The CBD entourage effect: what studies actually tested

The term appears in a paper from Raphael Mechoulam's team, first-authored by Shimon Ben-Shabat and published in 1998 in the European Journal of Pharmacology. The authors were studying 2-AG, an endocannabinoid, meaning a molecule the body makes that activates cannabinoid receptors. Alongside it they found two related esters that do not bind to the receptors but boost the activity of 2-AG, in cell tests and in mice. They called this the "entourage effect".

In other words, the first description concerns the endocannabinoid system, not hemp, CBD or terpenes. The shift to the plant came later.

In 2011, the neurologist Ethan Russo published a review titled "Taming THC" in the British Journal of Pharmacology. He gathered laboratory data and proposed that cannabis phytocannabinoids and terpenes act together, each modulating the effect of the others. The paper is a review of hypotheses, not a trial: it lists leads to be tested.

Terpenes are the plant's aromatic molecules: myrcene, limonene, linalool, β-caryophyllene, the pinenes. Our article on hemp terpenes describes their chemistry. Russo's hypothesis has since become a common sales argument for so-called full-spectrum oils.

A hypothesis gets tested. If terpenes strengthened cannabinoids, it should show up on their targets, the CB1 and CB2 receptors in particular. Four studies checked this in cells.

So β-caryophyllene does bind to the CB2 receptor: it is the only common terpene for which this link is established. Our article on β-caryophyllene covers the research. For the others, the Australian team concluded in 2019 that if an entourage effect exists, it does not act through the CB1 or CB2 receptors. The New Zealand team reached the same finding in 2020 with a different method, receptor binding assays, and a third study found no effect on the TRPA1 and TRPV1 channels.

These results do not bury the hypothesis: terpenes might act elsewhere, for instance on absorption or on other targets. But the mechanism brands cite most often, terpenes that "boost" CBD at its receptors, has not been observed.

The only controlled human study we found on a terpene and a cannabinoid was run at Johns Hopkins University and published in 2024. Twenty healthy adults inhaled, double-blind, vaporised THC alone (15 or 30 mg), limonene alone (1 or 5 mg), both together, or a placebo. Twelve of them completed a tenth session with 30 mg of THC and 15 mg of limonene.

Limonene alone changed nothing compared with placebo. Combined with the 30 mg of THC, 15 mg of limonene lowered self-reported anxiety and paranoia scores compared with THC alone. THC's other effects, on cognition, psychomotor performance and vital signs, stayed the same. Two caveats: the study concerns THC, not CBD, and Ethan Russo is a co-author, with a patent application filed by the university on the use of limonene.

On the CBD side, we found no published controlled trial comparing a full-spectrum oil and an isolate at the same dose. The most cited data is a 2018 meta-analysis covering 670 patients with treatment-resistant epilepsy, drawn from 11 observational studies. Patients on CBD-rich extracts reported improvement more often (71% versus 46% on purified CBD). With the usual clinical criterion, a reduction of at least 50% in seizures, the difference disappears: 37% versus 42%. Two of the three authors were affiliated with companies in the sector, one of them called Entourage Phytolab, and the paper itself concludes that the effect remains to be confirmed in controlled trials.

The only authorised CBD medicine, Epidyolex, is in fact a purified CBD, not a full-spectrum extract.

The differences between formats are real, but they are chemical. They are read on a certificate, not in a promise of effect.

Two facts weigh more than the entourage effect. First: the THC traces in a full-spectrum product can show up in a drug test, even under 0.30%. Second: EFSA, the European Food Safety Authority, set in February 2026 a provisional safe level of about 2 mg of CBD a day for a 70 kg adult, for supplements containing CBD at least 98% pure. The benchmark therefore only applies to a highly purified, isolate-type CBD. Our article on EFSA's reference dose explains it.

To compare formats, read our comparison of full spectrum versus isolate and the definition of broad-spectrum CBD. The CBD oils in the catalogue state their format on each product page.

MILDECA, the French interministerial mission against drugs, points out that products containing CBD may not make therapeutic claims, under penalty of criminal sanctions, unless they are authorised as medicines. Presenting the entourage effect as a proven benefit amounts to promising a health effect.