Choosing a balm for a baby, and the ingredients that do not belong near one
Bland occlusives are reasonable on dry patches and the nappy area. Camphor, menthol and essential oils are not, and the oil massage evidence is weaker than claimed.
Most of what an infant's skin needs is to be left alone, kept clean and kept out of contact with urine and faeces. That leaves a narrow set of jobs a balm can usefully do, and a much longer list of materials sold for babies that should not go near one. This page sets out what the neonatal oil trials measured, why the exposure arithmetic differs from an adult's, and which ingredients are a hard no rather than a matter of taste.
A plain, fragrance free occlusive is reasonable on dry patches and as a nappy area barrier. Routine whole body oiling is not supported: in the OBSeRvE trial, 115 term newborns randomised to olive oil, sunflower oil or no oil twice daily for four weeks, both oil groups developed lamellar lipid structure less well than the no oil group. Camphor, menthol, eucalyptus and methyl salicylate have no place on an infant at any concentration.
Why the same film is a bigger dose on a baby
Two independent facts make infant skin a different problem, and neither is about gentleness.
The first is barrier maturity. A term newborn is born with a functioning but unfinished stratum corneum. In vivo imaging work by Stamatas and colleagues found infant stratum corneum roughly 30 percent thinner than an adult's, with corneocytes about 20 percent smaller and a higher water content, and the maturation continues through at least the first year. Surface pH starts close to neutral at birth and falls toward the adult 4.5 to 5.5 band over the first weeks, which matters because the enzymes that build and shed the barrier are pH sensitive. The structural background is in stratum corneum and barrier lipids. In a preterm infant the barrier is dramatically more permeable, with water loss measured several times higher than at term, and functional competence takes two to three weeks of postnatal life or longer.
The second is geometry. A 3.5 kg newborn has a body surface of roughly 0.21 square metres, about 0.06 square metres per kilogram. A 70 kg adult has about 1.8 square metres, roughly 0.026 per kilogram. Coat both from head to foot with the same thickness of the same product and the baby receives somewhere between two and three times the dose per kilogram of body weight. Nothing about the product changes; the denominator does.
Both facts have a documented history behind them. Hexachlorophene bathing of newborns produced neurotoxicity through the skin, boric acid nappy powders caused poisonings, and aniline dyed nappies caused methaemoglobinaemia. None of those materials was meant to be systemic. On an infant, treat a leave-on product as a low but real systemic exposure rather than as something that stays on the surface.
What the newborn oil trials actually found
Olive oil and sunflower oil have been recommended for baby massage and dry skin by midwives and health visitors for decades, largely on tradition. The trials that finally tested it did not support the practice, and the way the results are usually summarised inverts them.
| Study | Participants | What was applied | Result that matters |
|---|---|---|---|
| Danby and colleagues, Pediatric Dermatology, 2013 | 19 | Olive oil to one adult forearm, sunflower seed oil to the other, twice daily, four weeks | Olive oil reduced stratum corneum integrity on tape stripping and produced mild erythema. Sunflower did neither. Adult skin, chosen because neonates cannot be tape stripped |
| OBSeRvE, Acta Dermato-Venereologica, 2016 | 115 | Olive oil, sunflower oil or no oil, twice daily, four weeks, from birth | Hydration improved in both oil arms. Lamellar lipid structure developed better in the no oil arm than in either oil arm. Authors advised caution rather than endorsement |
| BEEP, The Lancet, 2020 | 1394 | Daily emollient for the first year in newborns at high familial risk of eczema, versus standard skin care | No reduction in eczema at age two. More parent reported skin infections in the emollient group |
| PreventADALL, The Lancet, 2020 | 2397 | Oil baths and facial cream from two weeks of age, with or without early food introduction | No reduction in atopic dermatitis at twelve months from the skin intervention |
| Cochrane review, Kelleher and colleagues, 2022 | 33 trials | Pooled skin care interventions in healthy term infants for preventing eczema and food allergy | Probably little or no effect on eczema, probably an increased risk of skin infection, and a signal of possible increased food allergy |
Two things are worth pulling out. The OBSeRvE finding that gets quoted is the hydration improvement, because it is the one that sounds like a benefit. The finding that gets dropped is the lamellar one: the ordered lipid lamellae that constitute the barrier developed better with nothing applied at all. The pilot was not powered for clinical outcomes, and it has not been followed by a definitive trial that overturns it, so it remains the best direct evidence on healthy term newborns. The honest reading is a ranking rather than a recommendation: sunflower beats olive, and no oil arguably beats both for routine use on skin that is not dry.
The second is that the prevention trials answered a related but separate question and answered it clearly. Applying an emollient daily from birth does not prevent eczema in babies at high risk, and it may cost something. That is why a balm is a reasonable thing to reach for on a dry patch and an unreasonable thing to apply as a daily prophylactic ritual.
None of this transfers to the emollient trials in preterm and low birthweight infants in low income settings, where topical sunflower seed oil has been associated with reduced infection and improved survival. That is a different population, a different baseline barrier and a different endpoint. It is not evidence for oiling a healthy term baby in a heated house.
The oleic acid argument, and how far it actually goes
The mechanism usually offered for the olive oil result is its fatty acid profile. Under the Codex and International Olive Council trade standards olive oil runs roughly 55 to 83 percent oleic acid with only about 3.5 to 21 percent linoleic. Conventional sunflower seed oil inverts that at roughly 48 to 74 percent linoleic. Oleic acid is a well characterised penetration enhancer: it partitions into the intercellular lipid lamellae and disorders their packing, which is what a drop in measured stratum corneum integrity looks like. Linoleic acid is the acid the skin esterifies into the acylceramides that hold the lamellae together. The full comparison is at olive oil versus sunflower oil.
Where the argument stops is worth stating plainly. The disordering work is largely in vitro and on adult skin. Olive oils vary widely in composition between cultivars and grades, so "olive oil" is not one material. And in OBSeRvE the sunflower arm also lagged the no oil arm on lamellar development, which a simple oleic acid story does not explain. Something about applying any oil twice daily to a newborn appears to interfere with a process that runs perfectly well unassisted.
There is also an inversion people miss. Most sunflower oil sold for skincare is now the high oleic type, at roughly 75 to 91 percent oleic under Codex CXS 210-1999, because it keeps far longer. That is a profile closer to olive oil than to the oil the trials used. If you buy a sunflower baby oil on the strength of the research, check the grade: the trade off is oxidative stability against the fatty acid profile the argument was built on.
The nappy area, where a barrier balm genuinely earns its place
Nappy dermatitis is an irritant contact dermatitis with a well described mechanism. Occlusion overhydrates the stratum corneum and weakens it mechanically. Urease from faecal bacteria liberates ammonia from urea, raising local pH. Faecal proteases and lipases are more active at that raised pH, and they attack overhydrated skin. Friction from the nappy finishes the job. Every step of that chain is interrupted by putting a water repellent film between the skin and the mixture.
This is the one infant application where an anhydrous product does something no amount of careful washing does. Under the US over the counter skin protectant monograph at 21 CFR part 347, zinc oxide is permitted at 1 to 25 percent and white petrolatum at 30 to 100 percent, which tells you the levels that regulators consider effective and safe for that use. In the UK and EU there is no equivalent monograph: a plain barrier balm sold with cosmetic wording is a cosmetic, and the same product sold to treat nappy rash is likely a medicine. If you make and sell, the line is set out in cosmetic versus drug claims.
Thin and frequent beats thick and occasional. A thin film reapplied at every change keeps continuous coverage, comes off with an ordinary wipe, and lets you see the skin. A thick paste has to be scrubbed off to inspect the rash, and that scrubbing is itself a cause of the rash. At a change, wipe away the soiled surface only, leave the clean residue in place, and top it up.
Two further points. Petrolatum is the least reactive occlusive available, which is why it is the standard vehicle in patch testing, and a defensible default for inflamed skin. A heavy oil rich balm also ruins cloth nappies by making the fibres hydrophobic, so a family using cloth needs a liner. A worked barrier formula, with and without zinc, is at nappy rash balm. What fixes more rashes than any product is frequent changing and time out of the nappy.
The hard limits: camphor, menthol, eucalyptus and wintergreen
These are not a matter of dilution. They are a category of material that should not be applied to an infant, and the reasoning differs for each.
Camphor is absorbed readily through skin and mucosa, and small ingested doses cause seizures in young children. US over the counter rules cap camphor at 11 percent in topical products, and camphorated oil, a 20 percent preparation once kept in domestic medicine cabinets, was taken off the market after paediatric poisonings. A jar of camphorated balm within reach of a toddler is a bigger problem than the film on the skin.
Menthol and 1,8-cineole, the main constituent of eucalyptus oil, are associated with reflex apnoea, laryngospasm and bronchospasm when applied near the face of an infant or small child. The mechanism is a reflex response to a strong trigeminal stimulus in an airway with a small diameter, so it does not scale down with concentration in a comfortable way.
Methyl salicylate, sold as oil of wintergreen or sweet birch, is a salicylate at very high concentration. One teaspoon, about 5 mL of the near pure oil, contains roughly 7 g of salicylate, which is comparable to around twenty one adult aspirin tablets of 325 mg. Fatal ingestions in small children are documented. There is no formulation of it that belongs in a household with a crawling baby.
Never apply a camphor and menthol vapour rub under or inside a child's nose, at any age. The cooling is a TRPM8 sensation, not decongestion, and airway work published in Chest in 2009 found such a rub increased mucin secretion and slowed ciliary beat frequency in ferret trachea, with a case report of respiratory distress in a toddler after application under the nose. If you use a chest rub at all, it goes on the chest, on an age appropriate formulation, and the age bands are in chest rub balm.
The concentrations, monograph levels and toxicity numbers for all of these sit in menthol, camphor and other actives. The same restrictions drive the ingredient rules in lip balm for children, where swallowing is an additional route.
Essential oils, fragrance, and the case for leaving both out
Fragrance is the commonest cause of allergic contact dermatitis from cosmetics, and essential oils are fragrance with a better story attached. Lavender, chamomile, tea tree and citrus oils all contain identified sensitisers, and the EU list of fragrance allergens that must be named individually on a label was expanded from 26 substances to roughly 80 by Regulation (EU) 2023/1545, which is a fair indication of how the evidence has moved. The materials and their thresholds are covered in fragrance allergens.
Three things stack against essential oils on an infant specifically. Dermal limits for them are derived from adult data, so the per kilogram exposure is two to three times what the limit assumed. Sensitisation is cumulative and lifelong, and the only lever anyone has over it is reducing early exposure. And a baby cannot tell you that something stings. Set against that, the benefit is a smell that the adult applying it enjoys. The dilution arithmetic and the material by material limits are at essential oils in balms, and the answer for infants is zero rather than a small number.
Botanical extracts and infused oils fall the same way, for a duller reason: a wide, variable allergen set you cannot characterise, in exchange for nothing measurable on infant skin. Watch the label wording too. "Unscented" can mean a masking fragrance was added to cover a raw material odour; "fragrance free" should mean no perfuming material was added at all. How to read the rest of the list is in reading a balm label.
Nut oils on broken skin, and where the evidence genuinely stops
This is the question parents ask most often and the one with the least satisfying answer. In 2003 Lack and colleagues published a case control study in the New England Journal of Medicine finding an association between the use of skin preparations containing peanut (arachis) oil in infancy and later peanut allergy. It fed the dual allergen exposure hypothesis: sensitisation through inflamed skin, tolerance through the gut.
The evidence has not resolved since. The association is confounded, because those preparations were typically used on skin that was already eczematous, and eczema is itself the strongest predictor of food allergy. The oral half of the hypothesis was confirmed by the early introduction trials. The topical half has never been tested directly, and the 2022 Cochrane food allergy signal is a caution rather than a finding.
What that supports in practice is modest and specific. Refining removes most of the protein from a seed or nut oil, and food law treats fully refined oils differently for that reason, but cosmetics are not covered by the food allergen regime and a cosmetic grade oil carries no equivalent guarantee. Sweet almond is a genuine tree nut oil; shea is botanically a tree nut but reactions are close to unreported; coconut is a drupe and rarely implicated. The reasoning is worked through at nut oil allergy. On intact skin the risk is theoretical. On broken, weeping or eczematous skin, where a seed oil performs identically, there is no reason to take an unquantified risk.
What a sensible baby balm contains
The specification is short, and shortness is the specification.
- One structurant: beeswax, or petrolatum used alone.
- One or two bland lipids: a refined butter and a stable liquid oil.
- An antioxidant at 0.05 to 0.2 percent, for the oil, not for the skin.
- Nothing else. No fragrance, no essential oil, no botanical extract, no colourant, no active, no "soothing" anything.
Patch test before the first real use even with a list that plain: a rice grain quantity on a coin sized area of the inner forearm, checked at 30 minutes, 24 hours and 48 hours, uncovered. One negative test does not exclude sensitisation, which needs repeat exposure to develop, so keep watching over the first fortnight. The method is at how to patch test a balm. If you are making it yourself, baby massage balm is the worked version with the pour temperature and cure time.
If a baby has diagnosed eczema, the emollient regime comes from the clinician who prescribed it, and their instructions take precedence over anything here. A homemade balm is not a substitute for a prescribed emollient, and what the trial evidence does and does not support is set out in balm for eczema evidence.
Rashes that are not a balm job
Barrier products treat nothing. They keep irritants off skin and slow water loss, and a rash driven by anything else will sit under a balm unchanged or get worse while it is hidden.
| What you see | Why a balm is the wrong answer |
|---|---|
| Beefy red rash involving the skin folds, with small satellite spots at the edge | The pattern associated with candida. Irritant nappy rash typically spares the depths of the folds. Needs an antifungal, not an occlusive |
| Weeping, blistering, pustules or honey coloured crusting | Suggests bacterial infection. Occlusion over it delays recognition |
| No improvement after about three days of good barrier care | The mechanism a balm addresses is not the one operating |
| Fever, poor feeding, floppiness or an unwell baby | Systemic illness. Same day medical assessment, independent of the skin |
| A rash that does not fade under gentle pressure | Non blanching spots need urgent assessment. Nothing topical is relevant |
| Rash in a baby under three months, or one born preterm | Lower threshold for assessment across the board. Ask before applying anything |
| Thick scale on the scalp | Cradle cap is a different process. Oiling it has an equivocal evidence base and treating a named condition is a medicinal claim |
The decision rule, and what this page cannot settle
Use a balm on a baby for two things: a physical barrier in the nappy area, and an occlusive film on a named dry patch. For both, take the shortest ingredient list you can find, fragrance free, applied thinly. Do not oil a healthy newborn head to foot on a schedule: the barrier develops better without it, and the prevention trials found no eczema benefit to buy with the risk.
What this page cannot do is examine a particular baby. The trials above were run on healthy term infants in one or two countries, with endpoints measured over weeks, and none of them tested a finished wax and butter balm; they tested oils. Extrapolating from a bottle of sunflower oil to a balm with a longer residence time on the skin is an assumption, not a result. Where a health visitor, midwife or doctor has given specific instructions for a specific child, those instructions win, and the general limits of what a reference page can tell you are set out in the safety disclaimer.
Frequently asked questions
Is olive oil safe to use on a newborn?
It is not acutely dangerous, but the evidence points away from it. A 2013 forearm study found olive oil reduced stratum corneum integrity and produced mild erythema where sunflower seed oil did not, and the 2016 OBSeRvE trial in 115 newborns found lamellar lipid structure developed better with no oil than with either oil. Sunflower is the better of the two and neither is recommended for routine daily use.
Can I put lavender or chamomile essential oil on my baby's skin?
No. Dermal limits for essential oils are derived from adult data, and an infant receives roughly two to three times the dose per kilogram from the same applied film because of the surface area to body weight ratio. Both oils contain recognised sensitisers, sensitisation is cumulative and lifelong, and a baby cannot report stinging. The benefit is a scent the adult enjoys.
What is the best barrier cream for nappy rash?
The active matters less than the routine. Under the US skin protectant monograph, zinc oxide is permitted at 1 to 25 percent and white petrolatum at 30 to 100 percent, and both work by keeping urine, faeces and friction off the skin. Apply a thin film at every change rather than a thick layer occasionally, and do not scrub the residue off to inspect.
Why should I not put a vapour rub under a baby's nose?
The cooling from menthol is a TRPM8 nerve sensation rather than decongestion, so there is nothing to gain. Work published in Chest in 2009 found a camphor and menthol rub increased mucin secretion and slowed ciliary beat frequency in ferret trachea, and a case report describes respiratory distress in a toddler after application under the nose. Camphor is also readily absorbed.
Are nut oils dangerous on a baby with eczema?
The honest answer is unresolved. A 2003 case control study associated peanut oil skin preparations in infancy with later peanut allergy, but it is confounded because those preparations were used on already inflamed skin. No trial has tested the topical question directly. On intact skin the risk is theoretical; on broken or eczematous skin a seed oil performs identically, so there is no reason to take it.
How long should I try a balm before seeing a doctor?
About three days of consistent barrier care is a reasonable trial for straightforward nappy rash or a dry patch. Seek advice sooner if there is weeping, blistering, pustules, honey coloured crusting, satellite spots at the edge of a fold, a rash that does not fade under pressure, or if the baby is feverish, feeding poorly or under three months old.
Sources and further reading
- Cooke A, Cork MJ, Victor S, Campbell M, Danby S, Chittock J, Lavender T, Olive oil, sunflower oil or no oil for baby dry skin or massage: a pilot, assessor-blinded, randomized controlled trial (the Oil in Baby SkincaRE, OBSeRvE, study), Acta Dermato-Venereologica 96(3), 2016.
- Danby SG, AlEnezi T, Sultan A, Lavender T, Chittock J, Brown K, Cork MJ, Effect of olive and sunflower seed oil on the adult skin barrier: implications for neonatal skin care, Pediatric Dermatology, 2013.
- Chalmers JR and colleagues, Daily emollient during infancy for prevention of eczema: the BEEP randomised controlled trial, The Lancet, 2020.
- Kelleher MM and colleagues, Skin care interventions in infants for preventing eczema and food allergy, Cochrane Database of Systematic Reviews, 2022.
- US Food and Drug Administration, 21 CFR Part 347, Skin protectant drug products for over-the-counter human use, eCFR.
- National Institute for Health and Care Excellence, Postnatal care, NG194, NICE, 2021.
- Lack G, Fox D, Northstone K, Golding J, Factors associated with the development of peanut allergy in childhood, New England Journal of Medicine, 2003.
Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.