THC potency and driving: what a 2026 controlled study found.

AI-assisted literature summary of Brands and colleagues (2026), selected from the supplied cannabinoid research collection. No personal review or adoption of this article by Okorie Okorocha is represented. Editorial policy.

The short answer: this trial supports a relationship between smoked cannabis, product potency, and several measures of impaired simulated driving under its study conditions. It does not establish that every measured THC concentration proves impairment. It also does not support the opposite claim that blood THC can never correlate with driving performance. Both overstatements discard part of the evidence.

What the researchers actually tested

Brands and colleagues conducted a randomized, double-blind, placebo-controlled crossover trial in Toronto. Participants were eligible if they were 19 to 45 years old, used cannabis one to five days a week, and met the driving and health criteria. People with cannabis use disorder and several potentially confounding conditions or medications were excluded.

The paper reports 54 participants randomized, 49 receiving at least one drug condition, and 35 completing the per-protocol analysis. The completers were 19 to 34 years old. This is not a representative sample of every driver, every medical-cannabis patient, or every pattern of cannabis use.

Each participant was assigned sessions involving placebo and nominal THC potencies of 6.25%, 12.5%, and 22%. Driving assessments began 30 and 90 minutes after smoking. The researchers measured speed, lane-position variability, reaction time, blood cannabinoids, and participants’ perceptions of their driving. The original paper supplies the protocol and statistical details.

Do not confuse potency, dose, and blood concentration

Potency is the THC proportion in the cannabis product. The estimated amount consumed in this study was calculated from the cigarette’s change in weight multiplied by its potency. That estimate is not a direct measurement of the amount absorbed into the bloodstream.

The procedure also changed during the trial. Initially, participants followed a paced smoking schedule. Because that procedure was poorly tolerated, later sessions allowed participants to smoke to their usual desired effect. The paper reports that participants generally smoked less as potency increased, but the estimated THC amount consumed still increased. Smoking less material did not make exposure equivalent across the conditions.

For a separate discussion of administration route, see smoked versus vaporized cannabis. Do not transfer this smoked-cannabis trial directly to edibles or concentrates.

The primary outcome was negative; several secondary outcomes were not

The prespecified primary outcome was mean driving speed. The trial did not detect a significant potency effect on that outcome. That finding belongs in any fair account of the paper.

Several secondary measures told a different story. Lane-position variability increased in all active-cannabis conditions compared with placebo. The highest-potency condition produced more lane-position variability than the lower two active conditions. Reaction time was longer in the medium- and high-potency conditions than with placebo. Maximum speed also increased in those two conditions relative to placebo.

The correct reading is not “all driving measures worsened,” and it is not “the trial found no impairment.” Name the outcome. State the comparison. Keep the distinction between the primary and secondary analyses visible.

The blood-THC associations must be reported, not explained away

The researchers found positive associations between whole-blood THC and both lane-position variability and reaction time in their statistical models. These findings are relevant evidence. A categorical claim that blood THC never correlates with performance would be inconsistent with this study.

But a study-level association and an individual classification rule are different products. The paper did not establish a universal blood-THC cutoff with validated error rates for deciding whether a particular driver was impaired at an earlier event. Its regression relationships should not be silently converted into such a rule.

That boundary is an interpretation of the study’s design and reported analyses, not a claim that its positive findings disappear. To use the study in a particular evaluation, compare the person’s specimen, collection timing, route, use pattern, and observed performance with the conditions the researchers actually studied. The companion guide on specimen-matrix differences explains why naming the specimen comes first.

Limitations that change how the findings should be used

• Selected population and attrition: 35 per-protocol completers are not the same as 54 randomized participants. The paper also reports additional intention-to-treat analyses, but the supplementary tables were not independently reviewed for this brief.
• Multiple comparisons: the authors report unadjusted pairwise p-values. Secondary comparisons should not be presented as if multiplicity had been eliminated.
• Changed smoking procedure: paced and self-directed smoking were both used. Statistical adjustment does not make the protocols identical.
• Product preparation: the preparation of the lower-potency cigarettes changed during the study, and individual cigarettes were not each checked for THC content.
• Two laboratories: blood specimens were analyzed using GC-MS and LC-MS/MS methods at different laboratories. The authors identify that methodological difference as a limitation.
• Limited observation period: simulator testing began at 30 and 90 minutes. This study does not determine when driving performance returned to baseline after that period.
• Simulator, not a crash study: these performance measures are not direct measurements of real-world collision probability or an individual’s legal responsibility.

Questions to put beside a THC report

1. Does “high THC” refer to the product’s percentage, an estimated consumed amount, or a measured concentration in a named specimen?
2. What were the use time, driving or event time, and actual specimen-collection time?
3. Is the person reasonably comparable to the study population, or are frequency of use, age, medications, and route materially different?
4. Which driving outcome is being cited, and was it a primary outcome, secondary outcome, or exploratory association?
5. Is the opinion explaining a group association, predicting an individual result, or applying a legal rule? What evidence supports that extra step?
6. Does the opinion preserve both the positive findings and the study’s limitations?

Continue the reading path

Use the cannabis and THC evidence hub for specimen, method, and interpretation questions. For case organization, see four separate questions in cannabis DUI evidence. For a distinct analytical issue, the CBD-to-THC derivatization article concerns a particular sample-preparation method, not driving performance.

Source and review scope

Brands B, Zaweel A, Wright M, et al. Potency-related effects of smoked cannabis on simulated driving performance: a randomized, controlled crossover trial. Scientific Reports. 2026;16:12961. doi:10.1038/s41598-026-43045-2. PubMed record.

The main paper’s methods, results, limitations, tables, and figures were consulted. Figure 1 separates the driving outcomes; Figure 3 displays the reported concentration-performance relationships. The supplementary files were not independently reviewed. This is a focused appraisal of one trial, not a systematic review of the entire supplied collection. No source PDF, figure, or table is reproduced here.

Educational research material, not legal or medical advice, a case-specific opinion, or a recommendation about when it is safe to drive. A 90-minute observation point is not a safe-driving waiting period.