In Some States, Cannabis Users are Always Driving Illegally
Imagine - you’re driving to work one morning, completely sober, and you get a DUI because you used legal marijuana to help you sleep 2 days ago. That’s happening today, thanks to THC per se limits and oral fluid testing devices.
Across the U.S., several states have enacted per se limits for Delta-9 THC (Δ⁹-tetrahydrocannabinol) in drivers’ blood or saliva, much like the alcohol per se limits (0.08% BAC) that exist throughout the country. Under these laws, any driver whose THC exceeds a set threshold (regardless of actual impairment) is presumed to be “driving under the influence.”
These limits were created based on the legal framework for alcohol impaired driving in an effort to police legal cannabis users and prevent impaired driving. While that limit is scientifically valid for alcohol, given its linear metabolism and predictable effects based on blood-alcohol content, the same is not true for virtually any other drug, including cannabis. In practice, THC content in the body (including blood, breath, saliva / oral fluid, urine, etc.) is totally uncorrelated with the amount of impairment being experienced, if any. THC remains in the body far longer after cannabis use, up to several weeks, depending on the frequency of use. All chemical tests for THC are tests for prior use, not impairment, and cannot tell us anything about the mental state of the test subject.
The per se limits set by states in a misguided attempt to police impaired driving range from 1.0 to 5.0 ng/ml of body fluid. Unfortunately, regular cannabis users can easily have this amount of THC in their body for many days since the last use. In practice, this means that a regular or recent cannabis user can easily test over the limit, even when completely sober.
Below we summarize the states with THC per se laws, review the science on THC detection in body fluids, and discuss why fixed limits are a poor proxy for impairment. We then discuss how impairment-based testing – notably the Gaize device that automates Drug Recognition Expert (DRE) eye exams – offers a fairer path to road safety.
States with THC Per Se Limits
Currently, the following states have per se THC thresholds (beyond zero-tolerance) for driving, typically defined in ng/mL of whole blood. The links below point to the enabling statutes. These laws often apply to all THC, or to the active Δ⁹-THC and sometimes its metabolite (THC-COOH) in combination:
THC Per Se Limits by State
| State | Per Se Limit | Code | Link |
|---|---|---|---|
Illinois | 5.0 ng/mL | 625 ILCS 5/11-501.2 | |
Montana | 5.0 ng/mL | Mont. Code §61-8-1002 | |
Nevada | 2.0 ng/mL | Nev. Rev. Stat. §484C.110(4) | |
Ohio | 2.0 ng/mL | Ohio Rev. Code §4511.19 | |
Pennsylvania | 1.0 ng/mL | 75 Pa. C.S. §3802(d) | |
Washington | 5.0 ng/mL | RCW 46.61.502 | |
West Virginia | 3.0 ng/mL | W.V. Code §16A-5-10 | code.wvlegislature.gov |
| Colorado | 5.0 ng/mL (Reasonable Inference) | § 42-4-1301 (6), C.R.S. | leg.colorado.gov |
As one recent study notes, these cutoffs were set “without strong evidence linking them to impairment.” For example, Illinois and Nevada (2–5 ng/mL) were explicitly modeled on old alcohol rules, not on rules created for drug-impaired-driving, or on any science. In practice, they expose chronic or recent cannabis users to consistent legal risk: THC can linger above 5 ng/mL for days in a steady user, even when they are completely sober.
Washington v. Frazer - An Egregious Court Error
Washington v. Frazer is a Washington State DUI case in which a legal cannabis user was charged with impaired driving based on Washington state’s nonsensical per se limit for THC, despite the defendant not being impaired at the time of arrest. The defendant challenged the constitutionality and scientific validity of the state’s per se 5 ng/mL THC blood limit, arguing that it does not reliably indicate actual impairment. The court ultimately upheld the law, finding that the legislature had a rational basis for adopting a numeric threshold even amid conflicting scientific evidence about the relationship between THC concentration and driving performance. The decision reinforced the state’s ability to prosecute cannabis DUI based solely on blood THC levels, without requiring proof of real-time functional impairment.
The ruling in Washington v. Frazer reads less like a careful evaluation of scientific evidence and more like a judicial shrug in the face of decades of pharmacological research showing there is no reliable correlation between blood THC concentration and functional impairment. By treating an arbitrary per se number as if it were equivalent to alcohol’s well-validated BAC framework, the court effectively blessed a policy built on political expediency. Its suggestion that a numeric THC threshold can serve as a stand-in for real-time impairment ignores the overwhelming consensus that THC pharmacokinetics are wildly variable across individuals, routes of administration, and time since use. Most egregious is the court’s reliance on the notion that the limit is justified because it is “rational” and promotes “uniform enforcement” — as if administrative simplicity can substitute for scientific validity, and as if uniform injustice is somehow better than evidence-based standards. In doing so, the decision elevates a legal fiction over measurable human performance, undermines the credibility of impairment enforcement, and entrenches a standard that researchers have been dismantling for decades.
This case serves as a critical example of why per se limits should be vigorously and consistently challenged. If it’s the aim of politicians to prosecute safe and responsible cannabis users, per se limits are the way they’ll do it. If, however, safety is the goal, these laws should be rejected and rational, measurable impairment standards used instead.
THC in Saliva: Not A Proxy for Impairment
You might think that relying on so called “recent use tests” (typically oral fluid) are a way to allow law enforcement to enforce safety in legal marijuana states. However, researchers have quantified how long THC stays detectable in oral fluid and there’s no evidence that these tests are useful as proxies to impairment either.
In controlled studies, after a single cannabis cigarette (with just 6–7% THC), saliva THC spikes immediately and then decays slowly, over many hours. For example, Anizan et al. found that all subjects (frequent or occasional smokers) had THC-positive saliva for up to ~13.5 hours after smoking (pubmed.ncbi.nlm.nih.gov), and the decline extended beyond 30 hours when using lower cutoff levels. Modern cannabis, with 15-30% THC, will significantly extend the detection window and THC concentrations.
Moreover, screening devices (point-of-collection testing devices) lack precision. Field studies of the Dräger DrugTest 5000 and the Abbott SoToxa, two common saliva analysis tools, found they gave a false positive for THC in about 15% of cases. That is, drivers with only trace oral THC (or none) often tested “positive” on the device. Some false negatives also occurred (the Dräger missed ~13% of blood-confirmed DUI cases). Other devices have similar issues: a review notes that any oral fluid test at 25 ng/mL cutoff still yields “a large number of ‘false positives’ for illegal THC” when used against strict state limits (sciencedirect.com, pubmed.ncbi.nlm.nih.gov). In practice, saliva tools often confuse residual mouth contamination or passive exposure with impairment, casting doubt on their accuracy (pubmed.ncbi.nlm.nih.gov, sciencedirect.com).
How long can saliva tests detect THC?
Studies consistently show THC from smoked cannabis can linger in mouth fluids long after use. A 2019 study of oral fluid testing found median detection windows of 6–30+ hours for ∆⁹-THC. Even casual smokers may test positive 3–8 hours post-dose, while chronic smokers often test positive well into the next day or two (aaafoundation.org, adrasandaltiglaw.com). By contrast, the actual effects of one cannabis dose on driving performance last only a few hours. This mismatch is why researchers warn that oral fluid tests measure recent use, not real-time impairment (pubmed.ncbi.nlm.nih.gov, mgaleg.maryland.gov).
States’ rush to deploy oral-fluid THC devices as if they were THC breathalyzers is not just bad science, it is public policy malpractice. These tests do not measure impairment, cannot quantify dose or recency of use, and have no validated relationship to driving performance; they merely detect the presence of a molecule that can linger long after any psychoactive effect has vanished. Independent data from Michigan showed false-positive rates for the Abbott SoToxa ranging from roughly 11% to 24% when compared to blood confirmation, meaning officers were being pointed toward arrest decisions by a device that was wrong as much as one out of every four times.
Edibles and Oral Fluid Tests: a Huge Blind Spot
Orally consumed THC from edible cannabis is even harder to spot with saliva tests. When someone eats a THC brownie or gummy, the drug is absorbed through the gut, not the mouth. There is limited, if any exposure of THC to saliva. Therefore, there is little or no immediate THC residue in saliva after edibles. THC does not readily pass from the bloodstream to saliva, meaning that orally consumed THC can be missed entirely by oral fluid tests. One study explicitly found that the parent THC was undetectable in saliva of oral-cannabis users at all post-dose times.
In plain terms: an edible-induced high often produces a negative mouth swab drug test. This means that even an irresponsible user taking legal cannabis orally could drive totally impaired, yet pass an oral fluid drug test - potentially allowing them to escape undetected from a law enforcement interaction.
Scientific reviews underscore this limitation. The Lambert Initiative (Univ. of Sydney) pooled data on driving and found that THC blood or saliva levels correlate weakly or inconsistently with actual impairment, in this study. A driver might ingest a large edible dose and feel “extremely unfit to drive” yet register negligible or zero THC in saliva. Hence saliva tests give a false sense of security: they can miss impaired edible users altogether, or catch non-impaired smokers who happen to have lingering THC.
Logical Consequences: Fairness and Public Safety
Given these scientific facts, the logical conclusion is clear: using THC detection as a proxy for impiarment is fraught. Tests cannot detect intoxication, but they will ensnare sober users. Research and expert commentary draw this conclusion in no uncertain terms. For example, a detailed review of U.S. laws notes that “per se laws are particularly unfair to habitual cannabis users,” since THC can remain above the legal limit long after any impairment (adrasandaltiglaw.com). Indeed, a chronic user who smokes nightly will likely always exceed a 2–5 ng/mL cutoff, effectively rendering them guilty every time they drive.
Several studies find no clear impact to crash risk increase below about 5 ng/mL of blood THC (adrasandaltiglaw.com), suggesting these low per se thresholds are arbitrary. State-level analysis echoes these concerns. The National Traffic Safety Administration warns that low THC cutoffs could label “unimpaired individuals” as DUI offenders (sydney.edu.au). In another example, Lambert Initiative researchers explicitly conclude:
“Our results indicate that unimpaired individuals could mistakenly be identified as cannabis-intoxicated when THC limits are imposed by law.” (sydney.edu.au)
In other words, many “failed” drivers under per se laws would not be actually impaired. Conversely, impaired drivers (especially frequent users) often test below the limit and slip through. This undermines public safety and erodes trust.
Impairment Testing: A Better Way Forward
Rather than relying on chemical thresholds, traffic safety experts advocate impairment-based detection for drugged driving. The U.S. Drug Recognition Expert (DRE) program – founded by NHTSA – trains officers to use validated ocular and psychomotor tests to assess impairment from drugs (nhtsa.gov). A trained DRE checks a driver’s gaze tracking, pupil response, coordination, vital signs, and more. This takes about 30 minutes and has proven to be an accurate technique for detecting drug induced impairment.
DRE’s can’t be everywhere though. New technologies like Gaize are now automating these DRE methods. Gaize is a wearable headset that conducts fully automated DRE eye exams, measures subtle eye movements at high precision, and records video evidence. In effect, it can act like an electronic test for marijuana impairment. Within minutes it evaluates ocular indicators (smooth pursuit, gaze nystagmus, pupil reactivity, etc.) that are scientifically linked to impairment. Because it assesses actual physiological signs of intoxication, rather than presence of THC, Gaize can detect a genuinely impaired driver (regardless of whether they smoked or ate THC) and not falsely accuse a sober, responsible cannabis user.
Legally and practically, impairment tests have major advantages. They focus on how a driver is acting now, not what they consumed earlier. Gaize thus aligns with the scientific consensus: the goal is to catch impairment, not legal consumption. Early field reports (e.g. in Australia) indicate that relying on saliva THC for roadside stops has already missed real drugged drivers and wrongly flagged sober users (sydney.edu.au, Washington v. Frazer). In contrast, an impairment-screening device could preserve justice and public safety by ensuring only truly unsafe drivers are charged.
Conclusion
Per se THC limits and oral fluid tests might seem like straightforward ways to create public policy around THC impaired driving, but in practice they can both ensnare safe and sober drivers but totally miss impaired ones. Scientific studies and policy analyses all point to the same conclusion: THC levels do not reliably indicate impairment. In fact, the following governmental organizations and puplic policy groups have said that per se limits for THC should not be used:
NHTSA (again) - https://www.nhtsa.gov/sites/nhtsa.dot.gov/files/812117-drug_and_alcohol_crash_risk.pdf
AAA - https://aaafoundation.org/wp-content/uploads/2017/12/CannabisUseAmongDriversInWashington.pdf
Reason - https://reason.com/wp-content/uploads/2019/05/CRS-report-on-marijuana-and-highway-safety-2019.pdf
The solution is to shift to impairment-based standards. Technologies like Gaize – which objectively apply the proven DRE ocular tests promise a fair alternative. By measuring how a person’s body is responding to a drug, rather than measuring the drug itself, such systems can target only truly impaired driving and protect all others from wrongful DUI charges. For legislators, the message is clear: Replace chemical per se limits with validated impairment testing to keep roads safe without criminalizing safe and responsible cannabis use.
