Cannabis for Dementia and Alzheimer’s: What THC, CBD, CBG and CBN Can and Can’t Do

By Blazin Bill • September 27, 2026 • 18 min read • Sources checked against the primary studies; a second, independent review pass by an outside model

Updated the same day it went up: added an Alzheimer’s-specific section, the Avidekel oil protocol exactly as the trial ran it, and the doses CBG and CBN have actually been studied at.

This one was requested. A budtender at RISE Cleveland, at Brookpark and Pearl, had found the site, and the question they get most from families at the counter isn’t about terpenes or vaporizers — it’s some version of “my mother has dementia, will this help?” They asked for something honest they could point people to. Here it is: every claim below cites the actual trial, the numbers are the trial’s numbers, and where the answer is “we don’t know,” it says so. Thank you for asking — this page exists because you did.

The honest hierarchy

Before the details, the shape of the evidence. “Certainty” in the right-hand column is the language the Cochrane review and the trials themselves use.

Alzheimer’s specifically: which trials were which, and what the lab work does and doesn’t show

“Dementia” is the umbrella; Alzheimer’s is one disease under it. The Alzheimer’s Association puts Alzheimer’s at 60–80% of dementia cases (the WHO says 60–70%); vascular dementia, Lewy body disease and frontotemporal degeneration make up most of the rest, and in people in their 80s more than one is often present at once. That matters here for two reasons. A trial that enrolled “dementia” can’t tell you what happens in Alzheimer’s in particular, and a claim about amyloid or tau — the plaques and tangles that define Alzheimer’s — has to be tested in people who actually have them. None of the trials below confirmed Alzheimer’s with a PET scan or spinal-fluid test; they used the ordinary clinical diagnosis, which is how most families get it too.

So the strongest Alzheimer’s-specific evidence is for agitation, from two THC-type drugs, over six and three weeks respectively — details in the next section. The one positive CBD-forward trial was not an Alzheimer’s trial. No trial in any population measured whether the disease itself changed.

One Alzheimer’s-only trial looked at cognition rather than agitation, and it deserves an honest paragraph because it will be quoted at you. In 2025 a Brazilian group published a 26-week, double-blind, placebo-controlled phase 2 trial of a microdose balanced extract — 0.35 mg of THC and 0.245 mg of CBD a day, a fraction of a single dispensary gummy — in people aged 60 to 80 with Alzheimer’s dementia. At 26 weeks the cannabis group scored significantly higher on the Mini-Mental State Exam than placebo; there was no difference on any secondary outcome and no difference in side effects. It is the longest cannabinoid trial in Alzheimer’s so far. It is also small, single, and measured a bedside screening score rather than daily function or biomarkers; a later editorial in the same journal is titled “a promising trial, although not yet ready for the prescription pad,” and it cautions specifically against reading it as a reason to use unregulated cannabis products. A larger replication would change this page. Until then it is one data point, and the dose it used is not something you can measure out of a retail product.

The amyloid-and-tau story, and why it hasn’t left the lab

This is the part that gets repeated on product pages. In 2006 a Scripps group showed in a test tube that THC binds the enzyme acetylcholinesterase and blocks it from clumping amyloid-beta — a real biochemical result, at concentrations nobody has shown are reached in a living human brain. In 2014 a Florida group found very low THC concentrations lowered amyloid in a cultured cell line. In mice engineered to over-produce amyloid (the APP/PS1 model), a THC-plus-CBD extract given early preserved memory and lowered soluble amyloid-42 (Aso, 2015), and eight months of daily oral CBD at 20 mg/kg prevented a social-recognition deficit without changing amyloid load at all (Cheng, 2014). The same Australian lab later found that a medium-dose THC-plus-CBD combination had no therapeutic effect in older APP/PS1 mice and actually disrupted object memory in the healthy controls (Aumer, 2025), and that even 100 mg/kg of CBD restored one memory measure while leaving the amyloid markers unchanged (Watt, 2026). The evidence about tau is thinner still; there is no human data on cannabinoids and tau at all.

Why hasn’t any of it translated? The usual reasons, all of which apply here: cell-culture concentrations and mouse doses don’t translate directly (20–100 mg/kg of CBD in a mouse is roughly 110–570 mg a day for a 70 kg adult by the standard body-surface conversion, or 1.4–7 grams by naive weight scaling — and the Avidekel trial, at 527 mg a day, measured no disease effect at all); amyloid-overproducing mice model one piece of a disease that takes decades in people; a lab marker moving is not memory or independence improving; THC’s sedation, dizziness and confusion cap how far a dose can be pushed in an 85-year-old; and every human trial so far has been a symptom trial or, in one case, a 26-week cognition-score trial; none measured amyloid, tau or daily function. A credible disease-modifying claim would need a long, adequately powered randomized trial in biomarker-confirmed Alzheimer’s showing slower decline in cognition and daily function, backed by amyloid or tau measures. No cannabinoid has been put through that. The Alzheimer’s Association, in a 2020 statement still on its site: “we still don’t understand the benefits and risks of taking cannabis-derived products.”

Alzheimer’s trials still running

Two registered studies to watch, both symptom trials: NAB-IT (NCT04516057), nabilone versus placebo for eight weeks in a target of 112 people with Alzheimer’s agitation, still recruiting with primary completion estimated April 2027; and a small open-label McLean Hospital study (no placebo) of a high-CBD, low-THC sublingual solution twice daily for behavioral symptoms in adults 55–90 with mild cognitive impairment or Alzheimer’s (NCT04075435), about a dozen participants, no longer recruiting, primary completion estimated December 2026. The THC-AD registry entry (NCT02792257) lists 84 enrolled; the paper analyzes 75 randomized, and a separate, industry-sponsored phase 2/3 trial of dronabinol as an oral solution (NCT07422311, sponsor Benuvia Therapeutics, target 140, one site in Adelaide, Australia) is registered but not yet recruiting, with completion estimated 2028. Nothing registered tests whether any cannabinoid slows Alzheimer’s.

THC-type drugs and agitation: the strongest signal, and its price

Agitation and aggression are often the part of dementia that breaks families. So it matters that this is where the evidence is least bad — and that every trial that found something also found sedation.

Nabilone, 2019. Herrmann and colleagues at Sunnybrook and Baycrest in Toronto ran a randomized, double-blind crossover trial: 39 people with moderate-to-severe Alzheimer’s and clinically significant agitation, average age 87, six weeks on nabilone (a synthetic cannabinoid; mean dose 1.6 mg a day) and six weeks on placebo. Agitation on the Cohen-Mansfield scale was about 4 points lower on nabilone (95% CI −6.5 to −1.5, p = 0.003); overall neuropsychiatric symptoms and caregiver distress also improved. The cost: sedation in 45% on nabilone versus 16% on placebo. On one cognition test scores favored placebo; on another they favored nabilone — no consistent cognitive benefit. A larger, 112-person nabilone trial (NAB-IT) is still running as of this writing.

Dronabinol, 2026. The THC-AD trial — Rosenberg, Forester and colleagues, five U.S. academic centers — randomized 75 people with moderate-to-severe Alzheimer’s and agitation to three weeks of dronabinol (5 mg a day for a week, then 10 mg in split doses) or placebo. Agitation on the Pittsburgh Agitation Scale improved moderately more with dronabinol than placebo (effect size 0.53, p = 0.015 — the trial’s own numbers; press coverage translates that to roughly a 30% greater reduction); a second agitation scale trended the same way without reaching significance. Sleep, daily functioning, cognition and intoxication scores didn’t differ. The only notable side effect was sleepiness. It’s the biggest THC-type trial yet, and it’s three weeks long.

Very low-dose THC, 2015. Van den Elsen’s group in the Netherlands gave 50 people with dementia 1.5 mg of oral THC three times a day for three weeks. Neuropsychiatric symptoms didn’t improve more than placebo (mean difference 3.2, 95% CI −3.6 to 10.0). Side effects were similar in both groups. What this shows is that this dose didn’t work. It does not show that a higher dose would — a higher dose could just as easily bring more confusion, dizziness or falls.

A THC:CBD oil in severe dementia, 2026. The newest trial cuts the other way. Bianchi and colleagues ran a double-blind crossover in five Geneva nursing homes: 25 people with severe dementia, an oral THC:CBD oil titrated up to 20 mg of THC and 40 mg of CBD a day, two eight-week periods. Agitation on the Cohen-Mansfield scale was no different from placebo. Staff used fewer as-needed psychiatric drugs during the active period, and no serious side effect was attributed to the oil — but the headline result is null, at doses higher than the trials that were positive.

Appetite, 1997. Volicer’s crossover trial of low-dose dronabinol in 15 people with Alzheimer’s who were refusing food reported weight gain and less disturbed behavior. Fifteen people, one study. Loss of appetite in dementia also has a list of fixable causes — swallowing problems, depression, pain, constipation, dental trouble, medication effects — that should be checked before any of this.

The trade-off in a frail older adult

The harms aren’t theoretical. Sedation showed up in every positive trial. Confusion and disorientation were among the most common events in the CBD-oil trial (below). THC can lower blood pressure on standing and cause dizziness; in someone who is already unsteady, that is a fall waiting to happen — and the dementia trials were far too small and short to measure fall or fracture risk. One small crossover study measured what THC does to balance in people with dementia: 1.5 mg twice a day increased postural sway when standing and walking — no one fell during the study, but the direction is not reassuring. The 2023 AGS Beers Criteria, the standard reference on medications to avoid in older adults, don’t list cannabinoids at all — which means there is no geriatric guidance either way, not that they’re cleared. What Beers does say is to avoid stacking three or more brain-active drug classes — it names antidepressants, antipsychotics, benzodiazepines, sleep medicines, opioids and antiepileptics, not cannabis. The inference that adding a cannabinoid to that list adds impairment is common sense and pharmacology, not a Beers recommendation.

Two Alzheimer’s-specific cautions. Most people with Alzheimer’s are on donepezil, rivastigmine or galantamine, and there are no clinical studies of how THC, CBD, nabilone or dronabinol interact with them — we searched. What is known is that those drugs can slow the heart, cause fainting and cut appetite, and cannabinoids can add dizziness, low blood pressure on standing, sedation and poor intake on top; the risk is practical, not a proven metabolic interaction, but it is the kind of stacking a pharmacist should see before the first dose. One more data point, for what it is: a 2026 post-hoc analysis of the 39-person nabilone trial found that people not taking a cholinesterase inhibitor were among those more likely to respond to nabilone, alongside higher pain scores, appetite problems, apathy and milder cognitive impairment. That is a pattern-search inside a small trial, not a reason to stop donepezil so that a cannabinoid “works” — the authors themselves say the profile needs confirming. And if the diagnosis is dementia with Lewy bodies, or Alzheimer’s with Lewy body features, the bar is higher again: those patients are unusually prone to hallucinations, delirium, blood-pressure drops and sedation-related falls, and there is no cannabinoid safety trial in that population at all.

CBD: one real trial, one real interaction problem

The trial. Hermush and colleagues in Israel ran a 16-week, double-blind, placebo-controlled trial of Avidekel oil — about 30% CBD and 1% THC — in 60 people with dementia and agitation. By week 16 the average daily dose had climbed to roughly 527 mg of CBD and 22 mg of THC. 60% on the oil had at least a 4-point drop in agitation, versus 30% on placebo (p = 0.03); half versus 15% had an 8-point drop. It is the best CBD-forward evidence there is, and it has four asterisks: the oil contained THC, so it can’t tell you what CBD alone does; sleepiness (49% vs 30% on placebo) and confusion or disorientation (46% vs 30%) were common and every side effect trended against the oil without reaching significance in a 60-person trial; there were nine serious adverse events on the oil versus four on placebo, none attributed to it, and all eight dropouts were on the oil; and the study was funded by the oil’s maker. On the plus side of the ledger, caregiver-reported sleep disturbance fell more on the oil (66% to 28%) than on placebo (75% to 60%).

The Avidekel protocol, exactly as the trial ran it — and why there is no kitchen version

Readers asked for a recipe. There isn’t one, and the arithmetic below is why. What there is is the exact protocol the trial used, which is the only dosing schedule for a CBD-rich oil that has ever been tested against placebo in people with dementia. It is here so you can hand it to a physician, not so you can run it yourself.

Why you can’t make it. A home infusion is flower steeped in oil; the oil ends up holding whatever cannabinoids leach out of the plant matter. Take an absurdly generous case — a full ounce (28 g) of 15% CBD flower in one cup (237 mL) of oil with perfect, lossless extraction — and you get 4,200 mg of CBD in 237 mL, or about 18 mg per mL. Avidekel is 295 mg per mL. To match it in a cup of oil you’d need roughly 70,000 mg of CBD going in: more than a pound of high-CBD flower, before losses, in a single cup. It is made by concentrating an extract, not by infusing flower, and it is not on any Ohio or Michigan menu we track. The other problem is bigger than potency: a homemade oil has an unknown CBD number, an unknown THC number that could easily exceed the trial’s 0.5 mg per drop, no batch consistency, and no pesticide, solvent or microbial testing — given to someone who can’t tell you they feel drugged. That is the one product category where “lab-tested and labeled” is not optional.

What Ohio dispensaries actually sell, in the same units. Ohio menus carry CBD-dominant tinctures, and they are legitimate lab-tested products — but ratio and concentration are two different things, and none of them is Avidekel-strength. From the menus in our Edible Decoder as of this writing:

The point of the table is not that you should buy four bottles of tincture a week. It is that the “CBD for dementia” trial ran at doses that are an order of magnitude above what a bottle of dispensary tincture is designed to deliver, with THC riding along at 22 mg a day — a genuinely psychoactive amount in an 80-year-old. If a physician wants to try a CBD-dominant product, the honest framing is that it will be a much lower dose than the only positive trial used, nobody knows whether lower doses do anything, and a higher dose is not the answer either, given the sedation and fall risk.

The interaction. CBD is generally less intoxicating than THC, but at real doses it changes how the liver metabolizes other drugs. The cleanest documented example: a man on warfarin started pharmaceutical CBD for epilepsy and his INR (a measure of how thin the blood is) climbed from a stable ~2.0–2.6 to a peak of 6.86; his warfarin dose had to be cut by about 30%. The FDA label for prescription CBD (Epidiolex) calls it a moderate inhibitor of the liver enzyme CYP2C19 — the pathway for clopidogrel, some antidepressants, some benzodiazepines and acid reducers — and warns that combining it with other sedating drugs adds sedation. Clobazam, valproate, tacrolimus and other narrow-margin drugs are also documented concerns, though the case reports involve doses far above what a gummy delivers. The right move is not “avoid CBD” — it’s a pharmacist reviewing the whole list before the first dose, and closer monitoring after any change.

CBG: interesting mice, two tiny human studies, no dementia data

Cannabigerol has genuine laboratory results — anti-inflammatory and neuroprotective-like effects in cell cultures and in animal models of neurodegenerative disease. What it does not have is a single human trial in dementia — none registered on ClinicalTrials.gov, none published that we could find. The entire controlled human literature is two small studies in healthy adults. A 2024 crossover trial gave 34 people a single 20 mg dose of hemp-derived CBG or placebo over Zoom and found lower self-rated anxiety across the session and lower stress before, but not after, a lab stress test, plus a small verbal-memory boost; no intoxication and no impairment. A 2026 Johns Hopkins / University of Maryland dosing study gave 12 people single doses of 25, 50, 100 and 200 mg of CBG isolate: no drug-related adverse events, liver tests normal, blood levels rose with dose but varied a lot between people, and the subjective effects were minimal. A 2026 survey of 239 CBG users found they typically take 10–12 mg. That is the dosing knowledge base: single doses, young healthy people, no repeated-dose safety, no drug-interaction studies. The one human CBG trial did find a small verbal-memory boost in healthy adults after a single dose; that is not dementia evidence, and a product that says CBG “supports memory” in dementia is extrapolating from a dish and one afternoon in healthy volunteers.

CBN: sold for sleep, untested in dementia

Two things are true about cannabinol. First, the sleep evidence in people is modest and not about dementia: the best-known trial, a seven-night study of 321 adults who rated their sleep as poor, run and funded by the cannabis company Canopy Growth, found that 20 mg of CBN a night reduced night-time awakenings and overall sleep disturbance versus placebo but did not help people fall asleep faster, and the main sleep-quality endpoint only trended in CBN’s favor. A larger six-week trial in 2024 (1,020 adults, manufacturer-funded) found 25–100 mg of CBN reduced reported sleep disturbance about as much as 4 mg of melatonin — and no better. Second, a 2022 Salk Institute study found CBN protects cultured nerve cells from a kind of oxidative death by acting on mitochondria — a real finding, in a dish, with mouse work since and no human follow-up.

The practical caution matters more for a frail person than a healthy one. Sleep products are loosely regulated, and when an independent lab tested 52 CBD “sleep” products in 2022, about 60% were mislabeled (graded on a tiered scale whose best tier was within 10% of the label) and nearly half of the CBN-labeled ones were inaccurate. A 2025 peer-reviewed analysis of 97 hemp-derived products found only 13% within 10% of their labeled CBD content, and a “third-party tested” claim on the package did not improve the odds. Some CBN is made by chemically converting other cannabinoids. A seller’s own certificate is not the same as independent testing of the lot in your hand.

“So what dose?” — what has been studied, and what hasn’t

Several people asked for CBG and CBN dosing recommendations. We can’t give one, because no CBG or CBN dose has ever been tested in anyone with dementia, and a number we made up would look exactly as authoritative on this page as a number from a trial. What we can do is put the doses that have been studied in people, in one place, with who they were studied in:

Retail products sit in the same range — CBG gummies and tinctures are usually 5–25 mg a serving, CBN sleep products 2.5–10 mg, sometimes 20–25 — but those are marketing conventions, not evidence-based doses, and the label-accuracy problem above applies to all of them.

If a physician agrees to a monitored trial of a cannabinoid anyway, the two published dosing frameworks worth knowing are both explicitly not about dementia. A widely cited 2018 clinical review (MacCallum and Russo) says to start low, titrate over as much as two weeks, and keep total THC at or under about 30 mg a day, preferably alongside CBD, to avoid psychoactive effects and tolerance. A 2021 international expert consensus on chronic pain (Bhaskar and colleagues) wrote a conservative protocol for patients “more sensitive to drug effects,” naming “clinically frail patients” with “complex comorbidities” and “polypharmacy” as appropriate for it: a CBD-dominant product at 5 mg CBD once a day, raised by 10 mg a day every two to three days to a ceiling of 40 mg a day; THC considered only after that, at 1 mg a day, raised by 1 mg a week. Notice how far that is from Avidekel’s 527 mg of CBD and 22 mg of THC: the profession’s cautious starting point and the one positive dementia trial are a hundredfold apart on CBD and twentyfold on THC, and nothing in between has been tested. That gap, not a dose, is the honest answer.

Before any of this: rule out the fixable causes

Every geriatrician will tell you the same thing, so I’ll say it plainly. A sudden change in behavior — new agitation, sleepiness, hallucinations, confusion — is a medical warning sign, not a reason to reach for a gummy. Pain, constipation, urinary retention, infection, dehydration, hunger, a hearing aid that died, loneliness, overstimulation, a new medication: all of these produce “agitation,” and all of them are treatable. The trials above enrolled people whose symptoms persisted after that kind of workup.

The Ohio part

Alzheimer’s disease is a qualifying condition under Ohio’s medical marijuana law (Ohio Revised Code 3796.01, current as of March 20, 2026). Dementia in general is not on the list — vascular, Lewy body and frontotemporal dementia do not qualify by name, though chronic pain and Parkinson’s disease do. A chart that says “unspecified dementia” does not meet the Alzheimer’s line; mixed dementia does only if Alzheimer’s disease is actually documented as part of the diagnosis and the recommending physician is willing to certify it. A card means a recommending physician who can help with dosing, a registered caregiver who can legally buy, store and give the product to someone who can’t manage it themselves, and access to lab-tested, labeled products from a licensed dispensary rather than a gas-station CBN gummy. Since March 2026 the medical and adult-use programs share one regulator and one chapter of law; the rules on possession, purchase limits and where you can use are in our Ohio weed laws explainer. Qualifying status is a legal category, not evidence that it works — and dispensary flower, vapes and edibles are not the studied products. Nabilone, dronabinol and Avidekel oil are standardized oral products; a 25%-THC flower or a 100 mg gummy bag is not a substitute with known dosing.

If you’re still considering it: a caregiver checklist

1. Get sudden changes evaluated first. New agitation, sleepiness, hallucinations or confusion can be delirium from infection, pain, dehydration, constipation, urinary retention or a medication.
2. Loop in the doctor and the pharmacist before the first dose. The pharmacist review is the one people skip. Name every prescription and supplement, and flag warfarin, seizure drugs, valproate, tacrolimus, opioids, benzodiazepines, antipsychotics, sleep medicines, gabapentin and blood-pressure drugs.
3. Pick one measurable target. Aggressive episodes per day. Night-time awakenings. Refusals of care. Not “seems calmer.”
4. One oral product, one variable at a time, lowest dose. Oral cannabinoids take hours to peak and can last a long time; redosing too soon is how people get too much. Don’t change two things in the same week.
5. Write it down. Dose, time, the target symptom, alertness, walking, appetite, sleep. A notebook is how you’ll know whether anything is happening.
6. Set stopping rules before you start. Stop and call for more confusion, hallucinations, excessive sleepiness, new unsteadiness, fainting, vomiting, poor intake, trouble swallowing, or any fall.
7. Know the milligrams per mL, not just the ratio. A “20:1” tincture can be 30 mg of CBD per mL or 300; the one positive CBD trial used 295. Write the per-drop dose on the bottle.
8. Verify the exact lot. An independent certificate of analysis showing measured THC, CBD and CBN, plus pesticide, solvent, microbial, mycotoxin and heavy-metal testing. Keep products locked away.
9. Reassess in weeks, not months, and stop if there’s no clear benefit. The positive trials lasted three to sixteen weeks. Indefinite use without documented benefit isn’t supported by anything above.

Questions families actually ask

Sources

1. Bosnjak Kuharic D, et al. Cannabinoids for the treatment of dementia. Cochrane Database Syst Rev. 2021;9:CD012820. doi:10.1002/14651858.CD012820.pub2 (PMID 34532852). Four trials, 126 participants; very-low to low certainty.
2. Herrmann N, Ruthirakuhan M, Gallagher D, et al. Randomized placebo-controlled trial of nabilone for agitation in Alzheimer’s disease. Am J Geriatr Psychiatry. 2019;27(11):1161–1173. doi:10.1016/j.jagp.2019.05.002 (PMID 31182351). n=39; CMAI b = −4.0 (−6.5 to −1.5), p=0.003; sedation 45% vs 16%.
3. Rosenberg PB, Forester BP, et al. A randomized controlled trial of the safety and efficacy of dronabinol for agitation in Alzheimer’s disease (THC-AD). Am J Geriatr Psychiatry. 2026;34(2):167–179. doi:10.1016/j.jagp.2025.10.011 (PMID 41350162); trial registration NCT02792257. n=75, 3 weeks, 5→10 mg/day; Pittsburgh Agitation Scale effect size 0.53, p=0.015.
4. van den Elsen GAH, Ahmed AIA, Verkes RJ, et al. Tetrahydrocannabinol for neuropsychiatric symptoms in dementia: a randomized controlled trial. Neurology. 2015;84(23):2338–2346. doi:10.1212/WNL.0000000000001675 (PMID 25972490). n=50, 1.5 mg THC three times daily, no significant NPI difference.
5. Volicer L, Stelly M, Morris J, McLaughlin J, Volicer BJ. Effects of dronabinol on anorexia and disturbed behavior in patients with Alzheimer’s disease. Int J Geriatr Psychiatry. 1997;12(9):913–919. PMID 9309469. n=15 crossover.
6. Hermush V, Ore L, Stern N, et al. Effects of rich cannabidiol oil on behavioral disturbances in patients with dementia: a placebo controlled randomized clinical trial. Front Med (Lausanne). 2022;9:951889. doi:10.3389/fmed.2022.951889 (PMID 36148467; PMC9486160). n=60 (2:1), mixed dementia types; oil 295 mg CBD / 12.5 mg THC per mL, one 0.04 mL drop = 11.8 mg CBD + 0.5 mg THC; start 1 drop three times daily, +1 drop per dose every 2 days, max 21 drops per dose; mean end dose 527.5 mg CBD + 22.3 mg THC/day; CMAI ≥4-point responders 60% vs 30%, p=0.03; funded by TO Pharmaceuticals. Registration NCT03328676.
7. Grayson L, Vines B, Nichol K, Szaflarski JP. An interaction between warfarin and cannabidiol, a case report. Epilepsy Behav Case Rep. 2017;9:10–11. PMC5789126. INR peaked at 6.86; ~30% warfarin dose reduction.
8. Bianchi F, Broers B, Desmeules JA, et al. A randomised, double-blind, placebo-controlled crossover trial on cannabinoid-based medication for behavioural disorders in old patients with severe dementia: the study results. Age Ageing. 2026. doi:10.1093/ageing/afag239 (PMID 42613673). n=25; up to 20 mg THC / 40 mg CBD daily; no CMAI difference vs placebo.
9. Pan TJ, Wang HJ, Siddiqui A, et al. Efficacy and safety of cannabinoids for neuropsychiatric symptoms of dementia: a systematic review with meta-analysis. CNS Drugs. 2026. doi:10.1007/s40263-026-01277-w (PMID 41748849). 10 RCTs, 328 participants; agitation SMD −0.52 (−1.00 to −0.05), not significant after excluding high-risk-of-bias trials; sedation RR 2.09.
10. da Silva AMP, de Siqueira Lima DV, Figueiroa Souza R, et al. Cannabinoids treatment for agitation in Alzheimer’s disease: a systematic review and meta-analysis with Bayesian and sequential trial analyses. Am J Geriatr Psychiatry. 2026. doi:10.1016/j.jagp.2026.05.016 (PMID 42315374). 7 studies, 221 participants; NPI agitation/aggression SMD −0.47; somnolence RR 2.25; falls “imprecisely estimated.”
11. van den Elsen GAH, et al. Effects of tetrahydrocannabinol on balance and gait in patients with dementia: a randomised controlled crossover trial. J Psychopharmacol. 2017. doi:10.1177/0269881116665357. n=18; THC 1.5 mg twice daily increased postural sway; no falls.
12. American Geriatrics Society Beers Criteria Update Expert Panel. 2023 updated AGS Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2023;71(7):2052–2081. doi:10.1111/jgs.18372. Cannabinoids are not listed; cited for its general guidance on combining CNS-active drugs.
13. Bonn-Miller MO, et al. A double-blind, randomized, placebo-controlled study of the safety and effects of CBN with and without CBD on sleep quality. Exp Clin Psychopharmacol. 2024;32(3):277–284. doi:10.1037/pha0000682. n=321, 20 mg CBN, 7 nights; funded by Canopy Growth, authors company-employed. Sleep quality OR 2.26 (0.93–5.52), p=0.082; awakenings and overall disturbance reduced; no effect on sleep onset.
14. Kolobaric A, et al. A randomized, double-blind, placebo-controlled trial to assess the effectiveness and safety of melatonin and three formulations of Floraworks proprietary TruCBN for improving sleep. Pharmaceuticals. 2024. PMC11357382. n=1,020, 6 weeks; CBN 25–100 mg comparable to 4 mg melatonin; manufacturer-funded.
15. Liang Z, Soriano-Castell D, Kepchia D, et al. Cannabinol inhibits oxytosis/ferroptosis by directly targeting mitochondria independently of cannabinoid receptors. Free Radic Biol Med. 2022;180:33–51. PMID 34999187. Cell-culture study, Salk Institute.
16. Leafreport / Infinite Chemical Analysis Labs. Most CBD sleep products are mislabeled (52 products tested; 60% inaccurate; almost half of CBN-labeled products inaccurate; tiered grading, best tier ±10%). October 2022. leafreport.com. Industry-commissioned consumer testing, not peer-reviewed.
17. Dowd AN, et al. Cannabinoid content and label accuracy of hemp-derived products. Cannabis Cannabinoid Res. 2025. PMID 39029473. 97 products; 12.7% within ±10% of labeled CBD; third-party-testing claims did not improve accuracy.
18. U.S. FDA / DailyMed. EPIDIOLEX (cannabidiol) prescribing information, revised May 2026 — drug interactions (CYP2C19 inhibition; CNS depressants). dailymed.nlm.nih.gov.
19. Ohio Revised Code § 3796.01 (definitions; qualifying medical conditions), effective March 20, 2026. codes.ohio.gov.
20. Alzheimer’s Society (UK). Cannabis, CBD oil and dementia. alzheimers.org.uk. “Currently there is no evidence to show that cannabis or cannabis oil (CBD oil) can stop, reverse or prevent dementia.”
21. Alzheimer’s Association. Cannabis: helpful or harmful? (March 2020). alz.org. “We still don’t understand the benefits and risks of taking cannabis-derived products.”
22. ClinicalTrials.gov NCT04516057, Nabilone for Agitation Blinded Intervention Trial (NAB-IT), target n=112, recruiting; primary completion estimated April 2027; no results posted. clinicaltrials.gov.
23. Alzheimer’s Association. What is Alzheimer’s disease? / What is dementia? alz.org. “Alzheimer’s disease accounts for 60–80% of cases.” World Health Organization, Dementia fact sheet: “may contribute to 60–70% of cases.” who.int.
24. Eubanks LM, Rogers CJ, Beuscher AE, et al. A molecular link between the active component of marijuana and Alzheimer’s disease pathology. Mol Pharm. 2006;3(6):773–777. doi:10.1021/mp060066m (PMID 17140265). In vitro: THC competitively inhibits acetylcholinesterase and AChE-induced amyloid-β aggregation.
25. Cao C, Li Y, Liu H, et al. The potential therapeutic effects of THC on Alzheimer’s disease. J Alzheimers Dis. 2014;42(3):973–984. doi:10.3233/JAD-140093 (PMID 25024327). Cell culture (N2a/AβPPswe); low THC concentrations lowered Aβ.
26. Aso E, Sánchez-Pla A, Vegas-Lozano E, Maldonado R, Ferrer I. Cannabis-based medicine reduces multiple pathological processes in AβPP/PS1 mice. J Alzheimers Dis. 2015;43(3):977–991. doi:10.3233/JAD-141014 (PMID 25125475). THC + CBD extracts preserved memory and reduced soluble Aβ42 in transgenic mice.
27. Cheng D, Spiro AS, Jenner AM, Garner B, Karl T. Long-term cannabidiol treatment prevents the development of social recognition memory deficits in Alzheimer’s disease transgenic mice. J Alzheimers Dis. 2014;42(4):1383–1396. doi:10.3233/JAD-140921 (PMID 25024347). Oral CBD 20 mg/kg daily for 8 months; deficit prevented; no change in amyloid load or oxidative damage.
28. Aumer B, Rosa Porto R, Coles M, et al. Combination treatment with medium dose THC and CBD had no therapeutic effect in a transgenic mouse model for Alzheimer’s disease but affected other domains. Pharmacol Biochem Behav. 2025. doi:10.1016/j.pbb.2025.174101 (PMID 40976394). 3 mg/kg THC + 20 mg/kg CBD, 3 weeks, 14.5-month-old APP/PS1 females: no therapeutic effect; disrupted object recognition in controls.
29. Watt G, Olaya J, Muench G, Garner B, Karl T. Effects of chronic 100 mg/kg cannabidiol treatment in male double transgenic APPSwe/PS1ΔE9 mice. Pharmaceuticals (Basel). 2026;19(3):374. doi:10.3390/ph19030374 (PMID 41901221). Social-recognition deficit restored; no change in soluble or insoluble Aβ42.
30. Watt G, Karl T. In vivo evidence for therapeutic properties of cannabidiol (CBD) for Alzheimer’s disease. Front Pharmacol. 2017;8:20. doi:10.3389/fphar.2017.00020 (PMID 28217094). Review of the preclinical (animal) CBD literature.
31. Cuttler C, Stueber A, Cooper ZD, Russo E. Acute effects of cannabigerol on anxiety, stress, and mood: a double-blind, placebo-controlled, crossover, field trial. Sci Rep. 2024;14:16163. doi:10.1038/s41598-024-66879-0 (PMID 39003387). n=34 healthy adults; single 20 mg hemp-derived CBG; lower anxiety ratings across the session and lower stress before (not after) the stress test; enhanced verbal memory; no subjective drug effects or impairment.
32. Wolinsky D, Srungaram D, Zhang A, et al. Safety, pharmacodynamics, and pharmacokinetics of oral cannabigerol in healthy adults. J Pharmacol Exp Ther. 2026. doi:10.1016/j.jpet.2026.104984 (PMID 42575778). n=12; single ascending doses 0/25/50/100/200 mg CBG isolate; no drug-related adverse events; liver tests normal; dose-orderly pharmacokinetics with large individual variability.
33. Lutz ETH, Sulak D, Vinocur L, Russo EB, Cuttler C. A survey of cannabigerol users’ patterns of use, perceived efficacy and satisfaction. J Psychoactive Drugs. 2026. doi:10.1080/02791072.2026.2727651 (PMID 42741890). n=239; typical self-reported dose 10–12 mg; onset 15–60 min, duration 4–6 h. Survey, not a trial.
34. MacCallum CA, Russo EB. Practical considerations in medical cannabis administration and dosing. Eur J Intern Med. 2018;49:12–19. doi:10.1016/j.ejim.2018.01.004 (PMID 29307505). Total daily THC “should generally be limited to 30 mg/day or less, preferably in conjunction with CBD”; titrate slowly over as much as two weeks. Not dementia-specific.
35. Bhaskar A, Bell A, Boivin M, et al. Consensus recommendations on dosing and administration of medical cannabis to treat chronic pain: results of a modified Delphi process. J Cannabis Res. 2021;3:22. doi:10.1186/s42238-021-00073-1 (PMID 34215346). Conservative protocol for frail, comorbid or polypharmacy patients: CBD-predominant 5 mg once daily, +10 mg/day every 2–3 days to 40 mg/day; THC only thereafter at 1 mg/day, +1 mg weekly. Chronic pain, not dementia.
36. McKeith IG, Boeve BF, Dickson DW, et al. Diagnosis and management of dementia with Lewy bodies: fourth consensus report of the DLB Consortium. Neurology. 2017;89(1):88–100. doi:10.1212/WNL.0000000000004058 (PMID 28592453). Cited for the sensitivity of DLB patients to sedating and psychoactive drugs.
37. ClinicalTrials.gov NCT04075435, Cannabidiol solution for the treatment of behavioral symptoms in older adults with mild cognitive impairment or Alzheimer’s disease; open-label single-arm (no placebo), McLean Hospital, ages 55–90; high-CBD/low-THC sublingual solution twice daily; estimated enrollment 12; active, not recruiting; primary completion estimated December 2026. clinicaltrials.gov. NCT02792257 (THC-AD): completed May 2024, 84 enrolled. clinicaltrials.gov.
38. Cury RM, da Silva T, Cezar-Dos-Santos F, et al. A randomized clinical trial of low-dose cannabis extract in Alzheimer’s disease. J Alzheimers Dis. 2025. doi:10.1177/13872877251389608 (PMID 41160460). Phase 2, 26 weeks, 0.350 mg THC + 0.245 mg CBD daily, ages 60–80; MMSE significantly higher vs placebo at week 26; no difference in secondary outcomes or adverse events; ReBEC U1111-1258-2058. Editorial: Dos Santos Silva J. Cannabinoids for Alzheimer’s disease: a promising trial, although not yet ready for the prescription pad. J Alzheimers Dis. 2026. doi:10.1177/13872877261452183 (PMID 42216668).
39. Feldman OJ, Herrmann N, Ruthirakuhan M, et al. Assessment of clinical factors that predict response to nabilone for agitation in Alzheimer’s disease: a post hoc analysis of a randomized placebo-controlled trial. Int Psychogeriatr. 2026. doi:10.1016/j.inpsyc.2026.100183 (PMID 41530008). n=39; five response predictors including no concomitant cholinesterase inhibitor (CMAI response difference −13.9 [SE 4.4]); top tertile 82% vs 40% responders.
40. ClinicalTrials.gov NCT07422311, dronabinol oral solution for agitation in Alzheimer’s disease, phase 2/3, estimated enrollment 140, not yet recruiting, primary completion estimated May 2028. clinicaltrials.gov.
41. Dispensary menu data: UB Good Relief 20:1 CBD:THC tincture (label: 2,200 mg CBD, 110 mg THC per bottle) and Dr. Solomon’s CBD Rich tincture (500 mg CBD per 15 mL) as listed on Northeast Ohio dispensary menus in the 420Blazin Edible Decoder feed, September 27, 2026. Menus change; check the current certificate of analysis.

With thanks to the budtender at RISE Cleveland (Brookpark & Pearl) who suggested this post. Budtenders hear these questions first; if you work at a dispensary and there’s a question you keep getting that deserves an honest, sourced answer, tell us.

Educational only, not medical advice. The author is a cannabis writer, not a clinician. Every study above was checked against its primary publication on September 27, 2026, and the draft was independently reviewed for accuracy by a second model before publication; trials in this area are small and the evidence can change. Talk to the person’s doctor and pharmacist before starting, stopping or combining anything. Ohio medical marijuana requires a physician recommendation and registration. Cannabis remains a Schedule I controlled substance under federal law.