A 70-year-old taking a CBD capsule and a 30-year-old taking the same capsule are not running the same experiment. The older body has a smaller liver, less blood flowing through it, and a slightly different mix of the enzymes that break drugs down. None of that is visible from the outside, and none of it is on the label. Yet almost everything we know about how much CBD actually ends up circulating in a person’s blood comes from studies in young, healthy volunteers — the group least likely to be buying CBD for arthritis or insomnia.
A team from the University of the Pacific and Washington State University has now tried to close that gap with computer simulation rather than with blood draws. Their paper, Predicting Cannabidiol Pharmacokinetics in Older Adults Via Physiologically Based Pharmacokinetic Modeling, was published in The AAPS Journal on 8 September 2026 by Lixuan Qian, Aiden-Hung P. Nguyen, Mary F. Paine and Zhu Zhou. It combines fresh laboratory work on human liver tissue with a virtual population of older adults, and it produces the first published prediction of how CBD exposure shifts with age.
Key Takeaways
The study reports that cytochrome P450 enzymes handle about 86% of CBD’s breakdown and UGT enzymes about 14% — a meaningful revision of the 70/30 split the same group had published earlier, and one that came from removing two common laboratory additives that turned out to be inflating the UGT share. CBD did slow down one enzyme, CYP1A2, in a lasting way, but since CBD is not itself processed by CYP1A2, this had no practical effect on its own clearance. When the corrected model was extrapolated to a virtual population aged 65 to 98, simulated CBD exposure rose relative to young adults in almost every scenario, with the oldest group showing the highest levels. The size of the increase depended heavily on how CBD was delivered: an oromucosal spray produced exposures 1.20 to 1.82 times those in young adults, while an oil-based oral solution absorbed partly through the lymphatic system showed a much smaller age effect. Variability between individuals also widened with age. Crucially, none of this was measured in actual older people — there are no published CBD pharmacokinetic trials in healthy older adults at all, which is precisely why the modelling was done.
Why Nobody Has Simply Measured This
The obvious way to find out how older bodies handle CBD would be to give it to older volunteers and take blood samples over many hours. That study does not exist. Proper pharmacokinetic work requires repeated blood sampling, and in older participants that raises safety and tolerability concerns, so trials that do include them tend to skip the full sampling schedule. The result is that the two numbers clinicians care about most — total exposure over time, and peak concentration — are usually missing.
This matters because CBD use in this age group is not marginal. A 2022 survey found that 14.3% of U.S. adults aged 65 and over had used CBD, typically for pain, low mood or sleep problems. Those are chronic complaints, which means repeated dosing over months or years, often alongside several prescription medicines. The broader evidence base for CBD is thin enough in younger adults; in older ones it is close to absent.
Physiologically based pharmacokinetic modelling — PBPK — is the workaround. Instead of measuring a drug in people, you build a mathematical body: organ sizes, blood flows, enzyme quantities, all drawn from published anatomical and physiological data, and then you release a virtual dose into it and watch where it goes. The approach is routine in drug development and is accepted by regulators for exactly this kind of question, where the population of interest is hard to study directly.
What CBD Does to the Liver’s Processing Machinery
Before the model could be trusted, the authors had to settle an open question about CBD’s own behaviour in the liver. CBD had previously been reported to cause time-dependent inhibition of several CYP enzymes — a form of interference where the drug does not simply block the enzyme but progressively disables it, so the effect deepens the longer the exposure lasts. If CBD were disabling the very enzymes that clear it, it would effectively be slowing its own elimination.
Testing all five major CYPs in human liver microsomes, the team found time-dependent inhibition of CYP1A2 only — not of CYP2C19 or CYP3A, as an earlier study had suggested. They attribute the discrepancy to different probe substrates being used in the two experiments. Reversible inhibition, the more familiar kind, was moderate to weak across the board: at a CBD concentration of 0.5 µM, remaining enzyme activity was 53% for CYP2C19, 75% for CYP3A, 77% for CYP2C9, 82% for CYP1A2 and 92.2% for CYP2D6.
The practical conclusion is reassuring in one direction and unhelpful in another. Because CBD is not a CYP1A2 substrate, its inhibition of that enzyme does not feed back into its own clearance — building it into the model changed nothing. But the reversible inhibition of CYP2C19 and CYP3A is still relevant to anyone taking CBD alongside other medicines, and it is a reminder that CBD is pharmacologically active well beyond the effects people buy it for. Interactions run in both directions, as work on the combined effects of CBD and THC has also shown.
The Laboratory Additives That Were Skewing the Numbers
The most consequential finding is technical and easy to skip past, so it is worth stating plainly: the standard way of running these assays had been producing the wrong answer.
UGT enzymes sit inside a membrane pocket in liver cells, with their active site facing inward, away from the cofactor they need. In a test tube, that barrier slows them down artificially. To compensate, labs routinely add alamethicin, a peptide that punches pores in the membrane, plus magnesium chloride. Both boost measured UGT activity. Neither exists in a living liver.
When the team ran the assay both ways, the additives changed the picture substantially. With alamethicin and MgCl₂ present, intrinsic clearance was 1.07 mL/min/mg via CYPs, 0.55 via UGTs and 2.19 with both cofactors together. Without them, those values were 18%, 73% and 51% lower respectively. In other words, removing the additives barely touched the CYP figure but cut the apparent UGT contribution by nearly three-quarters.
Which version reflects reality? The authors answered empirically, by running each version through the PBPK model and checking it against five real drug-interaction trials in which CBD was given with ketoconazole, omeprazole, rifampicin, stiripentol or valproate. The condition without additives won. For the three trials where the interaction shifted exposure by more than 25%, predicted-to-observed ratios were 0.71, 1.28 and 0.94 — all within 30% of observed values. Independent verification against nine further clinical trials covering 29 dosing regimens stayed inside the conventional two-fold accuracy criterion.
That is how the 70/30 split became 86/14. It is not a cosmetic correction: in older adults, where liver size falls by as much as 35% and hepatic blood flow by as much as 40% relative to young adults, an inflated UGT share would mask genuine age-related changes in clearance.
Borrowing THC to Check the Method
Extrapolating a model into a population you have never measured is an act of faith unless you can validate the extrapolation itself. With no CBD data in healthy older adults, the team used THC as a stand-in — a cannabinoid with similar physicochemical properties, and one that has been studied in older volunteers.
They took their published THC model, swapped in a virtual geriatric population, changed nothing else, and compared predictions against a clinical trial that had given 3 mg, 5 mg and 6.5 mg of oral THC to healthy adults aged 65 to 80. Predicted exposures for both THC and its main metabolite fell within two-fold of the observed values across all three doses, though the model slightly underpredicted — plausibly because participants were taking an average of two other medications. The extrapolation method held up, which is what licensed the same move for CBD.
What the Simulations Predicted for People Over 65
The virtual older population spanned ages 65 to 98 and was split into three bands: youngest-old (65–74), middle-old (75–84) and oldest-old (85–98), each compared against young adults aged 18 to 64. Simulations covered an oral sesame-oil solution at 200, 750, 1,500 and 6,000 mg, and an oromucosal spray at 5, 10 and 20 mg.
Exposure rose with age in nearly every scenario, and the oldest-old group consistently showed the highest values. For the 750 mg oral dose — the most common dose in clinical trials — predicted exposure in the oldest-old was 1.14 times that of the youngest-old and 1.28 times that of young adults. Modelled as a non-oil formulation, the same doses produced exposure increases above 50% in the middle- and oldest-old groups and above 25% in the youngest-old.
The oromucosal spray behaved differently again, with every older group landing 1.20 to 1.82 times above young adults. At the 10 mg spray dose, the oldest-old reached 1.82 times young-adult exposure.
Why the Delivery Route Mattered More Than the Dose
The split between formulations is the most practically interesting part of the paper. CBD is intensely fat-soluble, and when taken in oil it is partly absorbed through the lymphatic system rather than straight into the portal circulation. Lymphatic uptake bypasses some first-pass metabolism — and it appears to blunt the effect of ageing. A sensitivity analysis showed that as the lymphatic fraction was reduced, the gap between young and old exposure widened; with lymphatic absorption switched off entirely, the oil formulation behaved much like a non-oil one.
Put simply, the same milligram number can mean different things depending on the carrier and the route. That argues against any single CBD dosing rule for older adults, and it sits awkwardly alongside a retail market where formulation is inconsistent and label accuracy is not guaranteed — as testing of European CBD oils has repeatedly demonstrated.
Variability also widened with age. Mean values in a simulation describe a population, not a person, and the spread among older virtual subjects was consistently larger than among young ones — a pattern matching a real population study in older patients with poor appetite.
What This Does and Does Not Tell a 70-Year-Old Taking CBD
These are predictions, not measurements, and the authors are explicit about that. The extrapolation to older adults was never verified against clinical CBD data, because no such data exist. The liver-tissue experiments were run under a narrow set of conditions and need replication in human hepatocytes. Distribution was modelled using partition coefficients, which limits the model’s ability to capture age-related changes in how a highly fat-soluble compound spreads through tissue. And higher exposure is not automatically a problem — it becomes one only if it translates into more of the adverse effects associated with CBD accumulation, such as liver enzyme elevations, drowsiness and hair loss. Whether it does is a pharmacodynamic question this study cannot answer.
What the work does support is a shift in default assumption. If a dose was characterised in 25-year-olds, it should not be assumed to deliver the same internal exposure to an 85-year-old, especially via a spray or a non-oil oral product. In the United States, hemp-derived CBD containing no more than 0.3% THC became federally lawful under the Agriculture Improvement Act of 2018, but the FDA has not approved CBD as a dietary supplement or food ingredient, and the only approved CBD medicine remains a prescription treatment for rare seizure disorders. Consumer products therefore reach older users without the exposure data that would accompany any licensed drug. A review of cannabinoid pharmacokinetics in special populations by the same group makes the same point about other under-studied groups, and similar evidence gaps recur elsewhere — including in what is known about CBD and bone.
FAQ
In these simulations, yes — total exposure was higher in every older age band than in young adults for most formulations, and highest in the 85–98 group. The increase ranged from modest to roughly 80% depending on dose and delivery route. This is a modelled prediction, not a measured result.
Because no published pharmacokinetic study has tracked CBD in healthy older adults, and THC has been studied in that group. The two cannabinoids have similar physicochemical properties, so the authors first showed that their THC model extrapolated accurately to older adults, then applied the same extrapolation method to CBD.
It can. CBD showed moderate to weak reversible inhibition of all five major CYP enzymes tested, with the strongest effect on CYP2C19. Because older adults commonly take multiple prescription medicines, and because both liver size and blood flow decline with age, this is the population where such interactions matter most. Anyone taking CBD alongside prescribed medication should raise it with the prescribing clinician.
Hemp-derived CBD with no more than 0.3% THC is federally lawful following the 2018 Farm Bill, but legality of sale is not the same as regulatory approval. The FDA has concluded that CBD is excluded from the dietary supplement definition under the Federal Food, Drug, and Cosmetic Act, and has not authorised its addition to foods or supplements. State rules vary considerably, and the only FDA-approved CBD product is a prescription medicine for specific seizure disorders.
Legal Disclaimer
This article is journalistic coverage of published scientific research and is provided for informational purposes only. It is not medical advice, a treatment recommendation, or a dosing guide, and it should not be used to start, stop or adjust any medication or supplement. The findings described are computer simulations, not clinical measurements in patients. CBD can interact with prescription medicines, and the regulatory status of CBD products differs by country and, within the United States, by state. Readers with questions about their own health, medication or the legality of a product where they live should consult a qualified healthcare professional or legal adviser.
