CBG: measured effects, products and precautions in France
CBG, or cannabigerol, is a hemp cannabinoid that does not get you high, and its effects in humans have barely been measured. A 2024 trial in 34 adults recorded a slightly larger drop in anxiety than on placebo; a 2025 trial in 63 veterans found no difference. In France it is sold mostly as resins, at 16 to 20% CBG, and as flowers, at about 1%. CBG products meant to be swallowed are not authorised.
Key takeaways
- Human data come down to three controlled trials: 20 mg in a single dose, 25 to 50 mg a day for four weeks, then single doses of 25 to 200 mg; the first found a small gap with placebo, the other two no clear effect.
- CBG has no authorised health claim, and EFSA's 2 mg a day benchmark covers CBD only.
- In the Phytogrammes catalogue, the CBG-rich resins show 16.3 to 20.5% CBG and 0.1 to 0.2% Δ⁹-THC.
- In lab tests, CBG slows liver enzymes that clear many medicines.
- Since May 2026, a food or supplement whose packaging displays CBG can be withdrawn from sale.
What studies in humans have measured
CBG is the neutral form of CBGA, the molecule from which the plant makes CBD and THC. That chemistry is covered in our article on the definition of CBG. This section asks a practical question: what was observed in people who took it?
| Study | People | Dose | Method | Result |
|---|---|---|---|---|
| Cuttler et al., 2024 | 34 healthy adults | 20 mg, single dose | Placebo-controlled, double-blind | Anxiety down 26.5% vs 22.5% on placebo, no intoxication |
| Emerson et al., 2025 | 63 veterans | 25 then 50 mg a day, 4 weeks | Placebo-controlled, triple-blind | No difference on sleep, quality of life or post-traumatic stress |
| Wolinsky et al., 2026 | 12 healthy adults | 25, 50, 100 and 200 mg, rising single doses | Placebo-controlled, blinded | No product-related adverse event, no clear pharmacodynamic effect |
| Russo et al., 2022 | 127 users | Variable | Online survey | Effects reported by users themselves, no control group |
The first controlled trial, published in Scientific Reports, gave the same 34 participants 20 mg of CBG and a placebo, one week apart, in remote sessions. The gap in anxiety between the two conditions is small: 4 percentage points. Word memory improved slightly, by about one word on a list. The researchers measured no intoxication or impairment.
The second trial, published in 2025 in Medical Cannabis and Cannabinoids, followed 33 veterans on CBG and 30 on placebo for four weeks and found no difference on its outcomes. Participants tolerated CBG well. The maker of the product under study funded the trial.
The third, published online in July 2026 in the Journal of Pharmacology and Experimental Therapeutics, gave 12 healthy adults rising single doses of CBG, with a placebo placed at random in the sequence, blinded. It found no product-related adverse event and no clear pharmacodynamic effect, and subjective effects stayed minimal. One of the authors is affiliated with a CBD company.
The survey by Russo and colleagues asked 127 Americans who use preparations with more than 50% CBG about their habits. Participants report their own uses and impressions. Without a placebo or any measurement, such a survey tells you what people are looking for; it cannot show what the product does.
In the lab, CBG acts on many targets. A 2021 review in Frontiers in Pharmacology reports that it weakly activates the CB1 and CB2 receptors, activates α2-adrenergic receptors and blocks the serotonin 5-HT1A receptor. These results come from cells and animals and say nothing about what a dose does in humans.
The CBG formats sold in France, and their status
The format determines which law applies. Hemp falls under the order of 30 December 2021, which sets authorised varieties and a ceiling of 0.30% Δ⁹-THC in the finished product. Anything swallowed is also subject to food law.
| Format | Typical CBG level | Status in September 2026 |
|---|---|---|
| CBD-dominant flower | About 1% | On sale, if not presented as a food |
| CBG-rich resin | 16 to 20% in the catalogue batches (reference profile) | On sale, capped at 0.30% Δ⁹-THC |
| Oil, capsule or gummy to swallow | As labelled | Unauthorised novel food, withdrawn from sale |
| CBG cosmetic | Depends on the formula | On sale, with no medical care claim |
Since 15 May 2026, France's food directorate (DGAL) has run a control plan on products to swallow. The Provence-Alpes-Côte d'Azur chamber of agriculture summarises the rule: a food product whose packaging mentions CBD, THC "or any other cannabinoid" may be withdrawn, and CBG is one of them. Our update on CBD rules in 2026 covers the full framework.
CBG-rich resins: what the analyses show
A mature flower keeps little CBG, since the plant converts almost all of it into CBD or THC. Growers who want a lot of it choose varieties whose conversion enzyme works poorly, then concentrate the trichomes by sieving. Read on 23 September 2026, the catalogue product pages give the reference levels below; no laboratory certificate is published for them.
| Product | Format | CBG | CBD | Δ⁹-THC | Batch |
|---|---|---|---|---|---|
| Strawsift Cbg/cbd | Resin | 20.5% | 4.7% | 0.2% | PG-26-50C3 |
| Rosift Cbg/cbd | Resin | 20.2% | 4.6% | 0.2% | PG-26-3A2B |
| CBG 90µ | Resin | 19.3% | 4.2% | 0.1% | PG-26-83EF |
| Afghan | Resin | 18.5% | 4.0% | 0.2% | PG-26-ED80 |
| Rosift | Resin | 18.4% | 3.9% | 0.1% | PG-26-EF5C |
| Strawsift | Resin | 16.3% | 3.1% | 0.2% | PG-26-32AE |
| Junky Milk CBD-CBG | Flower | 1.0% | 8.3% | 0.2% | PG-26-BA8C |
Three points stand out. CBG weighs four to five times as much as CBD in these resins. Δ⁹-THC stays between 0.1 and 0.2%, under the 0.30% threshold. And the product name can mislead: the Junky Milk flower carries "CBG" in its name yet contains only 1.0%, and some resin names state a CBD level that differs from the batch's reference profile. Go by the figure on the batch line. All these products are in the resins, hash and pollens category.
Reading the analysis of a CBG product
Four lines are enough to sort products. The batch number must match the product sold. CBG is read together with CBGA, its acid form: after heating, the total is CBG + 0.878 × CBGA, because CBGA loses about 12% of its mass when it decarboxylates. Δ⁹-THC must be 0.30% or less. Last, you can check for the absence of HHC, H4CBD and THCP, which the French medicines agency (ANSM) classified as narcotics in 2023 and 2024, if the lab looked for them.
Each section is explained in our guide to the certificate of analysis, and the laboratory page describes the measurement methods.
Precautions: medicines, pregnancy, driving
There is no dose benchmark for CBG. The provisional safe level published by EFSA on 9 February 2026, about 2 mg a day for 70 kg, covers CBD at least 98% pure, and CBD alone. No agency has run the same assessment for CBG. Where that figure comes from is explained in our article on EFSA's reference dose.
Tests on human enzymes have shown that CBG, like other cannabinoids, inhibits CYP2C9, a liver enzyme that clears many treatments (Doohan et al., 2021). In human liver fractions, a CBG-rich extract also slowed CYP2C9 and CYP3A4 (Treyer et al., 2023). These in vitro results put no figure on the real risk. They are still a reason to talk to your doctor or pharmacist if you take a treatment. Our article on the grapefruit effect describes the mechanism.
No data exist on CBG during pregnancy or breastfeeding. For CBD, the French food safety agency ANSES proposed on 21 March 2025 classifying it as presumed toxic for reproduction. As a precaution, these periods rule out CBG as well, and the same reservations apply to minors.
On the road, the saliva test detects THC and does not look for CBG. A resin at 0.2% Δ⁹-THC still contains traces of THC, and a test can pick them up after use. Our drug testing guide gives the timings.
What marketing is not allowed to claim
CBG is often sold as the "focus" cannabinoid, a label no trial supports. Regulation (EC) No 1924/2006 allows health claims on foods only when they appear in the EU register, and none of them concerns hemp or its cannabinoids. Under French law, a product presented as treating or preventing a disorder becomes a medicine by presentation.
Epidyolex, prescribed for rare forms of epilepsy, is the only authorised medicine in this family, and its active substance is CBD, not CBG. A product page that promises a specific effect on mood, sleep or pain from CBG goes beyond what research has measured.
Frequently asked questions
Only three controlled trials exist in humans. In 2024, anxiety fell by 26.5% after 20 mg in 34 adults, against 22.5% on placebo, with no intoxication. In 2025, 25 to 50 mg a day for four weeks made no difference compared with placebo in 63 veterans. In 2026, rising single doses produced no clear effect in 12 adults.
No. In the 2024 trial, 20 mg caused no intoxication and no measurable impairment. The risk comes from what the product actually contains: Δ⁹-THC above 0.30%, or an added semi-synthetic cannabinoid.
No authority has set a dose for CBG. The trials used a single 20 mg dose, then 25 to 50 mg a day; these figures are neither advice nor a safety threshold. EFSA's 2 mg a day benchmark applies to CBD alone.
Yes, as flowers, resins or cosmetics made from authorised varieties with no more than 0.30% Δ⁹-THC. Under novel food rules, CBG oils, capsules or gummies meant to be swallowed are not authorised.
The test targets THC. A CBG-rich resin still shows 0.1 to 0.2% Δ⁹-THC in the catalogue batches (reference profile), and those traces can be enough for a positive test after use.
Written by the Phytogrammes team
Each journal article draws on cited primary sources: ANSM, EFSA, EUR-Lex and peer-reviewed publications. Product profiles are distinguished from documented laboratory results. Our approach.
Article published on . CBD is not a medicine.