How Long Does Cannabis Impairment Last? The Question the FAA Just Asked the National Academies to Answer

According to PubMed, the largest meta-analytic review of this question found that most driving-related cognitive skills recover within about five hours of inhaling 20 milligrams of THC, and nearly all within about seven hours, with oral routes taking longer. The authors recommended waiting at least five hours after inhaled cannabis before performing safety-sensitive tasks.

Blood and urine testing measures a different thing entirely. THC remains detectable long after functional recovery, particularly in frequent users, which is why a negative or positive test says almost nothing about whether a person is impaired right now.

Federal rescheduling changed the legal classification of certain marijuana products without changing Department of Transportation testing rules or FAA medical certification standards. That mismatch leaves patients in safety-sensitive occupations in a genuinely difficult position, and it is the reason the FAA asked for expert input in the first place.

For a clinician, the practical question is narrower and answerable. A patient wants to know how long to wait. The controlled literature supports a specific, conservative interval, and it also explains why the interval is not the same for an edible as for a vaporizer, or for a daily user as for an occasional one.

The National Academies of Sciences, Engineering, and Medicine project page describes a four-part expert meeting series titled Meeting of Experts: Discussion of Cannabis Use and Operational Fitness Policy for Pilots and Air Traffic Controllers. The sponsor is the Department of Transportation. The stated purpose is to enhance the FAA’s understanding of the impact of both recent and cumulative cannabis use and to inform the agency on approaches it might take to evaluate, update, or refine its current policies. The stated scope includes initial and ongoing screening, monitoring for cognitive functioning, and duration and variability of impairment.

Three things on that page deserve emphasis because they are frequently misreported. The meetings are scheduled across 2026 and 2027. They are not open to the public. No product will be generated, meaning there will be no consensus report of the kind the National Academies is usually known for. The project is administered through the Center for Health, People, and Places and is listed as in progress.

This is a convening, not a finding. It does not change any rule. Department of Transportation drug testing requirements and FAA medical certification standards for cannabis stand as they were. The federal rescheduling action published in the Federal Register on April 28, 2026 moved certain FDA-approved and state-licensed marijuana products from Schedule I to Schedule III, and did not alter workplace testing obligations for safety-sensitive transportation employees.

According to PubMed, McCartney, Arkell, Irwin, and McGregor at the University of Sydney published a systematic and meta-analytic review in Neuroscience and Biobehavioral Reviews in 2021 covering 80 publications and 1,534 outcomes, specifically examining the magnitude and duration of acute THC-induced driving and cognitive impairment.

Several domains showed impairment at peak THC effect, including lateral control, tracking, and divided attention. Meta-regression found that regular cannabis users experienced less impairment than occasional users, and that the magnitude of impairment depended on dose, time since administration, and which skill was being assessed. The model predicted that most driving-related cognitive skills would recover within about five hours of inhaling 20 milligrams of THC, and nearly all within about seven hours, with oral THC taking longer to subside. The authors’ stated recommendation was to wait at least five hours after inhaled cannabis before performing safety-sensitive tasks.

A separate line of evidence converges on a similar window in patients rather than volunteers. According to PubMed, Eadie and colleagues published a scoping review in Frontiers in Psychiatry in 2021 restricted to medical cannabis patients using cannabis for chronic non-cancer pain or spasticity. Seven controlled trials met inclusion criteria. Cognitive performance declined largely in a dose-dependent way and resolved steadily over the hours following administration, and in every study there was no difference between any THC group and placebo on any neurocognitive measure after four hours of recovery.

Oral cannabis behaves differently, and a recent simulator study makes the point concretely. According to PubMed, Won and colleagues published a study in Traffic Injury Prevention in 2025 in which 88 adults in Denver aged 25 to 55 were grouped by past 90-day use as daily, occasional, or no recent use, and tested on a validated driving simulator at baseline and at 52 and 119 minutes after consuming edible cannabis.

The occasional-use group showed the clearest impairment. Their lane departures rose from 0.17 per minute at baseline to 0.47 at the first post-test and 0.40 at the second, and their standard deviation of lateral position in the rural driving scenario increased from 25.45 centimeters to roughly 29.4 and 29.8 centimeters. The daily-use group drove more slowly than both other groups and showed significantly smaller increases in lane departures. The authors concluded that tolerance contributes to outcomes beyond THC concentration alone.

That result is counterintuitive and clinically important. The person least accustomed to cannabis is the one most impaired by a given exposure, which means the safest interval is not a single number applied to everyone. It also means that at two hours after an edible, impairment in an occasional user is still measurable, well past the point most people assume the effect has faded.

The FAA’s problem is that no validated biological test identifies the moment a person stops being impaired. The pharmacokinetics and the pharmacodynamics come apart almost immediately after use, and the gap widens with frequency of use.

According to PubMed, Hoffman and colleagues characterized ten cannabinoids in blood and oral fluid from 191 frequent and occasional cannabis smokers for up to six hours after ad libitum smoking, published in the Journal of Analytical Toxicology in 2021. Two findings matter here. THC was detectable for significantly longer in both blood and oral fluid from frequent users than from occasional users, which is the inverse of the impairment pattern. And peak blood THC concentrations were higher in subjects who smoked the 5.9 percent THC plant material than the 13.4 percent material, because participants self-titrated. A higher-potency product did not produce a higher blood level.

Timing corrupts the measurement further. According to PubMed, Price and colleagues at the National Transportation Safety Board examined 8,923 Wisconsin drivers involved in crashes and arrested for impaired driving between 2019 and 2021, reporting in Traffic Injury Prevention in 2024 that mean time from violation to blood collection was 1.80 hours, rising to 2.35 hours for fatal crashes, and that blood THC concentration was significantly negatively correlated with that delay. The concentration measured is not the concentration that was present while driving.

The literature supports a defensible clinical posture even though it does not support a test. For inhaled cannabis, a minimum of five hours before any safety-sensitive task is the interval the meta-analysis recommends, and seven hours is the more conservative reading of the same model. For oral cannabis, the interval is longer and less predictable, and the simulator data show measurable impairment at two hours in occasional users.

Individual factors widen the window. A patient who uses cannabis rarely should assume a longer interval than a daily patient, not a shorter one. Higher doses extend the window. Combining cannabis with alcohol or sedating medications changes the calculation entirely and is not addressed by any of this work.

None of that resolves the regulatory question, and clinicians should be explicit about the separation. A patient in an FAA-regulated or Department of Transportation-regulated role faces a testing requirement that operates independently of whether they are impaired. Waiting long enough to be unimpaired does not make a urine test negative, and counseling has to address both facts rather than blurring them together.

For the duration question, this is reasonably strong evidence. A meta-analysis of 80 publications and more than 1,500 outcomes, with meta-regression modeling dose, time, route, and user frequency, is about as good as the controlled human literature on this topic gets, and an independent scoping review restricted to medical cannabis patients arrives at a compatible four-hour figure.

For the policy question the FAA is asking, the evidence is considerably thinner. No study has tested pilots or air traffic controllers, no study has measured performance on aviation-relevant tasks after cannabis, and the entire duration literature rests on driving and laboratory cognitive measures. Extrapolating from a driving simulator to a flight deck is an assumption, not a finding.

The five-hour figure is a model prediction under specified conditions, principally 20 milligrams of inhaled THC in participants described as occasional users, with recovery defined by an effect size threshold. Real-world doses vary enormously, concentrates deliver far more than 20 milligrams, and a threshold-based definition of recovery is a convention rather than a physiological boundary.

The medical patient scoping review included only seven controlled trials, used THC doses lower than typical recreational studies, and reported substantial variation in which cognitive tests were administered and when. The simulator study followed participants only to 119 minutes, so it cannot say when the occasional-use group returned to baseline. None of these studies measured residual next-day effects, which is a distinct question with its own contested literature.

This work does not establish a blood, breath, or oral fluid concentration that corresponds to impairment, and the detection data argue strongly that no such threshold will be found. It does not show that a person who waits five hours is safe in every circumstance or at every dose.

It also does not address the questions specific to aviation: cumulative effects of regular use on the cognitive skills a pilot or controller needs, performance under hypoxia or fatigue, or interaction with the shift patterns worked by air traffic control specialists. Those are the questions the expert meetings exist to scope, and the literature currently has little to offer on any of them.

Cannabis policy has largely borrowed its enforcement architecture from alcohol, where a blood concentration maps onto impairment closely enough to legislate around. Cannabis does not behave that way. Tolerance moves in the opposite direction from detectability, self-titration breaks the relationship between product potency and delivered dose, and elimination kinetics stretch detection far past the point of functional recovery.

That mismatch is why agencies keep commissioning reviews rather than issuing thresholds. The honest summary is that the science supports time-based guidance and does not support concentration-based testing as a measure of fitness, and that most existing policy does the opposite.

The question I get is almost never about policy. It is a nurse on nights, or a commercial driver, or someone who just started flight training, asking how long they need to wait. They are asking in good faith and they deserve a number rather than a shrug.

The number I give is conservative and I say why. At least eight hours after inhaled cannabis, longer if the dose was substantial or the product was a concentrate, and considerably longer after an edible, because oral onset is slow and the tail is unpredictable. That sits deliberately beyond the five to seven hours the meta-analysis supports, because the studies used defined doses in laboratory conditions and my patient is using a product of uncertain potency at home.

Then I say the part people do not want to hear. If you hold a Department of Transportation-regulated or FAA-regulated safety-sensitive position, being unimpaired is not the standard you are held to. The standard is a negative drug test, and those two things can be weeks apart in a regular user. I will not pretend otherwise, and I do not think a physician should help anyone plan around a federal testing requirement.

Controlled human research supports waiting at least five hours after inhaled cannabis before a safety-sensitive task, with seven hours a more conservative reading and oral routes requiring longer. Occasional users are more impaired than daily users at the same exposure, and blood or urine concentration does not indicate impairment in either group. Anyone in a federally regulated safety-sensitive role should understand that testing rules operate independently of impairment and have not changed.

Separate three questions that get merged constantly. Whether a person is currently impaired is answerable by time since use, dose, and route. Whether a person will test positive is answerable by frequency of use and elimination kinetics, and the answer is often yes long after impairment ends. Whether a rule permits them to work is a legal question that tracks the second answer rather than the first.