Kawakawa balm: what is in it, and how thin the evidence is

Published 6 August 2026Zane Hitchcox, editor

This is an unusual case. Kawakawa is not an obscure plant with no science behind it: a research group at the University of Auckland has spent years running it through mass spectrometers and feeding it to volunteers. The science is careful. It is simply pointed somewhere other than skin, and the gap between what has been measured and what a balm implies is worth mapping precisely.

The plant, and the two names it answers to

Kawakawa is Piper excelsum, a shrub or small tree in the pepper family, Piperaceae, endemic to New Zealand and common from Te Paki south to about Ōkārito, North Canterbury and Banks Peninsula; its conservation status is Not Threatened. The New Zealand Plant Conservation Network lists Macropiper excelsum as a synonym: the species was moved out of Macropiper and back into Piper, and both names remain in circulation.¹

That taxonomic detail has a practical consequence on the back of a jar. The European Union's official glossary of cosmetic ingredient names — the INCI list, established by Commission Decision (EU) 2019/701 — carries the plant under the superseded binomial, as entries 13710 and 13711: MACROPIPER EXCELSUM LEAF EXTRACT and MACROPIPER EXCELSUM LEAF POWDER.² A reader searching an ingredient list for the modern name will not find it. That is a routine feature of INCI nomenclature rather than an evasion — ingredient names freeze at the point of listing while botany moves on — but it is the kind of thing that makes an ingredient list harder to read than it needs to be.

The chemistry is well described, and it varies wildly between batches

A 2022 analysis in Nutrients from the University of Auckland's Liggins Institute put aqueous and methanolic extracts of kawakawa leaf through LC-MS/MS and identified 64 compounds across five classes: phenylpropanoids, lignans, flavonoids, alkaloids and amides. Eight were quantified absolutely. The phenylpropanoids include myristicin and elemicin; the amides include pellitorine and fagaramide; the lignans include yangambin isomers and diayangambin; the flavonoid vitexin and the alkaloid dopamine are also present, dopamine reportedly for the first time in the Piperaceae. Alkaloids and amides together formed the largest cluster in the molecular network.³ An earlier compositional study in the Journal of Ethnopharmacology had converged on the same short list — diayangambin, elemicin, myristicin, plus unidentified lignans and amides.⁴

The more consequential finding is the variation. Commercially available dried leaf ran high in phenylpropanoids and flavonoids; field-collected fresh samples ran richer in amides, alkaloids and lignans. The published figures are blunt: myristicin at 5.28 µmol per gram of dry leaf in one commercial source against 0.88 µmol/g in one fresh field sample, and pellitorine at 43.2 µmol/g in that fresh sample against 0.99 µmol/g in the commercial one — more than fortyfold.³ No rule in New Zealand or EU cosmetic law requires a botanical infusion to be standardised, so two jars labelled identically may differ in their headline compounds by an order of magnitude, and nothing on either jar would reveal it.

Everything tested in a person was swallowed

As of 6 August 2026 a PubMed search for the term Piper excelsum returns six records; the older name Macropiper excelsum returns five. Combining either binomial with skin, topical, dermatitis, eczema, wound, balm or ointment returns nothing at all. Running those same terms against the common name returns exactly one record, and it is a 2025 commentary on medical imagery in neonatal intensive care, which matches because its authors are at He Rau Kawakawa, the University of Otago's school of pharmacy. There is no trial of kawakawa on skin to read.

The human work that does exist is a coherent programme about ingestion. A 2022 crossover study in Nutrients gave 250 mL of kawakawa tea before a high-glycaemic breakfast: a six-man pilot found lower plasma glucose and insulin at 60 minutes on the high dose, and a 30-person follow-up found lower insulin only.⁵ A 2024 paper in Molecular Nutrition & Food Research used urine and plasma from those two studies to show that kawakawa compounds are absorbed and undergo phase 1 and phase 2 metabolism, identifying 26 urinary metabolites.⁶ A 2024 analysis in Food Science & Nutrition went back to the same blood and reported shifts in nine circulating microRNAs and reduced gene expression of IL-8 and IL-6.⁷

That last paper is the strongest anti-inflammatory signal in the human literature, and it is worth being exact about what it is: transcripts in blood, after a drink, in 26 healthy people, over a few hours. Its own authors open by noting that no human study of kawakawa's anti-inflammatory effects had been undertaken. Nothing was applied to anybody's skin, no symptom was scored, and a change in IL-6 messenger RNA in circulating cells is several inferential steps from a red patch on a forearm getting better. The signal is also not internally consistent: the same paper reports that plasma IL-6 — the protein, not the transcript — rose significantly over 120 minutes in the kawakawa arm, which is the opposite direction to the gene expression result, and the authors do not reconcile the two.⁷

The rest is cell culture. A 2015 Planta Medica paper isolated twelve amides from kawakawa fruit and tested them for cytotoxicity, with piperdardine most active against HT-29 colon cancer cells at an IC₅₀ of 14 µM.⁸ A 2017 Phytochemistry paper characterised piperine analogues from the same plant and measured glucose uptake in Caco-2 intestinal cells, where piperdardine raised uptake by 83 ± 18% at 100 µM.⁹ Neither is a result about skin, and cytotoxicity in a cancer line is not a desirable property in something rubbed on a face.

The base is doing work the plant is being credited for

A balm sold as kawakawa balm is, structurally, an infused oil set with wax: leaf steeped in a carrier oil, strained, thickened with beeswax to a solid that melts at skin temperature. That base is not inert packaging. A 2023 narrative review in the Journal of Cosmetic Dermatology describes beeswax in skincare as an occlusive that forms a semi-occlusive film reducing transepidermal water loss, and as an emollient; the review found five clinical studies of topical beeswax in total — three in animals, two in humans — and concluded that further studies are warranted; two of its authors were affiliated with a beeswax-based skincare company at the time.¹⁰ A 2017 study in AAPS PharmSciTech put beeswax-based nanoparticles into a gel-cream and measured a simultaneous fall in transepidermal water loss and rise in stratum corneum water content after 28 days.¹¹ The evidence for beeswax specifically is thinner than the confidence with which it is usually asserted, but occlusion as a mechanism is one of the few things in skincare not seriously disputed, and it does not require a botanical.

How much of the observed benefit a vehicle can account for is not hypothetical. The best-designed trial of any New Zealand botanical on inflamed skin tested a 3% kānuka oil cream against its own vehicle in 80 adults with self-reported moderate-to-severe eczema, across 11 community pharmacies, applied twice daily for six weeks. Patient-Oriented Eczema Measure scores fell from 18.4 to 6.8 in the kānuka arm — and from 18.7 to 9.8 in the vehicle arm.¹² The control cream, containing no active botanical at all, delivered about three-quarters of the total improvement. The between-group difference was −3.1 points, with a 95% confidence interval running from −6.0 to −0.2, and the published minimal clinically important difference for that scale is 3.4 points.¹³ The trial's result was statistically significant and sat, at its point estimate, just below the threshold at which a patient would be expected to notice.

This trial is about kānuka, not kawakawa, and nothing in it transfers to a different plant. It is cited here for the vehicle arm — the number a reader most needs and the one no marketing page prints — and as a working demonstration of what a real test of a New Zealand botanical on skin looks like: two arms, a defined concentration, a validated instrument, a pre-specified primary outcome and a trial registration lodged before recruitment. Kawakawa has never been through it. The ceiling on what any of this can achieve is set by what a topical product can physically reach, and an occlusive film works precisely because it stays on top.

Safety: mostly assumed, partly measured, and measured orally

The compounds that attract toxicological attention are the phenylpropanoids. Myristicin and elemicin are alkenylbenzenes, the same structural family as safrole and methyleugenol, which are genotoxic and carcinogenic in rodents. A 2022 review in Foods by scientists at Germany's Federal Institute for Risk Assessment set out how incomplete the data are: myristicin forms DNA adducts in liver but has never been tested for mutation induction in vivo; there is no published literature on elemicin's mutagenic potential or subchronic toxicity at all; and the two-year combined toxicity and carcinogenicity studies the authors call mandatory for evaluating long-term effects exist for neither compound. Myristicin did produce centrilobular hepatocyte hypertrophy and necrosis in 90-day rodent studies.¹⁴ That review covers dietary exposure and does not address skin; no dermal data for either compound appear in it.

Against that, the exposure from kawakawa itself looks small. The Nutrients group calculated that a cup of kawakawa tea delivers under 1% of the myristicin and elemicin concentrations associated with toxicity.³ A 2019 safety evaluation fed kawakawa tea to Sprague Dawley rats at doses equivalent to one to four cups a day for 14 and 28 days, found no adverse effects on body weight or food intake beyond minor variations in organ weights and laboratory parameters, and concluded that consumption could be considered safe within the conditions of that study.⁴ Both findings are about drinking a water infusion. A wax balm is a fat-phase extract left in contact with skin — a different route, solvent and duration — and has not been evaluated.

Two more specific hazards deserve separating from vague ones. The first is the wax. Propolis — bee resin — is a documented contact allergen, in the European baseline patch test series since 2019. A 2025 study in Contact Dermatitis patch-tested dermatitis patients at two centres, finding 16 of 257 positive in Genova (6.2%) and 3 of 329 in Malmö (0.9%), and states plainly that propolis can be present as an impurity in beeswax.¹⁵ A person who reacts to a beeswax balm may be reacting to the beeswax.

The second is photosensitivity, where the honest answer is narrower than the usual warning. Phototoxicity from botanical oils is a furocoumarin problem: the IFRA standard that governs it restricts expressed bergamot, bitter orange, grapefruit, lemon and lime oils, plus angelica root, cumin, parsley leaf and rue.¹⁶ No furocoumarin appears among the 64 compounds identified in the Nutrients characterisation,³ and no phototoxicity testing of Piper excelsum has been published in either direction, so the concern has been neither demonstrated nor excluded. A balm containing kawakawa plus a citrus essential oil carries the citrus oil's risk regardless.

Rongoā Māori is a different kind of claim, and this page cannot settle it

Kawakawa's standing does not come from any of the papers above. It is a plant of long-standing food, medicinal and cultural importance to Māori — the framing used in the Journal of Ethnopharmacology paper⁴ — and rongoā Māori, the traditional healing system in which it sits, is not a historical curiosity. The Accident Compensation Corporation operates a Rongoā Māori Service available to people with an approved claim, which places it inside New Zealand's funded rehabilitation system rather than alongside it.¹⁹ The kānuka eczema trial above opens by situating its own intervention in that tradition.¹²

It is not, however, the same category of claim as a trial result, and the two should not be laundered into each other. A randomised trial is an answer to a narrow question — did this preparation, at this concentration, beat this control, on this scale, in these people — and it has nothing to say about meaning, relationship, practice or transmission. Mātauranga Māori carries its own standards of evidence, its own lines of authority and its own custodians, and a reference site that evaluates claims exclusively by controlled trials is not equipped to assess it on those terms and does not attempt to. This page reports one thing only: what the trial literature contains about kawakawa on skin, which is nothing. That is a statement about a literature, not a verdict on a practice, and the distinction is not a courtesy — it is the limit of what the method used here can produce.

The practical consequence is narrow. A jar of balm bought in 2026 is a commercial cosmetic sold under cosmetic law, and the questions that apply to it are the ones that apply to any commercial cosmetic: what is in it, at what concentration, and what has been tested. The answers are, respectively: an infused oil in beeswax, unstated, and nothing topical. Where kawakawa sits relative to the other native plants that turn up on New Zealand labels is covered in the survey of New Zealand botanicals in skincare, and the pattern there is much the same — strong traditional record, decent phytochemistry, almost no clinical work.