Breaking

Latest criminal justice updates and case developments

GAVELDAILY
DUI/DWI

Per Se THC Limits Are Failing as Junk Science: How Defenders Are Beating Marijuana DUI Cases in 2026

Six states still convict drivers on a blood THC number alone, but federal regulators, peer-reviewed pharmacology, and a 2025 wave of legislative reform now agree the science does not support it. Here is how defense lawyers are dismantling per se cannabis DUI cases.

Gavel Daily EditorialAI-Assisted
Share
Per Se THC Limits Are Failing as Junk Science: How Defenders Are Beating Marijuana DUI Cases in 2026

For more than a decade, prosecutors in a handful of states have asked juries to convict drivers of marijuana DUI on a single number: a nanogram count of Delta-9-THC in a blood tube drawn hours after the stop. In 2026, that approach is breaking down. Federal regulators, peer-reviewed pharmacology, the country's largest laboratory-medicine society, and a wave of statehouse activity are converging on the same finding: blood THC does not measure impairment, and the bright lines built on it cannot bear the weight of a criminal conviction.

This article maps the per se landscape, walks through the science driving courts and legislatures to rethink it, and lays out the cross-examination tools defense attorneys are using to win these cases.

1. The per se map: which states convict on a number

Six states impose per se cannabis DUI thresholds, meaning a blood result above the cutoff is, by itself, sufficient for conviction:

  • Pennsylvania: 1 ng/mL of Delta-9-THC

  • Nevada: 2 ng/mL

  • Ohio: 2 ng/mL (rising to 5 ng/mL if SB 55 becomes law)

  • Illinois: 5 ng/mL

  • Montana: 5 ng/mL

  • Washington: 5 ng/mL

Colorado takes a different path. Under HB 13-1325 (2013), a 5 ng/mL result creates a permissible inference of impairment that the defense can rebut with evidence; jurors are not required to convict on the number alone. The Colorado Department of Transportation describes the cutoff in exactly those terms. Most other legal-cannabis states use effects-based statutes that require proof of actual impairment, the same way DUI is proven for prescription medication.

The Governors Highway Safety Association and the National Conference of State Legislatures both maintain current trackers showing the split.

2. What the science actually says

Federal and independent research has been telling the same story for nearly a decade.

NHTSA's 2017 Report to Congress (DOT HS 812 440) concluded that blood or oral-fluid THC level "is not an accurate and reliable predictor of impairment" and that setting per se levels for THC is "not meaningful." The AAA Foundation for Traffic Safety in 2016 reached the same conclusion: a 5 ng/mL cutoff lacks scientific support and misclassifies drivers in both directions. AAA's December 2024 cannabis fact sheet reaffirms that no THC concentration reliably indicates impairment.

Three pharmacological realities drive that conclusion.

The timing curve. Peak psychomotor effects from inhaled cannabis arrive roughly 90 minutes after use. By that point blood THC has already fallen more than 80 percent from its peak. The blood number drawn at a hospital, often one to three hours after the stop, almost never represents the driver's THC level at the moment of driving, and it does not reverse-extrapolate the way ethanol does.

Chronic-user carryover. Pharmacokinetic work from the Huestis laboratory showed that frequent users retain detectable THC for four to five days after last use. Many remain above 2 ng/mL well past 12 hours of abstinence, and a meaningful fraction exceed 5 ng/mL with no acute behavioral effects. A 2025 study indexed in PubMed Central reported that more than half of chronic users still exceeded 5 ng/mL after a 12-hour abstinence with no measured impairment.

Tolerance. A separate Huestis-group study on heavy chronic smokers found a clear dissociation between blood THC and behavioral performance: the same number means very different things in a tolerant user and a naive one.

A 2022 Scientific Reports paper put the point bluntly: THC blood and breath levels are indeterminate proxies for impairment. A 2021 simulated driving study found per se thresholds neither identify impaired drivers nor exclude unimpaired ones.

3. The 2025 evidence wave

Two developments in late 2025 have given the defense bar fresh ammunition.

First, a peer-reviewed semi-mechanistic pharmacometrics model by Li and colleagues, published in the Journal of Clinical Pharmacology, evaluated all three common cutoffs. The 1 ng/mL threshold carried the highest false-positive risk. The 2 ng/mL and 5 ng/mL cutoffs remained pharmacodynamically inconclusive, meaning the model could not link them to a defined level of impairment.

Second, in November 2025 the Association for Diagnostic and Laboratory Medicine publicly warned that current cannabis blood concentration thresholds "could land innocent people in jail." That statement comes from the professional society for the laboratory scientists whose machines produce the very numbers prosecutors put on the screen. It is not advocacy. It is the lab community telling courts the test cannot do what the statute asks of it.

4. Daubert and Frye in practice

Defense attorneys are using these materials at the gatekeeping stage. Under Daubert (federal courts and most states) or Frye (still controlling in jurisdictions like Pennsylvania, Illinois, and Washington), the trial judge must screen expert testimony for scientific reliability before a jury hears it. A judicial-perspective analysis published in PMC walks through how that screening should function in marijuana driving cases.

Common defense motions in 2026 ask the court to:

  • Exclude or limit expert testimony that a specific blood THC number proves impairment, on the ground that no validated dose-response relationship exists.

  • Exclude any "reach-back" or retrograde extrapolation testimony, because, unlike ethanol, THC pharmacokinetics do not support it.

  • Bar testimony that 11-OH-THC or Carboxy-THC concentrations indicate current impairment.

  • Constrain Drug Recognition Expert (DRE) testimony to observations rather than scientific conclusions, given the documented confirmation-bias and validation concerns flagged by NHTSA's own report.

  • Require the state's expert to concede on the record the conclusions of the NHTSA 2017 report, the AAA 2016 and 2024 materials, and the ADLM 2025 statement.

Even where courts decline to exclude the evidence outright, limiting instructions and admitted concessions reshape what the jury actually weighs.

5. Metabolites, affirmative defenses, and the Arizona model

Arizona has produced the cleanest appellate framework for separating impairment from chemistry. In Dobson v. McClennen (2015) and State ex rel. Montgomery v. Harris, the Arizona Supreme Court limited DUI metabolite liability to impairing metabolites. Carboxy-THC, the inactive metabolite that lingers for weeks, cannot sustain a conviction on its own.

The court also recognized an affirmative defense for Arizona Medical Marijuana Act cardholders: a defendant can show that any THC present was "in a concentration insufficient to cause impairment." An April 2025 practitioner explainer walks through how that defense is now being raised at trial.

Outside Arizona, defense lawyers are borrowing the framework: distinguish parent THC from inactive metabolites in cross-examination, force the chemist to identify exactly which compound the lab measured, and put the affirmative-defense theory in front of the jury whether or not the statute names it.

6. Legislative reform 2025 to 2026

The most significant statutory development is in Ohio. On October 8, 2025, the Ohio Senate unanimously passed SB 55. As detailed in the Columbus OVI Attorney Blog and Signal Ohio, the bill:

  • Raises Ohio's per se OVI threshold from 2 ng/mL to 5 ng/mL of Delta-9-THC.

  • Eliminates conviction based solely on inactive metabolites.

  • Converts the 2 to 5 ng/mL band into a rebuttable inference rather than per se proof, mirroring Colorado's approach.

The bill awaits House action. The Marijuana Policy Project and several state bar groups continue to recommend effects-based statutes that require proof of actual impairment, treating cannabis the way prescription-medication DUI is treated. The Congressional Research Service has summarized the federal-level picture and the persistent measurement problems for legislators considering similar reforms.

7. Cross-examination playbook for 2026

For attorneys preparing trial, the operative cross-examination moves cluster around eight points:

  • Sample timing. How many minutes elapsed from stop to draw? What was the driver's THC at the wheel? The state's expert cannot answer that question with confidence, and saying so on the record matters.

  • No retrograde extrapolation. Confirm the witness will not, and cannot, perform the kind of back-calculation routine in alcohol cases. AAA's own coverage of its findings makes this concession unavoidable.

  • Route of administration. Inhaled, edible, and sublingual products produce very different blood curves. The number on the page does not tell you which curve the driver was on.

  • Tolerance and chronic use. Use the Huestis chronic-user data to show that a frequent or medical user can exceed 5 ng/mL with no acute effect.

  • Parent THC versus metabolites. Force the chemist to break out Delta-9-THC, 11-OH-THC, and Carboxy-THC, and to concede which are inactive.

  • Lab uncertainty of measurement. Every assay has a confidence interval. At 1, 2, or 5 ng/mL cutoffs, that interval often straddles the statutory line.

  • DRE testimony. The 12-step protocol is an investigative tool, not a validated impairment quantifier. Limit it on the record.

  • Authoritative concessions. Walk the state's expert through NHTSA 2017, AAA 2016 and 2024, the 2022 Scientific Reports paper, the 2025 Journal of Clinical Pharmacology model, and the ADLM 2025 statement. Each concession narrows what the prosecution can argue in closing.

A recent practitioner overview from Carey Law Office (September 2025) sets out a similar post-legalization defense roadmap.

8. Where this lands next

Three things to watch in 2026 and beyond. First, oral-fluid roadside devices are spreading, and they raise the same scientific objections as blood with new chain-of-custody and calibration questions on top. Second, expect more states to follow Ohio toward rebuttable-inference frameworks and effects-based statutes, particularly where medical cardholder populations are large. Third, defense attorneys handling current per se cases should preserve every Daubert or Frye objection, every limiting-instruction request, and every metabolite distinction on the record now, because the appellate vehicles for striking down or narrowing these statutes are being built today.

The bright line was always a policy choice dressed as pharmacology. The pharmacology has finally caught up.

Sources

Note: This article contains AI-assisted content and has been reviewed by our editorial team.

IMPORTANT NOTICE: GavelDaily.com is an advertising and information service, not a law firm. Submitting this form does not create an attorney-client relationship. The information you provide will be shared with licensed criminal defense attorneys in your area who may contact you about your inquiry. All submissions are treated as confidential by our service, but attorney-client privilege does not attach until you formally retain an attorney. Any attorney you connect with through this service is independently responsible for their legal advice and representation. Free consultations are subject to individual attorney availability. If you are facing an immediate legal emergency, contact your local public defender's office or call 911.

Share

Reporting Note

Gavel Daily is an AI-operated publication. Articles may summarize statutes, court filings, or public reporting, but readers should verify time-sensitive legal details with primary sources or a licensed attorney.

Related Coverage

DUI/DWI

Case v. Montana Hands Police a New Way Into Your Home Without a Warrant or Probable Cause

On January 14, 2026, a unanimous Supreme Court ruled that police can enter a home without a warrant under the emergency-aid exception based only on an objectively reasonable belief that someone inside is seriously hurt or in danger. Probable cause is not required. Here is what the decision says, why defense lawyers fear it for DUI and welfare-check cases, and how to challenge an entry dressed up as a rescue.

Jun 13

DUI/DWI

Smith v. Arizona Is Killing DUI Blood Tests: Why Surrogate Lab Analyst Testimony Is Getting Suppressed in 2026

The Supreme Court's 2024 ruling in Smith v. Arizona closed a Confrontation Clause loophole that prosecutors had used for over a decade to introduce blood-alcohol results without the analyst who ran the test. Two years on, the fallout is reshaping DUI suppression practice.

Jun 10

DUI/DWI

Lange v. California Is Finally Killing Warrantless Home Entry for DUI: Why Misdemeanor Hot-Pursuit Suppressions Are Surging in 2026

Five years after the Supreme Court rejected categorical hot-pursuit entry for misdemeanors, state high courts are finally applying the rule. Pennsylvania's 2025 Hunte decision shows where DUI suppression motions are winning, and how bodycam timestamps have become the decisive evidence.

Jun 9