What Actually Reaches the Bloodstream: Measured Cannabinoid Exposure From Inhaled, Oral, and Topical Products

One study administered smoked, vaporized, and oral cannabis to the same participants and tracked whole blood cannabinoids for up to 72 hours. Smoked and vaporized routes produced few pharmacokinetic differences from each other. Oral dosing produced significantly greater concentrations of the inactive metabolite THC-COOH and its glucuronide, while minor cannabinoids that mark recent inhalation were absent entirely.

For topical products, the answer is narrower than either the marketing or the skepticism suggests. In 46 healthy adults using commercial hemp-derived topicals, cannabidiol was measurably absorbed from three of five products, while delta-9-THC and its metabolites stayed below the limit of detection in blood for every product tested.

Two practical questions hang on these numbers. A patient choosing a topical wants to know whether anything is reaching their bloodstream, for both efficacy and testing reasons. A patient or clinician looking at a blood THC result wants to know what it means, and the honest answer is considerably less than most people assume.

The gap between those two questions and the confident figures that circulate online is wide. Numbers get quoted for routes nobody has measured, and measured numbers get quoted as though they establish impairment, which they do not.

The comparison worth anchoring to is a controlled crossover study by Newmeyer and colleagues published in Clinical Chemistry in 2016. Frequent and occasional cannabis users received smoked, vaporized, and oral cannabis, and whole blood was sampled for 54 hours in occasional users and 72 hours in frequent users, with THC, its phase I and phase II metabolites, and several minor cannabinoids measured by mass spectrometry.

Three findings carry forward. First, smoking and vaporization produced few differences in blood cannabinoid pharmacokinetics, which the authors summarized as comparable cannabinoid delivery. Second, oral cannabis produced significantly greater THC-COOH and THC-COOH-glucuronide, the metabolite signature of a dose that passed through the liver before reaching circulation. Third, cannabigerol and cannabinol appeared frequently after inhalation with short detection windows and were not detected at all after oral dosing.

That third point is useful in a way that rarely gets stated. The route a person used leaves a chemical fingerprint in blood that is independent of dose. A blood panel showing cannabigerol and cannabinol points toward recent inhalation; their absence does not rule it out.

A companion study by Spindle and colleagues in the Journal of Analytical Toxicology compared smoked and vaporized cannabis at 0, 10, and 25 mg THC in adults with no past-month use. Blood concentrations were dose-dependent for both methods and higher after vaporization, with peak values typically at the first sampling point ten minutes after administration.

Until recently the topical question had no controlled human answer, which is why route tables that assign onset windows to topicals are reporting numbers the literature does not contain. That has changed for one specific question: systemic absorption.

Zamarripa and colleagues, in the Journal of Analytical Toxicology in 2024, gave 46 healthy adults one of five commercially available hemp-derived high-CBD topicals or a matched placebo: a cream, a lotion, a patch, a balm, or a gel. The protocol ran 17 days, with an acute laboratory session, nine days of twice-daily outpatient application, and a one-week washout. Blood, oral fluid, and urine were analyzed by mass spectrometry.

Transdermal cannabidiol absorption was observed for three of the five active products. Concentrations generally peaked after seven to ten days of use rather than acutely, and were highest for the lotion, which contained the most cannabidiol and included vitamin E as a permeation enhancer. Delta-9-THC and its metabolites remained below the limit of detection in blood for every product, and no urine specimen met United States federal workplace criteria for a positive cannabis result.

One unexpected result deserves attention. Nine participants, seven using the lotion and one each using the patch and gel, showed oral fluid THC concentrations at or above 2 ng/mL, which is the current federal workplace threshold for a positive oral fluid test. The likely explanation is transfer from hands to mouth rather than systemic absorption, and the practical implication is that a topical user facing oral fluid testing has a plausible route to an unexpected result even with nothing detectable in their blood.

This is where confident numbers fall apart fastest. In a randomized trial of 191 frequent and occasional users reported in the Journal of Analytical Toxicology in 2021, Hubbard and colleagues tested whether cannabinoid concentrations in blood, oral fluid, or breath could identify use within the previous three hours. Blood per se limits ranging from undetectable to 5 ng/mL offered limited usefulness as biomarkers of recent use, and frequent users frequently carried enough residual blood THC to be classified as recent users before they smoked anything.

Cannabinol in blood at a 1 ng/mL cutoff gave perfect specificity but only 31.4 percent sensitivity in that study, and its concentration varies with the cannabis chemovar. A 10 ng/mL oral fluid THC cutoff performed best overall for detecting use within three hours, with 99.7 percent specificity and 82.4 percent sensitivity, and it remained positive in 23.2 percent of participants roughly four and a half hours after smoking. The authors were explicit that detecting recent use does not equate to detecting impairment.

A pharmacometric modeling study published in the Journal of Clinical Pharmacology in 2025 reached a compatible conclusion for oral users specifically. Using data from ten published studies, the authors simulated plasma THC across doses from 2.5 to 100 mg and found the 1 ng/mL per se limit least effective because of false positives, while the 2 and 5 ng/mL limits remained inconclusive for lack of pharmacodynamic data linking blood levels to impairment.

The most instructive result on this point comes from the same NIDA dataset as the three-route comparison. Newmeyer and colleagues reported in the Journal of Applied Toxicology that occasional users given oral cannabis had roughly 6.4 times the odds of showing two or more impairment clues on standardized field sobriety tasks compared with placebo, while inhaled doses showed no significant effect at the times tested. Oral dosing produces lower blood THC than inhalation and, in that study, more measurable impairment.

Inhaled routes put THC into blood within minutes and produce the highest early concentrations, with vaporization and smoking behaving similarly enough that they can be treated as one category for pharmacokinetic purposes. The tradeoff is a steep decline, which is why blood levels after inhalation correlate poorly with how someone feels an hour later.

Oral routes produce lower peak THC, a much larger share of 11-hydroxy-THC and THC-COOH, a delayed peak, and a longer tail. The consequence that matters clinically is that a low blood THC after an edible does not mean a small exposure, because a substantial part of the dose is circulating as metabolites the test may not be reporting.

Topical products, on the current evidence, deliver measurable cannabidiol into blood with repeated use and no detectable delta-9-THC. Whether they deliver enough cannabinoid to the tissue underneath to produce a local effect is a separate question that these pharmacokinetic studies were not designed to answer, and the participants in the topical study showed no discernible subjective, cognitive, or physiological effects.

What none of this supports is a table with a specific onset time for topicals, a specific blood level for a given dose, or a threshold that separates impaired from unimpaired. Those numbers get repeated because the format demands a value in every cell, not because anyone measured them.

For the pharmacokinetic questions, this is about as good as the field gets. Controlled administration under blinded conditions, the same participants across routes, mass spectrometry rather than immunoassay, and sampling windows long enough to characterize elimination. The route comparisons and the topical absorption findings can be quoted with confidence.

For the interpretive question, the evidence is strong in a negative direction. Multiple independent groups using different designs have converged on the conclusion that blood THC concentration is a poor stand-alone indicator of recent use or impairment. That is a well-supported negative finding rather than an absence of research.

The topical study tested five specific commercial products in a market containing thousands. Absorption depended heavily on cannabidiol content and on whether a permeation enhancer was present, which means the results describe those formulations rather than topicals as a category. The authors said as much.

The oral fluid finding in that study has a mundane likely explanation, hand-to-mouth transfer, that the design could not exclude. It is an important practical observation and a weak mechanistic one, and the difference matters if someone wants to cite it as evidence of transdermal THC delivery, which it is not.

None of this shows that topical cannabis products work. Measuring cannabidiol in blood after nine days of lotion use says something about absorption and nothing about whether the product relieved anything; the participants in that study showed no discernible subjective, cognitive, or physiological effects, and they were healthy volunteers without a condition to treat.

It also does not establish onset or duration for topical products. No controlled human study has measured when a topical cannabinoid begins to act locally or how long that action lasts, so published onset windows for topicals remain unsupported. And nothing here identifies a blood concentration that separates impaired from unimpaired drivers.

Cannabis testing policy has borrowed its structure from alcohol, where blood concentration tracks impairment closely enough to support a legal threshold. The cannabinoid data do not support that borrowing. Frequent users carry residual blood THC for days, occasional users show more measurable impairment at lower blood levels after edibles, and the compounds distribute into fat rather than staying in water the way ethanol does.

For clinical practice the useful reframing is that route determines what is circulating, not just how much. A patient who took an edible and a patient who vaporized may show similar blood THC with entirely different metabolite profiles, different effect trajectories, and different time courses ahead of them.

Patients ask me whether a topical will show up on a drug test more often than they ask whether it will help their knee. The controlled answer is reassuring for blood and urine and less reassuring for oral fluid, and the oral fluid result probably comes down to washing your hands. That is an unsatisfying thing to tell someone, and it is what the data say.

On blood levels I try to move the conversation somewhere more useful. A number from a blood draw tells you a compound is present. It does not tell you when it was taken, how impaired someone was, or whether the dose was large, and for a patient who uses cannabis regularly it may tell you only that they used cannabis at some point recently. Treating that number as a measurement of impairment is a category error, and people have lost jobs and licenses over it.

Inhaled routes deliver the highest early blood THC and vaporizing and smoking behave nearly identically. Oral routes deliver lower blood THC with a much larger metabolite load and, in controlled testing, more measurable impairment. Commercial topicals delivered detectable cannabidiol into blood after a week of use and no detectable THC, though several users still produced positive oral fluid results. No blood THC threshold reliably identifies recent use or impairment.

Carry forward the route-specific metabolite signatures, the topical absorption result and its oral fluid caveat, and the repeated finding that blood THC is a poor stand-alone measure. Do not carry forward any specific onset time for topicals or any blood concentration presented as an impairment threshold; neither exists in the literature.