Evening primrose and borage oil: buying gamma linolenic acid, and what it costs

Evening primrose gives 8 to 10 percent GLA and borage 20 to 24 percent, with iodine values of 145 to 165. Use at 2 to 10 percent, plus the borage alkaloids.

Nobody buys evening primrose or borage oil for its skin feel. Both are thin, pale, faintly odorous seed oils that a cheaper base would out-perform on every sensory measure. They are bought for one molecule, gamma linolenic acid, and the interesting questions are how much of it you get per gram, how fast it dies in the jar, and what borage brings with it that the label does not mention.

• INCI: Oenothera biennis oil
• INCI: Borago officinalis seed oil
• GLA 8-10% vs 20-24%
• Shelf life 6-12 months
• Use 2-10%

What gamma linolenic acid is, and which seeds carry it

GLA is C18:3 n-6: an eighteen carbon chain with three double bonds, the first counted six carbons in from the methyl end. Alpha linolenic acid, the one in hemp, linseed and rosehip, has the identical molecular formula with the double bonds shifted to the n-3 arrangement. The two are not interchangeable, they are not the same nutrient, and a supplier sheet that reports only "linolenic acid" without saying which one is not telling you what you paid for.

Very few seed oils carry meaningful GLA. Evening primrose and borage are the two on general sale, blackcurrant seed sits between them at roughly 15 to 20 percent, and hemp and echium trail at low single figures. Everything else in the balm cupboard has none. That scarcity, not any demonstrated performance, is what sets the price.

The long chain monounsaturates are the practical way to confirm you have borage. Erucic and nervonic acids at a few percent each are unusual, and their absence from a certificate labelled borage is a reason to ask questions. For evening primrose, the European Pharmacopoeia carries monographs for the virgin and the refined oil, which set a fatty acid composition band including a floor on GLA. If you are buying at any scale, that is the document to name in your purchase specification rather than accepting whatever the supplier's own datasheet says.

Why the standard stability index understates both

The oil comparison table ranks carrier oils by bis-allylic index, linoleic percent plus twice linolenic percent, because the carbon sitting between two double bonds is where oxidation starts. That index is normally calculated from alpha linolenic acid, and both of these oils get filed under low or moderate linolenic content because their alpha linolenic is under one percent.

That reading is wrong for the wrong reason. GLA is methylene interrupted exactly as alpha linolenic is, so it carries two bis-allylic sites and should be weighted the same. Count it properly and borage runs at roughly 75 to 88 on the index rather than the 35 to 42 its alpha linolenic content implies, and evening primrose at 81 to 100. Both belong in the top tier of oxidation risk alongside standard sunflower and rosehip oil, not in the middle. The mechanism, and why the bis-allylic carbon matters more than the total double bond count, is set out in rancidity and oxidation.

What a gram of GLA actually costs

Because GLA is the only reason either oil is in the formula, the sensible unit of comparison is oil required per gram of GLA, not price per litre. Working from the midpoints of the bands above, the arithmetic is unflattering to evening primrose.

Both oils sit in the expensive speciality band, several times the cost per kilogram of a bulk emollient such as sunflower or sweet almond oil, and roughly comparable to argan. Between the two, cosmetic grade borage and evening primrose usually price within about a third of each other per kilogram. Since borage delivers two and a half times the GLA per gram, borage is close to half the cost per gram of GLA on almost any pair of quotations you will be offered. If the GLA is the point, borage wins on cost and it is not close. What follows is the reason people still buy evening primrose.

The pyrrolizidine alkaloid question

Borago officinalis produces pyrrolizidine alkaloids, principally amabiline, lycopsamine, intermedine and supinine. They are concentrated in the leaves and flowers rather than the seed, so alkaloid content in the oil is largely a function of how much plant material travelled into the press and how the oil was refined afterwards. Cold pressed unrefined borage oil is where the highest reported figures come from; refined oil is consistently lower.

This matters because the European Food Safety Authority's contaminants panel assessed 1,2-unsaturated pyrrolizidine alkaloids as genotoxic and carcinogenic, deriving a benchmark dose lower confidence limit of 237 micrograms per kilogram of body weight per day for liver tumours in rats and recommending a margin of exposure approach rather than a tolerable intake. The European Union subsequently set maximum levels for the sum of a named panel of alkaloids in specific foods, including herbal infusions and herb-based food supplements, under the contaminants regulation.

None of that is cosmetics law. There is no pyrrolizidine alkaloid limit in the EU Cosmetics Regulation, no equivalent in US or UK cosmetic rules, and every number circulating in the trade is borrowed from food legislation. That does not make the exposure irrelevant: dermal absorption of pyrrolizidine alkaloids has been demonstrated in laboratory work, which is the same reason the comfrey question exists at all and is covered under arnica, comfrey and salve herbs.

Evening primrose has no equivalent issue. That, rather than any composition advantage, is the honest case for paying twice as much per gram of GLA: it removes a contaminant question from the file. Whether that is worth the money depends on whether you can get documented low-alkaloid borage, which most reputable cosmetic suppliers now stock.

Shelf life, and what failure smells like

Treat both as 6 to 12 month oils from the pressing date, not from the day you opened them, and expect the lower end if the bottle lives on a warm shelf. Suppliers routinely print an eighteen month or two year best before because the peroxide value on despatch was compliant. Compliant is not fresh: Codex permits up to 10 milliequivalents per kilogram for a refined oil and 15 for a cold pressed one, while genuinely fresh oil tests at 0.1 to 1. A certificate showing 8 meq/kg is a legal oil that has already used most of its life.

Evening primrose fails with a faint fishy, wet cardboard note, which is the classic signature of oxidised n-6 and n-3 polyunsaturates breaking down into short chain aldehydes. Borage goes the same way with a heavier, more putty or crayon-like character over the top. In a finished balm the smell is easiest to catch on the first application of the day, before your nose has adapted, and it usually arrives before any change in colour or texture. If a batch has turned, the diagnosis and what to do about it are in rancid balm, and the general problem of a product failing inside its printed date is in balm went off before best before.

Buy small. A 100 mL bottle you finish in four months is a better purchase than a litre at a better unit price that spends nine months half empty. Refrigerate the stock bottle, decant into a small amber bottle for bench use, and keep the headspace down: the oxygen in the bottle is a reactant, so a half empty container ages roughly twice as fast as a full one. The wider argument for buying on turnover rather than unit price is in sourcing ingredients.

Where it goes in a batch, and how much

Use 2 to 10 percent by weight of the finished formula. Below 2 percent the GLA delivered is small enough that the ingredient is doing more work on the label than in the product. Above 10 percent you are loading a substantial fraction of the most oxidation-prone material available into a product you want to keep for a year, and the sensory return is nil, because neither oil contributes anything the base oils do not do better.

Sensible defaults: 3 to 5 percent in a face balm, 2 to 3 percent in a lip balm where turnover is fast but exposure to air and light is constant, 5 percent in a small-batch body balm with a six month date, and up to 10 percent only in something made to order. Worked examples of the surrounding formula are in barrier repair face balm and eczema-prone skin balm.

On timing: an anhydrous balm has no cool-down phase in the emulsion sense, because everything passes through the wax melt. What you can do is hold the melt off the heat, let it fall to 50 to 55 C (122 to 131 F), stir the GLA oil in, and pour immediately. That is above the crystallisation onset of every common balm wax, so nothing seizes, and it removes ten or fifteen minutes at 70 C or more from the oil's history. It is a small saving repeated on every batch. Where that sits relative to your pour window is covered in pour temperatures.

Protecting it: antioxidant, packaging and dating

An antioxidant defends the margin, it does not create shelf life. Mixed tocopherols at 0.02 to 0.1 percent of the total formula is the evidence-based window, and more is worse rather than neutral, because alpha tocopherol turns pro-oxidant above its optimum. Deodorised rosemary extract at 0.02 percent alongside it is a reasonable belt and braces. The numbers, and why the 0.5 to 1 percent figure circulating in craft sources is five to fifty times too high, are in vitamin E and antioxidants and rosemary extract.

Packaging does more than the antioxidant does. Oxygen, light and temperature are the three variables, and you control all of them: an opaque or amber container, filled to leave minimal headspace, in a jar or tube rather than a wide open pot the user leaves uncapped. Real-time storage work on comparably unsaturated oils has repeatedly found that the same oil in clear plastic in the light at room temperature fails several times faster than the same oil in amber glass in the dark and cool. The container choices are in the packaging guide.

Date it properly. Record the pressing or supplier batch date of the GLA oil in your batch record, not just the date you made the balm, because the oil's clock started at the press. Set the finished product's date to whichever expires sooner: your standard shelf life, or the remaining life of the shortest-lived input. A balm made in March with borage oil pressed the previous June does not get a two year date. See batch records for what to log and shelf life testing for how to substantiate the date rather than guess it.

What the skin evidence supports, and where it stops

The best documented evidence about these two oils is negative and it is about swallowing them. The Cochrane review of oral evening primrose oil and borage oil for eczema, covering a substantial set of randomised trials, concluded that neither oil taken by mouth is effective for eczema. That is a strong, well conducted piece of work and it is the single most important thing to know before writing any copy about GLA.

Topical evidence is a different and much thinner literature: small, short, mostly measuring instrumental endpoints such as transepidermal water loss or capacitance rather than disease severity, and rarely comparing against a properly matched vehicle. Some of it points in a favourable direction. None of it establishes that a GLA-containing balm outperforms an equally occlusive balm without GLA, which is the comparison that would matter. The mechanistic story, that atopic skin converts linoleic acid to GLA less efficiently and can be supplied downstream, is a plausible hypothesis with mixed support, not a finding.

What you can honestly say is what any competent emollient does: it reduces water loss and softens the stratum corneum. What you cannot say is that it treats, repairs, heals or relieves anything, because that is drug language in the US and a medicinal claim in the UK and EU regardless of the ingredient. The full picture on emollients and atopic skin is in balms and eczema-prone skin, and where the regulatory line sits is in cosmetic versus drug claims.

Is hemp seed oil a fair substitute?

It depends entirely on what you were buying. If you wanted GLA, no. Hemp carries 1 to 4 percent, so matching a 5 percent borage inclusion would need roughly 35 percent hemp, which is not a substitution, it is a different formula with a worse shelf life. Hemp seed oil earns its place on its own merits, a light dry feel and a genuinely useful n-6 to n-3 ratio, not as a cheap route to GLA.

If what you actually wanted was a light, fast-absorbing polyunsaturate with a fashionable name on the label, then hemp does that at a fraction of the cost and with the same short life, and so does standard sunflower for less again. Blackcurrant seed oil is the only real GLA substitute, at roughly 15 to 20 percent GLA plus stearidonic acid, and it is scarcer and dearer than borage in most markets. Whichever you choose, the reason to include it should survive the question of what the balm would feel like without it.

The decision rule

Buy borage if you want GLA and you can get a certificate of analysis with a pyrrolizidine alkaloid sum, a named panel and a stated limit of quantification. Buy evening primrose if you cannot get that documentation, or if the alkaloid conversation is one you would rather not have with a safety assessor or a retailer, and accept that you are paying roughly double per gram of GLA for that simplicity. Buy neither if the honest reason for including it is that GLA reads well on an ingredient list, because at 2 percent in a balm you are adding an expensive, short-lived oil to a product whose performance is set by the wax, the butter and the base oil.

The limits worth stating plainly: no compositional standard exists for either oil, so every number above is a supplier-typical band rather than a regulated specification; the printed best before on the bottle is not a measure of freshness; the clinical evidence that matters most is negative and concerns oral use; and a balm containing GLA is still, in every respect a regulator or a dermatologist would recognise, an occlusive emollient. It reduces water loss. That is a real and useful thing, and it is the whole of what it does.

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