Per Se DUI Thresholds for THC
Blood-THC limits vary by state, but the science linking concentration to impairment remains weak.
This article is general information, not legal advice. Laws change; check current rules with official sources or a lawyer.
States copied alcohol's per se model for THC without the science to back it up. A blood-THC number doesn't predict impairment the way BAC does—regular users can blow past the limit stone sober, while occasional users can be dangerously impaired below it. Federal health agencies have said so explicitly. Know your state's rule and don't drive high, but understand that a per se conviction is about what's in your blood, not necessarily how you were driving.
What 'per se' actually means
A per se DUI law makes it illegal to drive with a specified concentration of a substance in your blood, regardless of whether you appear impaired. For alcohol, every U.S. state uses 0.08% BAC (0.05% in Utah) as the per se threshold, backed by decades of dose-response research linking BAC to crash risk [1].
States have adopted per se models for THC by setting nanograms-per-milliliter (ng/mL) blood thresholds. The fundamental problem: THC pharmacokinetics behave entirely differently from ethanol. THC is highly lipophilic (fat-soluble), redistributes into fatty tissue, and reappears in blood long after acute effects fade [2] Strong evidence. A regular consumer can register above a per se limit while completely sober; a naive consumer can be acutely impaired while below it.
This article is informational only and is not legal advice. Cannabis driving laws change frequently. Verify the current statute in your jurisdiction before relying on this information, and consult a licensed attorney if you face charges.
Last verified: October 2026.
How states set the numbers
As of October 2026, U.S. states fall into roughly four categories:
Zero tolerance or any-detectable-amount rules: Arizona, Georgia, Indiana, Iowa, Michigan (for non-patients), Oklahoma, Rhode Island, Utah, and Wisconsin make it illegal to drive with any measurable THC in the blood or, in some cases, the inactive metabolite carboxy-THC [3] Strong evidence. Because carboxy-THC persists in blood for weeks after active use, these laws effectively criminalize driving long after any impairment has ended.
Numerical per se limits: Illinois, Montana, Nevada, Ohio, Pennsylvania, and Washington set specific blood-THC thresholds, typically 1, 2, or 5 ng/mL [3] Strong evidence. These thresholds typically operate as strict liability: if you are above the limit, guilt is presumed and the prosecution need not prove impairment.
Permissible inference: Colorado uses 5 ng/mL as a permissible inference of impairment rather than a strict per se rule — a jury may, but is not required to, infer impairment at that level [4] Strong evidence.
Impairment standard: The remaining states and federal jurisdiction rely on traditional impairment criteria: prosecutors must prove the driver was actually impaired. Evidence often includes officer observations, Drug Recognition Expert (DRE) evaluations, and field sobriety tests, without a numerical blood-THC threshold.
Federal law has no per se THC threshold for civilian drivers on federal land; federal DUI cases use an impairment standard. Commercial motor vehicle drivers under DOT regulation face a separate urine-testing regime that does not set a blood-THC concentration limit [5] Strong evidence.
Why the science doesn't support per se limits
Two major government reports have questioned the validity of per se THC thresholds:
The National Highway Traffic Safety Administration (NHTSA), in its 2017 report to Congress, stated that "a specific drug concentration in the blood is not a reliable predictor of impairment" for THC [6] Strong evidence. The AAA Foundation for Traffic Safety, in a 2016 evaluation, reached the same conclusion and explicitly recommended that states not adopt per se limits [7] Strong evidence.
Tolerance: Daily cannabis users exhibit substantially smaller psychomotor effects at the same blood-THC concentration as occasional users [2] Strong evidence. This means the same blood level can reflect profound impairment in a naive user but minimal impairment in a chronic consumer.
Rapid distribution and delayed peak effects: Peak blood THC occurs within minutes of inhalation and drops sharply within one hour. However, peak subjective and behavioral effects lag behind blood levels [2] Strong evidence. By the time blood is drawn at a police station—often 30 minutes to hours after arrest—the measured concentration may not reflect the driver's impairment at the time of driving.
Edibles present a different profile: Oral THC produces lower peak blood levels but causes longer-lasting impairment compared to inhalation, so a fixed per se threshold can systematically underdetect impairment from edibles Weak / limited.
Residual THC in chronic users: Chronic users have been measured with blood THC above 5 ng/mL after multiple days of abstinence [8] Strong evidence, meaning a per se threshold automatically criminalizes some users driving days after their last use and any impairment.
Crash-risk data: A large NHTSA case-control study of crash-involved drivers found that, after adjusting for age, sex, and alcohol use, THC-positive drivers did not have a statistically significant increased crash risk [9] Disputed. This does not mean cannabis driving is safe; rather, it indicates the signal linking blood-THC to crash risk is much weaker and noisier than alcohol's dose-response relationship. Other crash studies report mixed or positive associations.
Practical legal consequences
If you live in a per se or zero-tolerance state and you consume cannabis legally, you face real legal exposure even when you feel sober:
- Medical cannabis does not provide a DUI defense: A medical cannabis card or registration does not shield you from a per se DUI charge in most states. The card legalizes possession and use at home but typically does not create a legal right to drive impaired.
- Implied-consent penalties: Refusing a blood draw in most states triggers automatic license suspension, separate from any criminal conviction.
- Police can obtain a blood warrant: Officers can request a court warrant for a blood draw after roadside arrest based on observed signs (odor, red eyes, field sobriety test performance). The resulting blood test is admissible in criminal proceedings.
- Conviction without a blood test: In impairment-standard states, prosecutors can still secure a conviction without any blood analysis if police testimony and Drug Recognition Expert evaluation are persuasive to a jury.
- Defense considerations: Effective defenses typically focus on chain-of-custody challenges, the scientific validity of the blood assay, the time elapsed between driving and the blood draw, and evidence of tolerance—all matters requiring representation by an attorney qualified in DUI defense and, ideally, familiar with cannabis pharmacology.
How thresholds have been changing
State per se thresholds and rules have been shifting in both directions:
Tightening: Some states have considered narrowing permissible-inference rules into stricter per se status, though major changes have been limited as of October 2026.
Loosening: Michigan revised its zero-tolerance rule in 2021 (House Bill 4308) so that registered medical cannabis patients are judged by an impairment standard rather than any-detectable-amount [10] Strong evidence. Nevada in 2021 eliminated its 2 ng/mL per se threshold for non-fatal driving cases via Assembly Bill 400, retaining a per se rule only for fatal crashes [11] Strong evidence.
Federal research direction: NHTSA continues to fund research into oral-fluid (saliva) testing and impairment-based detection methods rather than advocating for a uniform national blood-THC standard [6] Strong evidence.
Last verified: October 2026. Statutes change with each legislative session. Always confirm the current text of the law and any case-law interpretations in your jurisdiction before acting. This remains informational only and not legal advice.
Sources
- Government National Highway Traffic Safety Administration. Alcohol-Impaired Driving. Traffic Safety Facts.
- Peer-reviewed Huestis MA. Human Cannabinoid Pharmacokinetics. Chemistry & Biodiversity. 2007;4(8):1770-1804.
- Government National Conference of State Legislatures. Drugged Driving: Marijuana-Impaired Driving Laws.
- Government Colorado Revised Statutes § 42-4-1301(6)(a)(IV). Permissible inference for THC concentration of five nanograms or more.
- Government U.S. Department of Transportation. 49 CFR Part 40 — Procedures for Transportation Workplace Drug and Alcohol Testing Programs.
- Government National Highway Traffic Safety Administration. Marijuana-Impaired Driving: A Report to Congress. 2017.
- Reported AAA Foundation for Traffic Safety. An Evaluation of Data from Drivers Arrested for Driving Under the Influence in Relation to Per Se Limits for Cannabis. 2016.
- Peer-reviewed Karschner EL, Schwope DM, Schwilke EW, et al. Predictive model accuracy in estimating last Δ9-tetrahydrocannabinol (THC) intake from plasma and whole blood cannabinoid concentrations in chronic, daily cannabis smokers administered subchronic oral THC. Drug and Alcohol Dependence. 2012;125(3):313-319.
- Government Compton RP, Berning A. Drug and Alcohol Crash Risk: A Case-Control Study. NHTSA Traffic Safety Facts Research Note, DOT HS 812 117. 2015.
- Government Michigan Legislature. House Bill 4308 of 2021, amending MCL 257.625 — eliminating zero-tolerance for registered medical marijuana patients.
- Government Nevada Legislature. Assembly Bill 400, 2021 Session — revising provisions related to driving with prohibited substances in blood.
How this page was made
Generation history
Drafted with AI assistance and checked by a separate automated fact-check pass. Not every article has been read by a person. See how articles are made.
Related
- Cannabis Laws in the Netherlands — Tolerated but not legalized: a guide to Dutch cannabis policy and coffeeshops.
- UN Drug Treaties and Cannabis — Three UN treaties control cannabis globally; legalization exists despite treaty language f...