug Recognition Expert (DRE) programs train law enforcement officers to conduct systematic behavioral evaluations that assess eye tracking, balance, coordination, vital signs, and other indicators of drug impairment. DRE evaluations are more comprehensive than standard field sobriety tests and are designed to identify the category of substance causing impairment.
The Advanced Roadside Impaired Driving Enforcement (ARIDE) program provides a shorter training for a broader pool of officers, focusing on recognizing signs of drug impairment during routine traffic stops.
Critics note that behavioral assessments are inherently subjective, creating concerns about inconsistent application and potential bias. They also require significant training investment — a full DRE certification involves 72 hours of classroom training plus supervised field evaluations, and there are currently only about 10,000 certified DREs in the United States, far too few to provide consistent coverage.
The Technology on the Horizon
Research is exploring more sophisticated approaches. Functional near-infrared spectroscopy (fNIRS) devices can detect changes in prefrontal cortex activation associated with cannabis impairment through a non-invasive headband-like device. While far from field-ready, this technology directly measures brain function rather than chemical presence, making it a fundamentally different approach to the impairment question.
Tablet-based cognitive testing apps that measure reaction time, divided attention, and decision-making speed are also being studied as supplements to traditional field sobriety tests. These tools produce quantitative data that is less subjective than officer observation, though questions remain about their legal admissibility and the establishment of impairment thresholds.
Where Policy Should Go
The honest assessment in 2026 is that no single test or technology can reliably determine cannabis impairment in the way that a breathalyzer determines alcohol impairment. Given this reality, the most defensible policy approach combines multiple evidence sources: oral fluid or breath testing to establish recent use, standardized behavioral assessment to evaluate actual impairment, and officer testimony and dashcam or body camera footage to document driving behavior that triggered the stop.
Per se THC limits based on blood concentration should be reconsidered by states that have them. The scientific consensus is clear that these limits do not reliably distinguish impaired from unimpaired drivers, and they disproportionately penalize medical cannabis patients and frequent consumers who may be completely sober at the time of testing.
The consequences of getting this wrong are significant in both directions. Failing to identify genuinely impaired drivers puts public safety at risk. Convicting unimpaired drivers undermines public trust in the legal system and in cannabis legalization itself. For an industry already facing banking challenges and marketing restrictions, the stigma of an unresolved impaired driving problem creates additional political headwinds.
The search for a cannabis breathalyzer is really a search for certainty in a domain where the biology simply does not cooperate. The sooner policymakers accept that and build multi-layered assessment systems rather than waiting for a silver bullet device, the sooner we can have impaired driving laws that are both effective and fair.