CLN - Feature
On the evening of July 11, 2017, Douglas Fraser was pulled over near Everett Mall in Washington state for speeding and erratic driving. According to the state trooper on the scene, Fraser had full body tremors and performed poorly on several field sobriety tests (FSTs), including the finger-to-nose test and accurately estimating time. Fraser was arrested for driving under the influence (DUI). A subsequent test showed he had a blood concentration of 9.4 +/- 2.5 ng/mL of delta-9-tetrahydrocannabinol (THC), the main psychoactive compound in cannabis. Fraser admitted to smoking cannabis but said it was at least 20 hours before the traffic stop, at which point cannabis typically no longer has any effect.
Nevertheless, Fraser was charged under the state’s per se DUI law, known as such because it deems a person intrinsically guilty of driving under the influence if they have a THC blood concentration of 5 ng/mL or higher, regardless of their behavior. Fraser sued up to the Washington Supreme Court, which upheld the per se law, citing research which says that 5 ng/mL of THC “appears to be related to recent cannabis consumption in most people,” which in turn is “linked to” impaired driving.
The State v. Fraser case illustrates the scientific and legal complexities of trying to determine whether a driver has been impaired by cannabis. For alcohol, there’s decades of research showing that a blood alcohol concentration of 0.08 strongly correlates with impaired cognitive and motor functions, as well as a simple tool — the breathalyzer — for measuring it at the roadside. Cannabis, on the other hand, has a more unpredictable effect on behavior and a more complicated metabolism in the body, which makes it difficult to detect the relevant molecules and to know how to interpret a positive test result. (The Washington Supreme Court justices acknowledged this discrepancy between alcohol and cannabis in their ruling.)There’s no dispute that cannabis can and often does slow reaction times and distort perceptions of time and distance; it’s easy to understand why someone who’s high shouldn’t get behind the wheel. And drivers who show obvious signs of impairment can be convicted of a DUI without a blood test or breathalyzer.
But in the absence of clear evidence, the thorny question is this: At what point does someone who’s consumed cannabis become impaired, and can this be tied to a number from a diagnostic test? Even the State v. Fraser verdict acknowledged that the logical progression of 5 ng/mL of THC in blood to recent use, and recent use to impairment, may not always hold true; research conducted since then further suggests that this chain of events is far from certain.
Situations like Fraser’s will likely become more common. According to the National Center for Drug Abuse Statistics, 61.5 million Americans, or nearly 1 in 4, consumed cannabis in 2024. And a survey from the American Automobile Association Foundation for Traffic Safety found that 53% of cannabis users have driven within an hour after consuming the drug.
At the same time, Washington and five other U.S. states have per se laws ranging from a 1 to 5 ng/mL THC limit, while a dozen others have zero-tolerance laws, which say that driving with any amount of THC in one’s system is illegal.
New diagnostic technologies currently in development, and a growing reliance among law enforcement on roadside testing, may help officers better identify drivers who’ve recently consumed cannabis. But can these efforts ultimately make roads safer while avoiding penalizing those who aren’t actually impaired?
Alcohol can intoxicate in a matter of minutes thanks to its water solubility. It is absorbed into the bloodstream through the digestive tract and travels to the brain shortly thereafter. Alcohol’s departure from the body also is predictable: the liver metabolizes ethanol at a rate of about one standard drink per hour and usually fully clears it 24 hours after consumption.
THC, in contrast, is highly lipid soluble, which presents a major challenge to efforts for detecting recent use based on body fluid concentrations. It’s absorbed by fatty tissue, including the brain, and released slowly into the bloodstream. Chronic users, including those who take cannabis for medicinal reasons, can have baseline blood THC concentrations that exceed 2 or even 5 ng/mL long after any impairment has waned.
Additionally, THC has a more complicated mechanism of action. Whereas alcohol is solely a central nervous system stimulant, THC can act as a depressant, stimulant, and hallucinogen; there’s far more variety in how people behave after cannabis use versus alcohol.
All these factors create uncertainty about how cannabis affects driving ability. To systematically address this question, Robert Fitzgerald, PhD, and Thomas Marcotte, PhD, who both have leadership roles at the Center for Medicinal Cannabis Research at the University of California San Diego, conducted a randomized controlled clinical trial. With 191 participants, the trial was the largest of its kind. Participants smoked either a placebo, 5.9% THC, or 13.4% THC cigarette and underwent a driving simulation meant to replicate city and country driving and common traffic challenges, such as freeway merging.
Results from the study were first published in JAMA Psychiatry in 2022. Individuals in the THC arms had significantly worse driving scores compared with those who smoked the placebo cigarettes. But interestingly, there was no relationship between the amount of THC in blood and the scores, and not everyone showed significant impairment.
“If you look at all the evidence, THC is clearly impairing, and particularly so for some individuals,” Marcotte said. “But concentrations in blood and oral fluid, two biofluids that are commonly utilized for per se DUI [determination], have no correlation with driving performance.”
During the trial, Fitzgerald and Marcotte also generated data on how effective FSTs might be in classifying cannabis-impaired drivers and the potential benefit of combining these tests with toxicology results.
When a police officer suspects that a driver has been impaired because of a substance, they’ll typically conduct several FSTs that assess balance, attention, and coordination. Some officers are certified as drug recognition experts (DREs), trained to recognize impairment due to substances other than alcohol, and they conduct a standardized 12-step evaluation, the last step of which involves a review of available toxicology information.
In the trial, 11 DRE-certified officers were asked to perform FSTs on the participants and classify them as impaired or not impaired. (The officers did not perform the full DRE exam.) The FSTs correctly classified 81% of the participants who received THC as being impaired. But 49% of the participants who received the placebo cigarettes were also deemed impaired on the basis of the FSTs. However, when a cutoff of 2 ng/mL THC concentration in oral fluid was imposed on the FST determination, the number of placebo individuals classified as impaired dropped to zero.
“Adding toxicology results may be helpful in increasing the level of suspicion that cannabis was involved in driving impairment, and conversely, in identifying those who have not consumed cannabis recently,” Fitzgerald, Marcotte, and co-authors wrote in a 2023 Clinical Chemistry publication. But these results don't demonstrate causality, they noted.
Blood samples, however, are hardly ever collected at the roadside. Blood draws for drivers suspected of being impaired typically take place at a police station, sometimes hours after a driving incident. In this span of time, blood THC concentrations can plummet, giving a potentially misleading picture of what was in a person’s system when they were driving.
Collecting oral fluid, in contrast, is noninvasive and can read out THC concentrations within minutes. The Dräger DrugTest 5000 analyzer, for example, takes a cheek swab for saliva and relies on an enzyme immunoassay strategy to detect several drug classes, including cannabinoids.
A few U.S. states, including Alabama, Indiana, Michigan, and Minnesota, have deployed statewide pilot programs that equip law enforcement with oral fluid roadside devices. Initial results from these programs show that the devices can perform well in this setting. Based on this, Alabama has since implemented oral fluid testing as part of its standard roadside DUI screening.
In Minnesota’s pilot program, 329 oral fluid tests (either using the Dräger DrugTest 5000 or Abbott’s SoToxa instrument) were obtained after a traffic stop from 268 drivers suspected of using drugs. THC was found 177 times, or in about 54% of samples. Blood or urine samples also were collected from patients, and they found 220 hits for THC, an overlap of 80.5%. Similarly, a study from Norway showed that the proportion of false negatives for cannabis from the Dräger device compared with blood were 13.4%, with a false positive rate of 14.5%.
Although it doesn’t linger in saliva for as long as it does in blood, oral fluid tests may not always yield a person’s systemic THC concentration, because saliva is susceptible to contamination. For instance, if a person smokes a cigarette and gives an oral fluid sample a couple of minutes later, the diagnostic device might detect a high THC concentration, but this likely doesn’t correspond to the concentration in other body tissues.
Breath samples, on the other hand, have been shown to contain traces of consumed drugs and their metabolites while potentially minimizing the issue of contamination, which could lead to consistent results regardless of the mode of cannabis consumption. It’s tempting to imagine law enforcement using a cannabis-detecting breathalyzer alongside the typical one for alcohol, or perhaps packaged into the same device, but technical challenges abound. THC concentrations in breath are several orders of magnitude lower than those of alcohol, and THC is far less volatile.
A team of researchers at the National Institute of Standards and Technology is undertaking the basic research that could one day form the basis for a cannabis breathalyzer used by law enforcement. Using an aerosol device that extracts cannabinoids from breath, the team was able to detect cannabinoids in breath after ingestion of cannabis-infused edibles — an important proof of concept.
Yet alongside these developments is a body of research — including Fitzgerald and Marcotte’s trial and work from the National Highway Traffic Safety Administration — showing a murky connection between THC concentration and driving impairment. Even those who believe in a strong correlation would concede that there’s no scientific basis for choosing 2 or 5 ng/mL as the limit, which per se laws do.
All of which begs the question: If, at least for now, THC concentration can’t be confidently tied to a level of impairment, what is the point of developing new diagnostic methods that might more accurately detect systemic THC levels?
Experts across all sides of the issue agree that toxicology testing should not be the only metric by which a driver is charged with a DUI. In combination with toxicology results, other factors such as driving behavior, officer observations (including FSTs), body camera footage, and witness observations should be considered.
And given that per se laws are on the books in several states, it’s important to refine diagnostic technologies so that test results correspond with actual body levels of THC, said Kara Lynch, PhD, co-director of the core clinical laboratory at San Francisco General Hospital. Additionally, better tests “protect people who use cannabis from being wrongly accused,” she said. “If someone used cannabis two or three days before their shift at work, and they test positive even though they’re not impaired, then they can’t work. There are consequences to society.”
Ultimately, experts are united in calling for more research in several areas, such as how different forms of cannabis affect driving ability and whether biomarkers outside of THC might correspond to driving impairment.
Yaakov Zinberg is a writer based in the Boston area. +LinkedIn: https://www.linkedin.com/in/yaakov-zinberg-276056198/