Propolis Benefits: What the Research Actually Shows

Bees settled the question of what propolis is good for long ago: they build with it, seal with it and defend the hive with it. What it does for people is a much younger question, and the honest headline from three decades of research is where that research happened — a systematic survey of the 1990–2018 literature, its search closing in March 2018, found the field dominated by cell dishes and rodents: 790 rats and 420 mice against 150 human participants. Underneath sits a genuinely interesting structural problem the field itself keeps naming: propolis is not one substance. This page walks through that problem, the laboratory record on its own terms, the human trials that exist — mostly small and mostly in the mouth — how the regulator read the evidence, and the one human effect documented firmly enough to carry numbers: contact allergy.

What is propolis?

The glue of the hive. Honeybees collect resins from plants and work them into propolis, a material they use to build, repair and protect the hive; it comprises resins, waxes, essential oils, pollen, flavonoids and minerals. Fuller background is on the propolis page; the common UK consumer format, an alcohol extract sold as drops, is described in the propolis tincture guide.

Why is propolis so hard to study?

Because the word names a category, not a compound. The foundational review of propolis chemistry (Bankova, 2005) sorts propolis into distinct chemical types according to which plants the bees collected resin from, and states the consequence plainly: reliable standardisation criteria are needed, and generally accepted criteria “do not yet exist”. Even for the best-studied type — the temperate-zone “poplar” propolis relevant to Europe — that review could propose only bulk markers (total flavones and flavonols, flavanones, total phenolics) as proxies for activity, and concluded that standardising the other types remained undone.

Seventeen years later the field’s stated position was unchanged. A 2022 review (Kasote, Bankova and Viljoen, Phytochemistry Reviews) reports over 800 compounds across the propolis types, composition varying with region and season, and calls this variability a source of “several challenges” to standardisation and quality control. The trial literature says the same: a 2019 systematic search (Braakhuis, Nutrients) notes that a sample’s exact composition varies between hives, locations and seasons, and that the different classifications and origins make it difficult to extrapolate health-related claims from one sample to another product. This is why the standardisation point is not pedantry — a finding about one sample of propolis is, strictly, a finding about that sample.

Plate I · IllustrationRecord

What do laboratory studies show?

A great deal of activity, in dishes and in animals — and it is worth being clear about what that kind of evidence is. A cell assay shows what a compound can do to cells at doses the experimenter chooses; it cannot show what a swallowed extract does in a living body, with digestion, dilution and metabolism in between. The standard pharmacology review (Sforcin and Bankova, Journal of Ethnopharmacology, 2011) records that propolis mechanisms have been widely investigated in vitro and in animal models, then states the field’s gap in its own words: there is a lack of clinical research on propolis, and clinical investigation is needed to evaluate its potential.

The proportions were counted in 2019. Braakhuis’s systematic search of the 1990–2018 literature located 63 publications: 24 purely cell-based assays, 27 animal studies (790 rats and 420 mice), two combined cell-and-animal studies, four compound-identification studies — and six human trials. The vast majority of papers were cell-based assays applying increasing doses of propolis constituents; the rodents outnumber the human participants roughly eight to one. The popular “natural antibiotic” reputation lives here: the antibacterial literature is overwhelmingly in-vitro, and no clinical antibiotic use is established. That survey’s search closed in March 2018; trials published since are not in these counts.

Fig. I · The literature, to scaleDiagram

63 publications · 1990–2018Six human trials

Cell-based assays24

Animal studies790 rats and 420 mice27

Combined cell-and-animal studies2

Compound-identification studies4

Human trials6

Participants

790

420

150

RatsMiceHumans

RodentsHuman participants

What has actually been tested in people?

Less than the laboratory literature implies, and mostly in the mouth. The six human trials that survey located enrolled 150 people between them: 118 healthy and 32 with diabetes. That is the survey’s human sample, not the whole human literature — its search was English-language only and closed in March 2018, and the mouthwash cluster below is not inside that count.

The densest human cluster is dental, and topical: propolis-based mouthwashes for plaque and gingival inflammation. A 2020 systematic review (Halboub and colleagues, BMC Oral Health) found nine randomised trials totalling 333 participants. Only two were at low risk of bias; the majority were at high risk. The trials also resisted the reviewers’ main tool: meta-analysis works by averaging trials that asked a similar enough question, and these did not — different formulations, durations of 5 days to 6 weeks, inconsistent outcome measures. Samples were small, and every trial’s follow-up fell short of the American Dental Association’s recommended six-month evaluation period.

A 2025 update (Ballouk and colleagues, BDJ Open) found the picture essentially unchanged: ten randomised trials, 453 patients, four at high risk of bias, follow-ups of two weeks to three months again all short of the six-month standard, and again no meta-analysis possible for heterogeneity. Despite positive individual results, the reviewers’ verdict is blunt: the studies “lack certainty and their power of evidence is low with no agreed gold standards”.

Two things sharpen the reading. The cluster is topical — mouthwash held in the mouth — so it is not evidence about swallowing propolis as a tincture or capsule. And the reason neither review could pool its trials is the standardisation problem surfacing inside the trial literature: the products tested were not the same substance. A later 2025 review did pool plaque and gingival outcomes from seven trials of alcohol-free propolis mouthwash.

How did the regulator read the evidence?

The regulator hit the same wall the chemists had been naming — which makes the regulatory file a data point about the science, not just a legal list. In 2010 EFSA assessed the propolis health claims submitted under the EU claims regulation — thirteen claim IDs covering propolis and its flavonoids, spanning immune support, throat and airway comfort, gut and liver claims, blood fluidity and oral health (EFSA Journal 2010;8(10):1810). None survived, and the ground for rejection, recorded verbatim against thirteen of the fifteen propolis entries in the Great Britain register, is the composition problem itself: “this food is not sufficiently characterised for a scientific assessment of this claimed effect and the claim could not therefore be substantiated”. The obstacle Bankova stated in 2005 is the one the regulator recorded in 2010. Two antioxidant claims were rejected separately, and on the merits: assessed against protection of DNA, proteins and lipids from oxidative damage, the claimed effect was not substantiated for this food.

The Great Britain nutrition and health claims register — the operative list for what may lawfully be said about food in England, Scotland and Wales — holds fifteen propolis entries, all non-authorised. Set against the popular claims, the record sorts like this:

One row deserves a pause: “Supports oral health. Maintains health of teeth and gums” (ID 1645) — the one indication with a genuine trial cluster is itself a rejected claim.

Fig. II · Fifteen entries, one columnDiagram

Great Britain nutrition and health claims register · propolis

The operative list for what may lawfully be said about food in England, Scotland and Wales.

Propolis entries15

Non-authorised15

AuthorisedNone — and only authorised claims may be used. For propolis, that set is empty.

IDEntry, as this page words itStatus

Ground recorded verbatim · thirteen of the fifteen“this food is not sufficiently characterised for a scientific assessment of this claimed effect and the claim could not therefore be substantiated”

1242Soothes the throat, supports the airwaysNon-authorised1244“Nature’s antibiotic”Non-authorised1245“Supports the immune system and the body’s defence”Also covers “Aide à combattre les refroidissements” — helps to fight coldsNon-authorised1246Soothes the throat, supports the airwaysNon-authorised1248“Supports the immune system and the body’s defence”Non-authorised1645“Supports oral health. Maintains health of teeth and gums”The one indication with a genuine trial cluster is itself a rejected claimNon-authorised3184Soothes the throat, supports the airwaysNon-authorised3526Healthy blood flowNon-authorised3527Liver supportNon-authorised3799“Supports the immune system and the body’s defence”Non-authorised

Three of the fifteen · this page names twelve IDs

Entry not named on this pageNon-authorisedEntry not named on this pageNon-authorisedEntry not named on this pageNon-authorised

Rejected separately, on the merits · two entriesAssessed by EFSA on the merits; not substantiated.

1243Protection of DNA, proteins and lipids from oxidative damageNon-authorised3797Protection of DNA, proteins and lipids from oxidative damageNon-authorised

What is actually well documented in humans?

Contact allergy. The firmest human literature on propolis concerns the people it sensitises. European patch-test series report that 1.2 to 6.6% of patients tested for dermatitis react to propolis. The main sensitisers in poplar-type propolis are caffeic acid esters — 3-methyl-2-butenyl caffeate and phenylethyl caffeate chief among them, with benzyl and geranyl caffeates and several cinnamic compounds also implicated. The record has an occupational profile — beekeepers can develop the allergy after years of handling hives; in one occupational series, onset averaged 9.5 years (range 0.1–35) — and a consumer one: propolis cross-reacts with balsam of Peru, with which it shares 13 constituents, and sensitisation has been rising, including in children, as propolis has spread through cosmetics and “natural” remedies. The quiet irony: the one human effect of propolis documented firmly enough to put a number on is the unwanted one.

What would have to change for the evidence to firm up?

The sources give their own answer, consistently, across twenty years. Bankova in 2005: reliable standardisation criteria are needed and do not yet exist, with much work still to be done beyond the poplar type. Sforcin and Bankova in 2011: clinical investigation is needed. Kasote, Bankova and Viljoen in 2022: the variability still presents challenges to standardisation and quality control. Ballouk and colleagues in 2025: future trials should use typified — chemically characterised — propolis. The regulator’s recorded threshold is the same: a food must be sufficiently characterised before any claimed effect can even be assessed. Each of those is one prescription: a defined, characterised material first.

So — what is propolis good for, on the published record? For bees: building, repairing and defending the hive. For people: a rich laboratory literature that clinical work had barely followed by the survey’s 2018 close; a clinical literature that has grown since, mostly small trials measuring blood markers; a topical dental cluster its own reviewers call weak; and one firmly documented human effect — the allergy. The next chapter is already prescribed, by the chemists and the regulator alike: characterise the material first. Its shelf-mates are read the same way across the evidence section.