Manuka Honey Benefits: What the Research Shows

Manuka’s best evidence is real, specific — and topical. As a sterile, licensed dressing applied to wounds in clinical care, honey healed partial-thickness burns about 4.7 days faster than conventional dressings in the Cochrane review, on high-quality evidence, with a more mixed record across other wound types. Eaten from a jar, manuka is a different subject, and a far thinner one: three randomised trials of swallowing it — 163 patients in radiation oesophagitis, 131 and 106 in radiation mucositis — were each null on their primary endpoint; the cough evidence belongs to ordinary honey; and the stomach story failed its only clinical test. Two different questions wear one name; this page separates them.

Plate I · IllustrationRecord

Fig. I · Two questions, one nameDiagram

On the skina sterile, licensed medical product, in clinical care

• Cochrane 2015 (Jull and colleagues)26 randomised trials · 3,011 participantsThe anchor evidence for topical honey
• Partial-thickness burnshealed about 4.7 days faster than conventional dressings2 trials · 992 people · high-quality evidence
• Infected surgical woundshealed more quickly than antiseptic washes followed by gauze1 trial · 50 people · moderate-quality evidence
• Leg ulcers, diabetic foot ulcers, chronic wounds“does not form a robust basis for decision making”Low or very low-quality evidence

The jar is not the dressing, and the dressing’s evidence is not the jar’s.

On the spoona food, eaten from a jar

• RTOG 1012 (Fogh and colleagues, 2017)a randomised trial of swallowing manuka — 163 patientsNegative on its primary endpoint
• Al Somal and colleagues, 1994the claim that manuka fights H. pylori traces hereA dish finding on biopsy isolates
• McGovern and colleagues, 199912 patients — a tablespoon four times daily for two weeksNone of the twelve eradicated
• GB nutrition and health claims registerno authorised claim for honey of any kindEight honey claims assessed, all refused

Why do hospitals use manuka honey?

Because a specific wound product exists, and it is not the jar from the shelf. In NHS wound care, medical honey means sterile dressings impregnated with honey and applied directly to the wound surface; NHS patient guidance (Oxford Health) describes a short prescribed course for wounds with high bacterial levels or local infection. The honey in these products is processed to a standard food honey never meets: Simon and colleagues’ 2009 review records a certified honey licensed as a medical product for professional wound care in Europe and Australia, with each batch gamma-irradiated to inactivate Clostridium spores. Even Carter and colleagues’ sympathetic 2016 review describes the clinically used material as a standardised product “registered as a wound care product with appropriate medical regulatory bodies” — and that review’s entire scope is topical; it makes no efficacy case for eating manuka at all.

The same NHS leaflet draws the line itself, warning against substituting supermarket “table honey”: in laboratory tests, table honey is less effective at reducing bacteria in wounds and carries a wider range of microbial spores than sterilised honey. The popular inference — hospitals use manuka, therefore eating it works — fails in both directions. The jar is not the dressing, and the dressing’s evidence is not the jar’s.

What did the Cochrane review of honey dressings find?

Two genuine positives, and a mostly low-quality field around them. The 2015 Cochrane review by Jull and colleagues — 26 randomised trials, 3,011 participants — is the anchor evidence for topical honey. Its strongest finding: honey healed partial-thickness burns about 4.7 days faster than conventional dressings (2 trials, 992 people), at high-quality evidence — Cochrane’s top grade, meaning further research is very unlikely to change the estimate. Its second positive rests on far less: a single trial of 50 people, at moderate-quality evidence, found infected post-operative wounds healed more quickly under honey than under antiseptic washes followed by gauze. Both findings are topical, in clinical settings, and many of the included trials used honeys of varying types, not only manuka.

Beyond those two, the field thins out. For venous leg ulcers, diabetic foot ulcers, minor acute wounds and mixed chronic wounds, the review rated the evidence low or very low quality — in its words, it “does not form a robust basis for decision making”. The review also records a comparison honey lost: early excision and grafting healed partial- and full-thickness burns 13.6 days faster than honey followed by grafting as necessary (one trial, 50 people, low-quality evidence). Its overall conclusion is guarded: “It is difficult to draw overall conclusions regarding the effects of honey as a topical treatment for wounds due to the heterogeneous nature of the patient populations and comparators studied and the mostly low quality of the evidence.” That is the strongest half of the manuka story speaking.

What does the laboratory evidence show?

The chemistry behind the numbers is real, and it lives in a dish. Mavric and colleagues identified methylglyoxal (MGO) in 2008 as the dominant antibacterial constituent of manuka: six New Zealand samples carried 38–761 mg per kilogram, up to 100-fold more than conventional honeys, and MGO at concentrations of about 1.1 mM inhibited Escherichia coli and Staphylococcus aureus in culture. The wider laboratory literature is broad: Carter and colleagues tabulate inhibition of dozens of bacterial species, plus drug-resistant variants of several — MRSA, VRE and ESBL-producing clinical isolates among them — with no resistance to honey yet identified.

An in-vitro result is a different kind of fact from a trial, and the difference is worth having clearly: a dish assay describes what a honey does to bacteria in culture, at dish concentrations, not what happens in a person. The clinical relevance of this chemistry runs through the wound products, not the spoon — inhibiting S. aureus in agar is not treating an infection in a person, and no ingestion trial bridges that gap.

Fig. II · The dish and the personDiagram

What the laboratory work measured, and where the measuring stops

In the dishBacteria in culture — measured

• Methylglyoxal (MGO), 2008Mavric and colleagues: the dominant constituent behind manuka’s inhibition of bacteria in culture
• Six New Zealand samples38–761 mg per kilogram — up to100-fold more than conventional honeys
• About 1.1 mM MGOinhibited Escherichia coli and Staphylococcus aureus in culture
• Carter and colleagues, tabulateddozens of bacterial species inhibited, plus drug-resistant variants — MRSA, VRE and ESBL-producing clinical isolates among them; no resistance to honey yet identified

No ingestion trial bridges that gap.

Clinical relevance runs through the wound products, not the spoon

In peopleInfection in a person — never shown

• McGovern and colleagues, 1999 — measured12 patients, a tablespoon four times daily for two weeks — all twelve still tested positive for H. pylori on urea breath testing. A finding of no effect, not an absence of data.
• In-vitro databacteria in culture, not outcomes in people
• No dose-response datain people, at any grade
• No grade ever comparedno human trial has compared grades of eaten manuka against each other, or against ordinary honey

Manuka has been swallowed in three further trials — 163,131 and 106 patients — but those measured radiation oesophagitis and mucositis, and each was null on its primary endpoint. Their record is Fig. I.

Inhibiting S. aureus in agar is not treating an infection in a person.

Is there any evidence for eating manuka honey?

Strikingly little, and what exists is null. The most substantial randomised trial of swallowing manuka this review located is NRG Oncology’s RTOG 1012 (Fogh and colleagues, 2017): 163 lung-cancer patients were randomised between best supportive care, liquid manuka honey, or manuka lozenges taken four times daily, to prevent radiation oesophagitis during chemoradiotherapy. The trial was negative on its primary endpoint, pain on swallowing at four weeks: honey as prescribed “was not superior to best supportive care”.

Two placebo-controlled trials in radiation mucositis landed in the same place. Bardy and colleagues (2012) randomised 131 head-and-neck cancer patients to 20 ml of manuka honey or a golden-syrup placebo four times daily for six weeks, and found “no significant difference between honey and golden syrup in their effects on mucositis”. Hawley and colleagues (2014) randomised 106 head-and-neck cancer patients to swish, hold and swallow 5 ml of manuka honey or a placebo gel four times daily through radiotherapy, and reported “no statistically significant difference between the honey and placebo arms in any of the outcome indicators” — with the honey “not tolerated well by our patients”, dropouts mostly due to nausea.

The best oral evidence for honey is not about manuka at all. Short-term cough relief in upper respiratory infections is the one field with respectable trial evidence: a Cochrane review of honey for acute cough in children (Oduwole and colleagues, 2018 — 6 trials, 899 children) found honey probably reduces cough frequency more than no treatment (2 trials, 154 children) and more than placebo (2 trials, 402 children), both moderate-certainty; a 2021 systematic review across upper respiratory infections (Abuelgasim and colleagues, 14 studies) found honey superior to usual care, its two placebo comparisons inconclusive, and called for placebo-controlled trials. The trials used ordinary honeys; a search of the trial literature found nothing manuka-specific.

The stomach story is older and thinner still, and it is a good lesson in how a laboratory fact travels. The claim that manuka fights Helicobacter pylori traces to a 1994 study by Al Somal and colleagues: H. pylori isolates from gastric biopsies were sensitive in agar to a 20% manuka solution, with growth prevented at 5% over 72 hours — a dish experiment on biopsy isolates. The one clinical test located since is a 12-patient trial (McGovern and colleagues, 1999): a tablespoon of manuka four times daily for two weeks, half the patients also taking omeprazole — and four weeks after treatment, all twelve still tested positive for H. pylori on urea breath testing. Three decades on, no later trial of manuka for H. pylori was located, and no major guideline recommends honey for H. pylori. (One 2016 uncontrolled study of ordinary honey combined with Nigella sativa reported a negative breath test in 8 of 14 patients — a different preparation, no control arm, and no test of manuka.)

How does the whole record stack up?

What do the MGO and UMF numbers on the jar mean?

Concentration and certification, not proven benefit. MGO counts milligrams of methylglyoxal per kilogram of honey — a concentration scale — and UMF is a licensed trade-association grade indexed to it. Neither certifies an outcome: no human trial has compared grades of eaten manuka against each other, or against ordinary honey, and there are no dose-response data in people at any grade. The full grading story — what the numbers measure, New Zealand’s five-attribute export definition, and why more honey has been sold worldwide as manuka than New Zealand produces — is in our guide to raw honey vs manuka.

How regulators judged the evidence

In Great Britain, food may carry only the health claims authorised on the GB nutrition and health claims register — and for honey, that register authorises nothing. Read row by row, the current edition holds eight honey-related entries, every one non-authorised, and no authorised claim for honey of any kind. Manuka, MGO and methylglyoxal do not appear in the register at all.

The refused entries are the shelf pitch, almost word for word, which makes the reasoning worth reading. “Soothing for the throat”: assessed and non-authorised (EFSA Journal 2010;8(2):1484). “Helps support and maintain a healthy digestive system”: non-authorised (EFSA Journal 2011;9(6):2243). Claims for the antioxidants in honey, and for honey helping “natural defences” through “natural antimicrobial action”: non-authorised. The stated reason recurs across the entries: on the evidence assessed, honey “is not sufficiently characterised for a scientific assessment of this claimed effect and the claim could not therefore be substantiated” — the food itself is too variable for any such effect to be pinned to it.

Fig. III · Eight rows, none authorisedDiagram

In Great Britain, food may carry only the health claims authorised on the register — and for honey, that register authorises nothing.

Claim wording, as assessedRegister status

1. “Soothing for the throat”EFSA Journal 2010;8(2):1484Non-authorised
2. “Helps support and maintain a healthy digestive system”EFSA Journal 2011;9(6):2243Non-authorised
3. Claims for one named group of compounds in honeynamed in the page’s own paragraph above — the term is kept off this drawingNon-authorised
4. “Natural defences”, through “natural antimicrobial action”assessed as a claim for honeyNon-authorised
5. Entry not named on this pagecounted in the register’s eight; wording not given hereNon-authorised
6. Entry not named on this pagecounted in the register’s eight; wording not given hereNon-authorised
7. Entry not named on this pagecounted in the register’s eight; wording not given hereNon-authorised
8. Entry not named on this pagecounted in the register’s eight; wording not given hereNon-authorised

Where a manuka entry would sit

Manuka, MGO and methylglyoxal do not appear in the register at all.

No entry — nothing assessed, nothing to stamp

The stated reason, recurring across the entriesHoney “is not sufficiently characterised for a scientific assessment of this claimed effect and the claim could not therefore be substantiated” — the food itself is too variable for any such effect to be pinned to it.

Honey entries 8Non-authorised 8Authorised 0Manuka entries 0

Does the evidence justify the price?

The evidence cannot say, because no study has asked: nothing in the literature cited here addresses price. What a grade certifies is methylglyoxal concentration, not benefit — no human trial has tested whether any grade of eaten manuka produces better outcomes than any other grade, or than ordinary honey.

What about safety?

The same rules as any honey, and grade changes none of them. NHS guidance says not to give honey to babies under 12 months, because honey occasionally contains bacteria that can produce toxins in a baby’s intestines and cause infant botulism; it also notes honey is a sugar, with tooth decay the concern. Bee products can provoke allergy. A high MGO number confers no exemption — and note the closing symmetry: the Clostridium spores medical wound-care honey is gamma-irradiated to destroy are the same genus behind the infant warning on the unsterilised jar.

What eating honey of any kind does and does not have evidence for — the sugar arithmetic included — is the subject of our companion review, is honey good for you; the rest of the honey literature, read the same way, starts at the evidence hub.