Nut and seed oil allergens: what refining removes and what a label must say

Protein residue levels in refined against cold pressed nut oils, the almond, peanut and sesame position, and how UK, EU and US labelling rules treat each of them.

Almond, hazelnut and macadamia oils turn up in balms constantly, and the same question follows all of them: does someone with a nut allergy have to avoid this? The answer depends on how the oil was pressed, how much protein survived, and which body of law you are reading, because food and cosmetics are regulated on entirely different principles. This page gives the measured residues, the legal position in the UK, EU and US, and how to word a label that is true rather than defensive.

• Refined oils 0.2-2.2 ug protein/mL
• Crude oils 100-300 ug/mL
• Peanut protein limit 0.5 ppm (EU, UK)
• Shea and coconut are not tree nuts

What is actually in a nut oil

A pressed seed oil is triglyceride, not protein. The allergens that matter in tree nuts and peanuts are storage proteins: the 2S albumins, 7S vicilins and 11S legumins that the seed lays down to feed the seedling. They are water soluble and they partition into the press cake, not the oil. What ends up in the bottle is a small amount of protein carried over as suspended particulate and as part of the mucilage, and how much survives depends entirely on what happens after pressing.

Cold pressing is one step: crush, press, filter. Nothing removes protein except the filter, which is not selective for a molecule. Full refining is four or five steps. Degumming removes phospholipids and the material bound to them. Alkali neutralisation moves free fatty acids and hydrophilic contaminants into the soapstock, taking residual protein with them. Bleaching passes the oil through activated clay, which adsorbs polar material. Deodorisation is steam stripping at 180 to 270 C under vacuum, which denatures whatever is left. The result is the grade sold as RBD, refined, bleached and deodorised.

Read that table for its shape, not its decimals. Only peanut is genuinely established: refined oil has been fed to allergic people under controlled conditions and did not provoke reactions, while crude oil did in a minority of the same group. The pattern is assumed to hold for other seeds because the refining chemistry is identical, but assumed is the word. Nobody has run the almond trial.

Why food law treats refined oil differently

Food allergen law works by naming foods that must always be declared, then carving out specific derivatives shown not to carry enough protein to matter. The carve outs are narrow and jurisdiction specific, which is where the confusion starts.

In the US, the Food Allergen Labeling and Consumer Protection Act excludes any highly refined oil derived from a major food allergen, and any ingredient derived from such an oil, from the definition of a major food allergen. Highly refined peanut oil therefore needs no "contains peanut" declaration on an American food label. Cold pressed, expelled or extruded peanut oil is not covered by that exclusion and must be declared.

In the EU and the UK, Regulation (EU) No 1169/2011 takes the opposite approach for peanuts. Annex II lists the fourteen allergens, and the only vegetable oil exempted from the list is fully refined soybean oil and fat. Peanut oil is not exempted, at any refining grade. The same regulation exempts nuts used for making alcoholic distillates, and nothing else in the nut family. So a jar of refined peanut oil sold as food in Britain declares peanut and the same oil sold in Ohio does not, from identical material. Both are correct within their own system.

Cosmetics are not a food allergen regime

This is the point stated wrongly most often, including by suppliers. There is no declarable allergen class for nut or stone fruit oils in cosmetics anywhere: not in the EU, the UK, the US, Canada or Australia. A cosmetic label does not carry "contains almond" the way a food label carries "contains milk", and no regulator expects it to.

What cosmetics law does instead is require the whole formula to be disclosed. Article 19 of Regulation (EC) No 1223/2009, retained in Britain as the UK Cosmetics Regulation, requires a list of ingredients in INCI nomenclature, in descending order of weight down to one percent, headed by the word "Ingredients". The US requirement in 21 CFR 701.3 is the same idea in different words. Because INCI names are binomial, the botanical origin is already on the label: Prunus Amygdalus Dulcis Oil is sweet almond, Corylus Avellana Seed Oil is hazelnut, Macadamia Ternifolia Seed Oil is macadamia, Arachis Hypogaea Oil is peanut. A reader who knows what to look for has everything. A reader who does not is the reason reading a balm label exists.

There are exactly two allergen specific mechanisms in EU and UK cosmetics law, and neither is about nuts as a class. The first is the fragrance allergen list in Annex III, which requires named volatile materials such as limonene and linalool to be declared individually above set thresholds. That list is separate and is covered in fragrance allergens. The second is a restriction, not a declaration: Annex III caps peanut oil, its extracts and derivatives at a maximum peanut protein level of 0.5 ppm, following the Scientific Committee on Consumer Safety opinion that recommended it. That is the only nut specific number in the regulation, it applies to one species, and it is a purity specification your supplier has to meet rather than something you print.

In the US, the Modernization of Cosmetics Regulation Act added registration, listing, safety substantiation and adverse event reporting, and directed FDA to make rules on fragrance allergen labelling. It did not import food allergen declarations into cosmetics. The practical duties are set out in MoCRA duties, and the EU and UK side in selling balms in the UK and EU.

Which balm oils carry a real tree nut association

Botanical family and statutory status do not line up neatly, so separate what is legally a tree nut, what is clinically associated, and what merely sounds like one.

Individual pages carry the formulation detail: sweet almond oil, macadamia oil, argan oil and apricot and peach kernel oil.

Shea and coconut, the two that get mislabelled

Shea butter comes from the kernel of a fruit, so calling it a nut butter is loose but not absurd. For allergen purposes it is not a tree nut: it appears on neither the EU Annex II list nor FDA's tree nut list, and testing of commercial shea for IgE binding soluble proteins has come back essentially empty, which is what you would expect from a fat that is boiled, roasted, pressed and often refined. Reported reactions are vanishingly rare relative to the volume used. That is not the same as zero risk, and unrefined grade A shea carries more residual non-fat material than refined does. The full picture is in shea butter.

Coconut is a palm drupe. It has been listed inconsistently over the years, which is why some American labels still treat it as a tree nut, but coconut allergy is uncommon and does not track tree nut allergy reliably. Someone allergic to cashew has no particular reason to avoid coconut oil unless they have been tested and told otherwise.

Sesame, and why it changed

Sesame is a seed, so it never belonged in the nut conversation, but it belongs in this one. The FASTER Act of 2021 made sesame the ninth major food allergen in the US from 1 January 2023, catching up with the EU, which had listed it since 2003. Prevalence estimates put sesame allergy at roughly 0.2 percent of the US population, comparable to several tree nuts.

Two details matter for a maker. First, the US highly refined oil exclusion applies to sesame too, now that sesame is a listed food, but most sesame oil in commerce is unrefined or toasted rather than highly refined, so the exclusion often does not bite. Second, cosmetic sesame oil, INCI Sesamum Indicum Seed Oil, is usually cold pressed, because that is what buyers of a massage or body oil want. If you use it, use it knowing it is the least refined common carrier in the cupboard.

Cross-reactivity, kept separate from the law

Statutory lists are about labelling. Cross-reactivity is immunology, and the two answer different questions. A customer asking "I react to cashew, is your almond balm a problem" is asking the second one, and no list will answer it.

The reliable clusters are botanical. Cashew and pistachio sit together in Anacardiaceae, walnut and pecan in Juglandaceae, and co-allergy within each pair is high. Almond is a Rosaceae drupe seed and correlates comparatively weakly with the others, which is why it is often tolerated by people who react to cashew. Hazelnut has two stories: severe systemic reactions driven by the storage proteins Cor a 9 and Cor a 14, and mild oral symptoms in birch pollen allergic adults driven by the labile PR-10 protein Cor a 1. The second kind does not usually survive refining or predict a reaction to an oil.

None of this transfers cleanly to skin. Ingestion challenge data tell you about a gut exposure to milligrams of protein. A balm on intact skin is a different route, a different dose and a different mechanism, and the clinically reported cosmetic reactions to nut oils are mostly contact urticaria or contact dermatitis rather than systemic events. Damaged or eczematous skin changes that calculation, because a disrupted barrier is a recognised route of sensitisation. That is a reason to take the question seriously on cracked hands and lips, not a reason to claim a mechanism you cannot demonstrate. If a customer reports stinging or a rash, work through balm stings or burns before assuming the nut oil is the culprit, because fragrance and essential oils are more frequent causes.

What dose does a lip balm actually deliver

Intuition here is usually wrong by several orders of magnitude. Take a lip balm with 30 percent sweet almond oil. The Scientific Committee on Consumer Safety uses 0.057 g per day as the exposure figure for lip care products, which assumes the product is largely ingested.

Two caveats. The cold pressed row applies a peanut figure to almond, because no equivalent almond measurement exists, so treat it as an order of magnitude and nothing finer. And a reference dose is a population statistic: the VITAL doses predict a reaction in about one percent of allergic individuals, a policy choice rather than a threshold of zero. Someone at the extreme tail is not covered by an averaged number.

Wording a label so it is accurate rather than defensive

The instinct when a customer asks about allergies is to add reassuring words to the packaging. Almost every available reassuring word makes the label worse.

The defensible position is a complete INCI list, an optional plain English gloss for the ingredients people ask about, and a written answer ready for the enquiry. Keep the claim reasoning in your product information file or safety substantiation record; safety assessment and the CPSR covers what that file has to hold in Britain and Europe, and labelling cosmetics for the US covers the American panel.

Cross contact in a small workshop

If you pour an almond oil balm and a sunflower oil balm through the same jug, the second contains almond oil. Not a trace: a measurable amount, because a warm jug retains a film of a gram or more. That is the dominant contamination route in a small studio, far ahead of airborne transfer.

Protein residue is not soluble in the oil, it is carried in it, so a solvent rinse that dissolves the fat does not necessarily carry the protein away. What works is what works on any food contact surface: hot water above 60 C, detergent to lift the fat film, mechanical scrubbing, then a rinse. Silicone spatulas, spouted jugs, stick blender shafts and tube filling trays hold the most residue, and porous or scratched plastic is worse than stainless steel or glass. The general regime is in workshop hygiene.

Answering the parent who asks if it is safe

You will get this question, usually about a lip balm for a child, and the honest answer is not the one being hoped for. You cannot assess anyone's allergy. You are not their clinician, you have no idea how they react or at what dose, and telling a worried parent a product is "safe for nut allergies" is outside your competence and, in Britain and Europe, an unsubstantiated claim.

What you can do is give them facts and let them decide with someone qualified. Tell them exactly which oils are in it and their botanical species. Tell them whether those oils are refined or cold pressed, which means asking your supplier and keeping the answer on file. Tell them what else is made on the same equipment. Point out that lip balm is partly ingested, which is why lips are the highest stakes application and why lip balm for children deserves more caution than a hand salve. Suggest they check with the child's allergy team, and suggest patch testing if the team is content for them to try it.

If you want a product you can hand over without any of this conversation, formulate it out. Sunflower, high oleic safflower, jojoba, castor, olive, rice bran, oat and hemp oils carry no tree nut association at all, and a worked formula is in nut free balm. That is a business decision rather than a safety one, and for a market stall selling to strangers it is usually the right one. Nothing on this page is medical advice, and the limits of what a reference site can tell you are set out in the disclaimer.

Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.