The authors searched Scopus, Google Scholar, and PubMed for literature on Cannabis sativa phytochemicals published between January 2010 and May 2025, using search terms built around non-cannabinoid compounds and their pharmacological activity. That search initially returned 187 articles.

After applying inclusion and exclusion criteria, 12 studies qualified for the final review. The review organizes its findings around four non-cannabinoid compound classes: terpenes, flavonoids, phenolic compounds (including phenolic acids and spiro-phenols), and alkaloids.

Over 100 distinct terpenes have been identified in Cannabis sativa, produced in the same glandular trichomes that make cannabinoids and responsible for much of the plant’s characteristic smell. The review highlights monoterpenes such as D-limonene, linalool, beta-myrcene, and alpha-pinene, and sesquiterpenes such as beta-caryophyllene.

In the cited preclinical work, beta-caryophyllene binds directly to the CB2 cannabinoid receptor, myrcene has been associated with reduced IL-6 expression in inflammation models, and terpene blends including alpha-pinene showed antimicrobial activity against several bacterial genera in laboratory testing.

About 20 flavonoids have been identified in Cannabis sativa. Cannflavin A and B are unique to the plant and, in the cited cell studies, inhibited 5-lipoxygenase and microsomal prostaglandin E2 synthase-1 while showing minimal inhibition of COX-1 and COX-2, a pattern the review links to a lower theoretical risk of the gastrointestinal side effects associated with traditional COX-inhibiting anti-inflammatory drugs. This has not been tested in patients.

Other flavonoids identified in the plant, apigenin, luteolin, kaempferol, and quercetin, were studied mainly in cancer cell lines and animal tumor models. The review describes a proposed entourage mechanism in which cannabinoids and terpenes converge on shared targets, including TRP channels and CB2 receptors, and notes that CBD and some flavonoids may inhibit fatty acid amide hydrolase (FAAH), the enzyme that breaks down the body’s own endocannabinoid anandamide.

Across the 12 included studies, the review reports laboratory or animal-model signal for anti-inflammatory activity (beta-myrcene, cannflavin A and B), anti-cancer activity (quercetin-rich extracts, apigenin, kaempferol, and a cannflavin B analog called FBL-03G tested in pancreatic cancer models), anti-diabetic activity (flavonoid-rich extracts affecting glucose measures in rodent and cell models), analgesic activity (myrcene and limonene acting on adrenergic and pain-signaling pathways), and antimicrobial activity (several terpenes against Gram-positive and Gram-negative organisms).

Every one of these findings comes from cell culture or animal-model experiments. None involved human participants, and none evaluated an isolated non-cannabinoid compound as a standalone treatment in people.

The review points to several concrete barriers to clinical translation. Non-cannabinoid compounds are typically present in low concentrations, and there is a lack of validated reference standards and analytical methods for identifying and quantifying them precisely, which complicates accurate dosing.

Plant chemotype and cultivation method also matter: the review cites a comparison in which the flavonoid content of one cannabis strain, Carmagnola Cs, ran up to 25 percent higher than in another strain, Kompolti. Most existing human data comes from trials of whole-plant or full-spectrum extracts that combine cannabinoids and non-cannabinoids together, which the authors note makes it difficult to isolate the individual contribution of any single non-cannabinoid compound.

The authors conclude that non-cannabinoid phytochemicals are integral to the pharmacological complexity of Cannabis sativa and warrant further investigation as candidates for drug development, particularly in formulations designed to optimize whole-plant therapeutic effects.

They call explicitly for further research into non-cannabinoid interactions, pharmacokinetics, and clinical efficacy, stating that such research is needed to understand whether the synergy they describe actually underlies the enhanced outcomes sometimes reported with full-spectrum extracts compared with isolated cannabinoids.

This review sits inside a much longer-running debate about the ‘entourage effect,’ a term first proposed by Mechoulam and Ben-Shabat in 1998 to describe synergistic interactions among cannabis compounds. The idea has since become a common marketing claim for full-spectrum products, often stated with far more confidence than the underlying evidence supports.

What this review adds to that debate is not new proof of the entourage effect in patients, but a more organized accounting of which non-cannabinoid compounds have documented laboratory activity, which molecular targets they might share with cannabinoids, and how far the field still is from testing any of this directly in people.

I have spent more than 20 years practicing evidence-based cannabis medicine, and ‘entourage effect’ is a phrase I hear used far more often than it is actually sourced. This review is useful to me precisely because it resists that temptation. It names specific compounds, beta-caryophyllene binding CB2 receptors, cannflavin A and B inhibiting inflammatory enzymes, myrcene’s association with reduced IL-6, and it is honest that every one of those findings comes from a dish or an animal model, not a patient.

When a patient asks me whether a full-spectrum product will work better for them because of its terpene or flavonoid content, this review tells me I can talk about plausible mechanisms and promising early laboratory signal, but I cannot yet tell them that science has confirmed a clinical benefit from any specific non-cannabinoid compound. That distinction matters. It is not a reason to dismiss whole-plant formulations, which many patients report benefiting from, but it is a reason to be precise about what is mechanism, what is preclinical signal, and what remains, for now, unproven in humans.

A Four-Step Reading Frame