Grapeseed Oil

Grapeseed oil is rated Caution, and the number attached to it is not Fulton's — neither his 1989 table nor his 1984 one has a grapeseed row. The 2–3 in circulation is Morris and Kwan's, from a separate 1983 rabbit-ear assay on a 0-to-5 scale, and it is the only measurement of this oil we can find. It sits mid-scale rather than in the insignificant band, and second from the top of the nine vegetable oils that lab tested. The Caution follows that measurement; the popular case for calling grapeseed clear rests on its linoleic content, which no study has connected to this oil in a follicle.

What it is

The oil pressed from the seeds of the wine grape, Vitis vinifera — largely a byproduct of wine and juice production, since the seeds are what remain after pressing. On a label it is Vitis Vinifera (Grape) Seed Oil. Its composition is what the reasoning below turns on, so it needs a composition source rather than an assumption. Garavaglia and colleagues' 2016 review in Nutrition and Metabolic Insights reports linoleic acid as the most abundant fatty acid in cold-pressed grape seed oils, contributing between 66.0% and 75.3% of total fatty acids; the composition it tabulates gives oleic acid 14.3%, palmitic 6.6% and stearic 3.5%. Those figures are for cold-pressed oil. The same review distinguishes filtered from turbid oils and records differences in their minor compounds, so processing moves the material, and a single range should not be read as a specification for every commercial grade. We set the composition out because the rating's reasoning depends on it — not because composition has been shown to determine how an oil behaves in a follicle.

Why we rate it Caution

Start with where the number is not. We rendered Fulton's 1989 Table I and read it row by row, and there is no grapeseed oil entry in it — not under grape, not under Vitis, not anywhere in the paper. Nor is there one in the earlier survey by Fulton, Pay and Fulton, whose Table I we also hold transcribed in full. Fulton is the source of nearly every comedogenic number in circulation, so "grapeseed per Fulton" is the phrasing a reader is likely to have met, and it describes a row that exists in neither of his rabbit-ear papers. The number that does exist belongs to a different laboratory. Morris and Kwan, in the Journal of the Society of Cosmetic Chemists in 1983, ran their own rabbit-ear assay — three rabbits, the material applied in the ear canal, undiluted unless the row says otherwise, fourteen applications, follicular keratosis then graded microscopically on a 0-to-5 scale — and scored Grape Seed Oil 2–3, with no dilution noted against it. Theirs is the only assay on that scale that is not Fulton's: it predates his 1989 paper, and its authors write in print that their finding of noncomedogenic activity for sodium lauryl sulfate "conflicts with the high scores published by Fulton et al." So grapeseed's number is real, it is not Fulton's, and it comes from a lab that was willing to disagree with him. One caution about that number, from inside the same paper. Our record of it notes that while the table prints 2–3, the caption to the paper's Figure 3 has been read as 4. We take the table as the record — it is the formal results matrix, it carries every material in the study, and captions are the likelier place for a labelling or typesetting slip — but we have not adjudicated the two against the original scan, and a reader should know the paper does not speak with one voice here. The two values are not averaged, and the lower one is the one we carry. What the 2–3 is worth cuts against the rating rather than for it. It is not in the 0–1 band, and it is not a clean pass. The descriptive key usually quoted for this scale — 0–1 "not considered significant," 2–3 "borderline," 4–5 "uniformally reproduceable and considered positive" — is Fulton's, printed in a later paper with a different protocol: his material was diluted to roughly 10% in propylene glycol, where Morris and Kwan's was applied neat. So grapeseed's 2–3 falls numerically in the band Fulton later called borderline; it was not labelled borderline by the laboratory that produced it, and we are not going to attribute his terminology to them. Inside Morris and Kwan's own set the comparison is cleaner, because all nine vegetable oils were tested the same undiluted way: olive 0, avocado 0, castor 0–1, corn 0–1, sesame 1, safflower 1–2, peach kernel 2, grape seed 2–3, sweet almond 3. Grapeseed is second from the top of that list. The case for calling it clear rests on composition instead, and there is a published line of evidence behind it. Downing and colleagues proposed in 1986 that a shortage of linoleate in sebum imposes a local essential-fatty-acid deficiency on the follicle and drives the hyperkeratosis that begins a comedone. Letawe and colleagues reported in 1998, from a double-blind, placebo-controlled randomised crossover in patients with mild acne, that topically applied linoleic acid achieved an almost 25% reduction in the overall size of follicular casts and microcomedones over a month, with no change at placebo-treated sites. Both of those tested linoleic acid as the free fatty acid, while most of grapeseed oil's linoleate is esterified in triglycerides, and neither study put the oil on anybody. Nor does the one table that did put oils on follicles carry the inference: we hold its rows, not a compositional analysis of the oils it tested, and a ranking is not a demonstration of what produced it. So the record, stated plainly. One direct measurement of this exact material exists. It is a rabbit's ear, it is unreplicated, it reads 2–3 of 5 undiluted with an unresolved possible 4 in the same paper, and the only oil in that lab's set to score higher is sweet almond, at 3. Against it sits an argument from composition that no study has connected to this oil. That is why the label is Caution and not Clear. It is the same rule the rest of the site runs on: a printed mid-scale score carries Caution, and the neighbours in this very table are rated that way — peach kernel oil at 2, sweet almond at 3, isopropyl lanolate at 3, all Caution. Grapeseed sits between the first two. Rating it Clear would mean setting aside the only measurement of the material in favour of an inference about one of its fatty acids, which is the move this site exists to argue against. The confidence is Low, because one rabbit-ear reading with an unresolved caption against it is not much of a record — but Low confidence is a reason to say so, not a reason to round the number down to zero.

What the evidence doesn’t tell you

The single measurement here is a rabbit's ear, and it is the only comedogenicity measurement of grapeseed oil we can find. We identified no human comedogenicity study of grapeseed oil — no count of comedones on a face, or a back, after using it. Everything the rating leans on beyond Morris and Kwan's 2–3 is inference from the oil's composition, not a test of the oil. Fulton, who has no grapeseed row but whose paper supplies the key everybody quotes, called his own survey "not at all definitive but simply designed to stimulate research" and warned that the model is sensitive enough that "not everything that irritates this model will also irritate human skin." That caution applies symmetrically: it undermines an alarming reading of a 2–3 exactly as much as a reassuring one. Even the one source is not internally unanimous. The table reads 2–3 and the Figure 3 caption has been read as 4; we carry the table, and the discrepancy is unadjudicated rather than resolved. The tested material is underspecified. Morris and Kwan's row says "Grape Seed Oil" and nothing else — no grape variety, no supplier, no extraction method, no refined-versus-cold-pressed grade, no filtration status, no fatty-acid analysis and no oxidation state. Linoleic content varies with variety and processing, and grapeseed is a high-polyunsaturated oil, which means it oxidises readily. Whether an oxidised sample behaves differently on a follicle is not something any study we found has measured, and we are not going to assume it in either direction. The linoleic-acid evidence is bounded and easy to over-read. Letawe measured existing microcomedones shrinking under topical free linoleic acid; that study did not test whether grapeseed oil forms comedones. Downing 1986 is a hypothesis-and-review paper about sebum chemistry, not a comedogenicity test of any oil. Neither is evidence that grapeseed oil is non-comedogenic; both are reasons to expect a high-linoleic oil might behave well, which is a different and weaker thing. Finally, one body of human data does exist and it measures something else. The Cosmetic Ingredient Review's 2017 assessment of 244 plant-derived fatty acid oils includes Vitis Vinifera (Grape) Seed Oil and concludes the group is safe as used in cosmetics. That is a safety conclusion drawn from irritation, sensitisation and toxicology data. It counts no comedones, and safety is not comedogenicity.

Where you’ll see it

On a label it is Vitis Vinifera (Grape) Seed Oil. The Cosmetic Ingredient Review files it among the plant-derived fatty acid oils used on skin for skin-conditioning, occlusive, emollient and moisturising properties, so a reader will meet it as a general-purpose emollient and carrier oil. It has a parallel life as a culinary oil, which is the context most of the published work on it — the Garavaglia review included — was written for.

Sources

Each source says what it actually is. Several widely-cited “sources” for comedogenic ratings are republishing the same 1989 assay, and counting them as independent agreement is how a thin evidence base gets made to look thick.

1. Morris WE, Kwan SC. "Use of the rabbit ear model in evaluating the comedogenic potential of cosmetic ingredients." J Soc Cosmet Chem 34:215–225 (1983) The actual source of grapeseed oil's number, and the only direct measurement of the material we have found. Method: three rabbits, material applied in the ear canal, undiluted unless the row states a dilution, fourteen applications, follicular keratosis graded microscopically on a 0–5 scale. Its Table III prints Grape Seed Oil at 2–3, with no dilution noted, among nine vegetable oils all tested the same way. It is the only 0-to-5 assay we hold that is not Fulton's, and it is genuinely independent of him — earlier than his 1989 paper, and contradicting him in print on sodium lauryl sulfate. Two cautions carried onto the page: it is a rabbit assay, not a human one, and our record notes that the paper's Figure 3 caption has been read as 4 against the table's 2–3, a discrepancy we have not adjudicated against the original scan.
2. Fulton JE. "Comedogenicity and irritancy of commonly used ingredients in skin care products." J Soc Cosmet Chem 40:321–333 (1989) The primary source of nearly every comedogenic score in circulation, and cited here for what it does NOT contain: there is no grapeseed oil row in Table I or anywhere else in the paper. Method: the ingredient at roughly 10% in propylene glycol, applied to the inner ear of New Zealand albino rabbits, three per assay, follicles scored for keratosis on a 0–5 scale. Used on this page only for two things it can support — the descriptive key that says what numbers on that scale meant to him (0–1 "not considered significant," 2–3 "borderline," 4–5 "considered positive"), and his own warnings about the model. It is not the source of grapeseed's number, because it holds no such number, and his key belongs to his protocol rather than to Morris and Kwan's.
3. Fulton JE Jr, Pay SR, Fulton JE 3rd. "Comedogenicity of current therapeutic products, cosmetics, and ingredients in the rabbit ear." J Am Acad Dermatol 10(1):96–105 (1984) The earlier rabbit-ear survey by the same lead author — two rabbits per test, one ear dosed daily for two weeks and the alternate ear the control — whose Table I we hold transcribed in full off the journal scan. Cited here only for an absence: it prints no grapeseed, grape seed or Vitis vinifera row, and gives the material no numeric grade and no prose classification. Its verdict key also differs from the 1989 one (1 to 2 not significant, 3 and above positive), which is a reason never to move a number between the two papers, and not relevant to grapeseed beyond confirming that neither paper measured it.
4. Garavaglia J, Markoski MM, Oliveira A, Marcadenti A. "Grape Seed Oil Compounds: Biological and Chemical Actions for Health." Nutr Metab Insights 9:59–64 (2016) A peer-reviewed review of grape seed oil chemistry and its studied health actions, in a nutrition journal, cited only for composition: linoleic acid as the most abundant fatty acid at 66.0–75.3% of total fatty acids in cold-pressed oils, with oleic 14.3%, palmitic 6.6% and stearic 3.5% in the composition it tabulates, and filtered and turbid oils differing in their minor compounds. It has nothing to say about pores or comedogenicity. It is here because a composition claim needs a composition source, and its figures are for cold-pressed oil rather than a universal specification for every cosmetic grade.
5. Downing DT, Stewart ME, Wertz PW, Strauss JS. "Essential fatty acids and acne." J Am Acad Dermatol 14(2 Pt 1):221–225 (1986) A hypothesis-and-review paper, not an experiment on any oil. It proposes that low linoleate concentration in sebum imposes a local essential-fatty-acid deficiency on the follicular epithelium and induces the hyperkeratosis that starts a comedone. Cited as the reasoning behind the idea that a linoleic-rich oil might be favourable — and characterised as reasoning, not as a comedogenicity measurement, because it is not one.
6. Letawe C, Boone M, Piérard GE. "Digital image analysis of the effect of topically applied linoleic acid on acne microcomedones." Clin Exp Dermatol 23(2):56–58 (1998) A double-blind, placebo-controlled randomised crossover in patients with mild acne: topically applied linoleic acid achieved an almost 25% reduction in the overall size of follicular casts and microcomedones over one month, with no change at placebo-treated sites, read by digital image analysis of cyanoacrylate follicular biopsies. Real human data — about linoleic acid as the free fatty acid, not about grapeseed oil, and about shrinking existing microcomedones rather than about whether an oil forms them. The abstract gives no patient count, so we do not state one. Cited with that boundary stated, because collapsing "linoleic acid helps" into "grapeseed oil is safe" is exactly the move this page is trying not to make.
7. Burnett CL, Fiume MM, Bergfeld WF, Belsito DV, Hill RA, Klaassen CD, Liebler D, Marks JG Jr, Shank RC, Slaga TJ, Snyder PW, Andersen FA. "Safety Assessment of Plant-Derived Fatty Acid Oils." Int J Toxicol 36(3 Suppl):51S–129S (2017) The final published Cosmetic Ingredient Review Expert Panel assessment — not a tentative report or a draft for comment — covering 244 plant-derived fatty acid oils, Vitis Vinifera (Grape) Seed Oil among them, and concluding they are safe as used in cosmetics. Independent of Fulton and of the charts that republish him. Cited for two things only: that grapeseed oil sits inside a reviewed safety assessment, and the cosmetic functions that assessment attributes to this class of oils (skin-conditioning, occlusive, emollient, moisturising). It assesses safety, not comedogenicity; it counts no comedones and says nothing about follicular plugging.

Others in the same family

Oils behave similarly enough that the evidence for one is often wrongly read across to the others. These are its structural relatives, not a guess. What the evidence says about oils as a family ›

• Argan Oil Clear Low
• Avocado Oil Caution Low
• Castor Oil Clear Moderate
• Coconut Oil Clogging Moderate
• Corn Oil Caution Low
• Cottonseed Oil Caution Low
• Evening Primrose Oil Caution Low
• Flaxseed Oil Caution Low

Last reviewed 2026-08-01 · How we decide

‹ Check a full ingredient list