Mineral Oil

Mineral oil is rated Clear, and the honest version of the case for it is decent rather than overwhelming. Fulton's 1989 assay does contain a mineral oil row — it prints 0–2, which on his own key runs from "not considered significant" into "borderline" — the earlier rabbit-ear work that condemned it was later called a false positive by the man who wrote it, and the human findings everyone quotes are a summary we have not been able to read in full.

What it is

A mixture of saturated hydrocarbons — alkanes and saturated ring structures — distilled from petroleum. The grades used in cosmetics and pharmacy are heavily refined fractions; refining is what removes the aromatic and polycyclic-aromatic content, and it is the whole reason the International Agency for Research on Cancer treats untreated and mildly-treated mineral oils as a different material from highly-refined ones. Formulators use it as an occlusive emollient. It sits on the stratum corneum and slows water loss. It is often described as inert, and in the sense that it is not metabolised by skin that is fair, but "inert" is doing too much work: Rawlings and Lombard's review notes that occlusion itself has biological consequences — measurable moisturisation and a dampened inflammatory response — so the effect is physical in origin without being biologically silent. Unlike most ingredients on any comedogenic list, it is typically used at high percentages rather than trace ones. On a label it is Mineral Oil or Paraffinum Liquidum.

Why we rate it Clear

We rate it Clear at Moderate confidence, and the rating stands. Start with the correction, because it is the most useful thing on this page and because we would not accept it from anyone else. Until 12 July 2026 this page carried four errors, all found by an external fact-check rather than by us. It said Fulton's assay had no mineral oil row — our own reference file, in this repository, contains the row. It said the rabbit-ear literature never condemned mineral oil — earlier rabbit-ear work condemned it explicitly. It said Fulton's Table II scored D&C Red #36 a 5 in mineral oil — Table II, which we had also already transcribed, prints 3. And it omitted a 2019 rabbit assay of liquid paraffin while asserting the evidence pointed uniformly one way. Three of the four contradicted files sitting in our own repository, which is worse than not knowing. The page below is the version that survives checking. What Fulton actually printed. Table I of the 1989 assay — ingredient diluted to roughly 10% in propylene glycol, applied to the inner ear of New Zealand albino rabbits, three rabbits per assay, follicles scored for keratosis — carries the row "Mineral oil", and the entry is 0–2. He tested it. The interesting thing is what 0–2 means on his own published key, which reads: 0–1 "not considered significant", 2–3 "borderline", 4–5 "considered positive". So the range he printed for mineral oil starts below the threshold of significance and ends inside the borderline band. It is a low result. It is not a clean one, and a page that rounds 0–2 down to zero is doing the same thing to Fulton that the comedogenic lists do. His text distinguishes among mineral oil sources — some acceptable, others not — which is what a printed range rather than a printed number is telling you. The history is the opposite of what this page used to claim, and it is a better story. Kligman and Mills, working the rabbit ear in the early 1970s, reported that petrolatums and mineral oils from different sources were uniformly comedogenic. Mineral oil was condemned, by the assay, by name. What happened next is that the assay was retracted by one of its own authors. In his 1996 paper Kligman writes, in as many words: "Using the original model, Kligman and Mills stated that petrolatums and mineral oils from different sources were uniformly comedogenic. With the updated model, it turns out that these were 'false positives'." He names the two defects that produced them — material placed inside the ear canal, a protected site where loose horny squames are retained and mimic comedones, and no biopsy of the untreated opposite ear as a control, without which, he writes, many materials will be falsely labelled comedogenic. That is the same retraction that clears petrolatum, and mineral oil is named in it. So the correct summary is not that mineral oil escaped the rabbit-ear literature. It is that mineral oil was convicted by it and the conviction was later withdrawn by the prosecution. The human-oriented literature is broadly reassuring and we are going to be careful about exactly how much it says. Mills and Kligman built the human comedogenicity model in 1982: the material applied under occlusion to the upper back of people with large follicles for a month, then read by cyanoacrylate follicular biopsy. Their abstract — which is as far as we could get — confirms the model, the one-month occlusion, the biopsy, and the finding that human skin was less sensitive than the rabbit ear. It does not expose a mineral oil row or score, and we are not going to claim it cleared mineral oil when we have not seen it do so. What is repeated everywhere, and what people are actually quoting when they say mineral oil was tested in humans, comes from DiNardo's two-page 2005 commentary in the Journal of Cosmetic Dermatology, "Is mineral oil comedogenic?": that 100% cosmetic-grade mineral oil scored zero in the rabbit ear across five studies, and that formulations containing up to 30% mineral oil tested non-comedogenic in humans. Rawlings and Lombard repeat those figures in 2012. Both are reporting results; neither is a trial report we can read; the underlying testing is largely industry work. Those are reported findings, not findings we verified, and the distinction is the entire point of this site. What is left is a rating held up by four things, none of them individually strong: an exact generic-name match in Fulton's table with a low-to-borderline rather than positive result; a documented retraction of the older assay that had condemned the ingredient; a reassuring but second-hand human record; and a single contrary study that does not survive being looked at. That adds up to Clear at Moderate confidence. It does not add up to certainty, and anyone telling you the evidence here is unusually strong or all points one way has not read it. It is decent by the standards of a field that has no standards.

What the evidence doesn’t tell you

The one directly contrary study is a rabbit assay, and it should be named rather than skipped. Francis and Shojan, "Comedogenicity of Oils," International Journal of Contemporary Medical Research (2019), tested liquid paraffin — mineral oil — in rabbit ears on a 0–3 scale and concluded it was comedogenic. We give it essentially no weight, and here is precisely why, so that a reader who finds it can weigh it themselves. IJCMR appears by name on Beall's list of predatory journals. Submission to publication took 22 days. Four of the 24 rabbits died and none were necropsied. The authors accepted the null hypothesis on four to six ears per group — "Since the p-value in each case is greater than 0.05, we accept the null hypotheses" — which is not what a non-significant result in a sample that size means. Liquid paraffin, their own comparator, produced comedones in three of five ears. There was no untreated ear anywhere in the study, so nothing in it has a baseline. And the paper acknowledges funding from Marico, maker of one of the test articles, beneath a footer declaring "Conflict of Interest: None." It is close to worthless as evidence. It is not nothing as a fact about the literature, and pretending it does not exist would be the same failure in a different direction. The interpretive threshold this page used to lean on, we could not read. The claim that a rabbit-ear score of 0–2 is unlikely to indicate comedogenicity in human skin is usually sourced to Strauss and Jackson's report of the 1989 American Academy of Dermatology symposium on comedogenicity. We could not inspect that paper, so we are not resting the reading of mineral oil's score on it. We use Fulton's own key instead, because we can read it — and on Fulton's key a 2 is borderline, not clean. Fulton himself also demonstrated that mineral oil as a vehicle can move another ingredient's score. His Table II tested D&C Red #36 in four vehicles: 3 in mineral oil, 2 in pentaerythrital tetra capra/caprylate, 1 in propylene glycol, 0 in PEG 400. Mineral oil produced the highest score of the four, and 3 is borderline on his key rather than positive. It is a smaller effect than we previously printed — this page said 5, twice, and the table says 3 — but the direction survives: the vehicle moves the number, and "mineral oil does not clog pores" is not the same claim as "a formula built on mineral oil cannot." The label name does not certify a grade. Cosmetic and pharmaceutical mineral oil is generally highly refined, and in regulated markets the material reaching a cosmetic label is the refined kind. But "Mineral Oil" or "Paraffinum Liquidum" on an ingredient list identifies an ingredient; it is not a declaration that the material meets a USP or BP monograph. Those are two different propositions and this page previously ran them together. Pomade acne cuts more than one way. Plewig, Fulton and Kligman described it in 1970 — closed comedones on the foreheads and temples of Black men using heavy hair pomades daily, with follicular plugging on biopsy. It is the strongest human signal in the vicinity of these products. But the detail usually left out is that the elaborate formulations produced pomade acne more than the simpler preparations such as mineral oil and Vaseline did, which points away from mineral oil rather than at it. We also cannot call it the one place a human population developed comedones under products of this type: Berlin's 1954 report in the AMA Archives of Dermatology and Syphilology, "Acne comedo in children due to paraffin oil applied on the head," is a second and earlier one. We have not been able to access it, so we do not know what it shows or how well it shows it. We are listing it rather than writing as though it were not there. One legacy score we could not resolve. Nguyen, Dang and Maibach's 2007 rabbit-ear paper, which uses a 0–4 scale, lists "paraffin" among the materials tested. We could not obtain the full table, and we cannot tell from what we could see whether that means liquid paraffin — mineral oil — or solid paraffin, which is a different material, or what it scored. That is an open gap in the evidence behind this page and it belongs here rather than in a footnote. Finally, the field-wide caveat, which applies here as everywhere. Draelos and DiNardo's 2006 piece in JAAD is a Current Issues and Opinion article that tested finished products on human upper backs, six subjects per group for four weeks; it is evidence that an ingredient's rabbit score does not transfer to the formula it ends up in, and it is not a same-material rabbit-versus-human validation, which is how it is usually cited. A 2025 review in JAAD Reviews, surveying the field from 1972 onward, finds that no standardised comedogenicity test exists at all and that "non-comedogenic" remains an unregulated label. The evidence on mineral oil leans one way. It is not proof, and the instrument it was measured with is not a good one.

Where you’ll see it

Baby oil, which is close to neat mineral oil with fragrance. Cleansing oils and balms, makeup removers, ointments, lip balms, body and hand creams, and many pharmacy-tier facial moisturisers. It is also used as a pressing agent in powder makeup and as a base in traditional hair pomades. Because it is used for bulk occlusion rather than for a trace effect it is generally present at high percentages — DiNardo's paper discusses human testing of formulations containing up to 30% — so it tends to sit high on an ingredient list rather than at the bottom. On a label, look for Mineral Oil, Paraffinum Liquidum, Liquid Paraffin, Paraffin Oil or White Mineral Oil. These are names for the ingredient, not certificates of its grade: cosmetic mineral oil is generally highly refined, but a label name is not a pharmacopoeial monograph and does not stand in for one. Petrolatum, the semisolid member of the same hydrocarbon family, is a different ingredient with its own entry here.

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. Fulton JE. "Comedogenicity and irritancy of commonly used ingredients in skin care products." J Soc Cosmet Chem 40:321–333 (1989) The primary source behind the 0–5 scale nearly every comedogenic list online still uses, including ours. Ingredient diluted to roughly 10% in propylene glycol, applied to the inner ear of New Zealand albino rabbits, three rabbits per assay, follicles scored for keratosis. Table I contains the row "Mineral oil" and prints 0–2. Fulton's own key reads 0–1 "not considered significant", 2–3 "borderline", 4–5 "considered positive" — so the range straddles the boundary between the two lowest bands and is not a clean zero. His Table II separately shows the vehicle moving a score: D&C Red #36 grades 3 in mineral oil, 2 in pentaerythrital tetra capra/caprylate, 1 in propylene glycol and 0 in PEG 400. This page previously printed that 3 as a 5, twice.
2. Kligman AM. "Petrolatum is not comedogenic in rabbits or humans: a critical reappraisal of the rabbit ear assay and the concept of 'acne cosmetica'." J Soc Cosmet Chem 47:41–48 (1996) The retraction, and the single most important source on this page. Kligman, who co-built the rabbit assay, writes: "Using the original model, Kligman and Mills stated that petrolatums and mineral oils from different sources were uniformly comedogenic. With the updated model, it turns out that these were 'false positives'." He names the method's two defects — material placed in the ear canal, where retained horny squames mimic comedones, and no untreated control ear. It is not a study of mineral oil, and its human arm tested petrolatum, not mineral oil. It is the document that withdraws the older rabbit-ear condemnation of mineral oil, by name, in the words of the author who made it.
3. Kligman AM, Mills OH. "Acne cosmetica." Arch Dermatol 106:843–850 (1972) The original condemnation, cited as the thing that was overturned rather than as support. Rabbit-ear-canal work on a 0–3 scale — a different scale from Fulton's — which reported that petrolatums and mineral oils from different sources were uniformly comedogenic. This is the paper whose method its own author later retracted. Anyone who says the rabbit-ear literature never condemned mineral oil, as this page did until 12 July 2026, has not read it.
4. Mills OH, Kligman AM. "A human model for assessing comedogenic substances." Arch Dermatol 118:903–905 (1982) The paper that established comedogenicity testing on human skin rather than rabbit ears: material applied under occlusion to the upper back of subjects with large follicles for a month, then read by cyanoacrylate follicular biopsy. Cited for the model and for one finding we can confirm from the abstract — human skin was less sensitive than the rabbit ear. We could not obtain the full text, and the abstract does not expose a mineral oil row or score. It is routinely cited as having cleared mineral oil. We have not been able to verify that, and we are not going to repeat it as though we had.
5. DiNardo JC. "Is mineral oil comedogenic?" J Cosmet Dermatol 4:2–3 (2005) The paper that asks our exact question, and the origin of the modern position. A two-page peer-reviewed commentary by an industry toxicologist — not a trial report with a methods section, a participant count or a control arm. It REPORTS that 100% cosmetic-grade mineral oil scored zero in the rabbit ear across five studies and that formulations containing up to 30% mineral oil tested non-comedogenic in humans, and argues mineral oil should be removed from the comedogenic lists. Those are reported findings summarising largely industry testing we have not read. We cite them as his report of the evidence, which is what they are, and not as evidence we have checked.
6. Rawlings AV, Lombard KJ. "A review on the extensive skin benefits of mineral oil." Int J Cosmet Sci 34:511–518 (2012) A peer-reviewed review, and the source of most of what circulates online about mineral oil's safety — including the second-hand summaries of Mills and Kligman and of DiNardo. Cited on this page for two things: it repeats DiNardo's figures, and it notes that occlusion is not biologically silent (moisturisation, a dampened inflammatory response), which is why this page does not describe mineral oil as doing "nothing biological." Its acknowledgements state that it was fully funded by Union Swiss and that Rawlings is a consultant to them; Union Swiss sells a mineral-oil-based product. That does not make it wrong, and we have not found a claim in it that contradicts the primary literature — but we have not been able to check its claims against primary sources we can read, either, and it is not a disinterested source.
7. Francis A, Shojan A. "Comedogenicity of Oils." Int J Contemp Med Res 6(8):H5–H7 (2019) The one study in the record that directly tests mineral oil and finds against it: a rabbit-ear assay on a 0–3 scale that tested liquid paraffin — mineral oil — among five oils and concluded it was comedogenic. Cited so that a reader who finds it knows what it is, and used as evidence for nothing. IJCMR appears by name on Beall's list of predatory journals. Submission to publication took 22 days. Four of its 24 rabbits died and none were necropsied. The authors accept the null hypothesis on four to six ears per group. Liquid paraffin, their own comparator, produced comedones in three of five ears, and no untreated control ear appears anywhere in the study. The paper acknowledges funding from Marico, maker of one of its test articles, beneath a footer declaring "Conflict of Interest: None." It is close to worthless. It is also real, and leaving it out — as this page did — is how a page ends up claiming the evidence is uniform when it is not.
8. Plewig G, Fulton JE, Kligman AM. "Pomade acne." Arch Dermatol 101:580–584 (1970) The human counter-signal, and it is more equivocal than either side usually admits. An observational report: closed comedones on the foreheads and temples of Black men using heavy hair pomades daily, with follicular plugging on biopsy. The pomades were mixtures and no component was isolated. Cited with the detail that is usually dropped — the more elaborate formulations produced pomade acne more than the simpler preparations such as mineral oil and Vaseline did, which is a point partly in mineral oil's favour rather than against it.
9. Berlin C. "Acne comedo in children due to paraffin oil applied on the head." AMA Arch Derm Syphilol 69:683–687 (1954) A second, earlier human report of comedones under a paraffin-oil product applied to the head. We could not access it, so we do not know its size, its method or the strength of its finding, and we make no claim about what it shows. It is listed for one honest reason: it is the reason we cannot describe pomade acne as the only occasion a human population developed comedones under products of this kind, which is what this page previously implied.
10. Nguyen SH, Dang TP, Maibach HI. "Comedogenicity in rabbit: some cosmetic ingredients/vehicles." Cutan Ocul Toxicol 26:287–292 (2007) A rabbit-ear assay on a 0–4 scale — a third scale, different again from Fulton's 0–5 and Kligman's 0–3 — which lists "paraffin" among the materials tested. We could not obtain the full table. We cannot tell whether "paraffin" here means liquid paraffin (mineral oil) or solid paraffin, which is a different material, and we do not know what it scored. Cited as an unresolved gap rather than as evidence in either direction.
11. Strauss JS, Jackson EM. "American Academy of Dermatology Invitational Symposium on Comedogenicity." J Am Acad Dermatol 20:272–277 (1989) The field's own attempt to say what a rabbit-ear number is worth, and the usual source for the rule that a score of 0–2 is unlikely to indicate comedogenicity in human skin. We could not inspect it. This page previously rested its entire reading of mineral oil's score on that threshold; it now rests on Fulton's own printed key instead, which we can read and which calls a 2 "borderline". Listed because a reader will meet the threshold quoted everywhere and should know we did not verify it.
12. Fulton JE, Pay SR, Fulton JE III. "Comedogenicity of current therapeutic products, cosmetics, and ingredients in the rabbit ear." J Am Acad Dermatol 10:96–105 (1984) Fulton's earlier rabbit-ear paper, on the same 0–5 scale, which tested finished acne products as well as ingredients; its headline finding was that the acne products themselves were comedogenic. Cited for context on the assay's history. This page previously claimed it recommended mineral oil as a pressing agent for makeup in place of isopropyl myristate. We could not verify that, and our own legacy-assay reference attributes the isopropyl-myristate substitution recommendation to octyldodecyl stearate, a different material. The claim has been removed.
13. Draelos ZD, DiNardo JC. "A re-evaluation of the comedogenicity concept." J Am Acad Dermatol 54:507–512 (2006) A Current Issues and Opinion piece in JAAD that tested FINISHED PRODUCTS on human upper backs — six subjects per group, four weeks — and found products testing non-comedogenic despite containing ingredients the rabbit assay had scored as comedogenic. Characterised narrowly on purpose: it is evidence that an ingredient's score does not transfer to the formula it ends up in. It is not a same-material rabbit-versus-human validation, which is how it is usually cited, including on this page previously.
14. "Comedogenicity in cosmeceuticals: A review of clinical relevance, regulatory gaps, and future directions." JAAD Reviews (2025) Recent peer-reviewed review of the whole field, covering literature from 1972 to 2025. Finds that no standardised comedogenicity test exists and that "non-comedogenic" is an unregulated label. It is a synthesis of the legacy literature, not a new measurement, and independent of Fulton only in authorship, since the corpus it surveys includes his. Cited for the field-wide caveat, which applies to this page as much as to any other.
15. IARC. "Mineral oils: untreated and mildly-treated oils (Group 1); highly-refined oils (Group 3)." IARC Monographs, Supplement 7 (1987) The formal basis of the grade distinction, and probably the buried root of the fear. Untreated and mildly-treated mineral oils are classified as carcinogenic to humans on the evidence of skin cancers among mulespinners and metal machinists; highly-refined oils — the cosmetic and pharmaceutical grades — are not classifiable. This is an evaluation of CARCINOGENICITY from occupational exposure to a different, less refined material. It is not a comedogenicity finding and it says nothing about pores in either direction. It is cited to mark that boundary, and kept strictly separate from everything else on this page.
16. acne.org comedogenic ingredient list Republishes the legacy assay scores. Not an independent measurement, and neither are CosDNA, Face Reality, Banish or Sofie Pavitt — they are a small number of old papers quoted many times, not several sources agreeing. Listed here for one reason: mineral oil does not appear on their comedogenic list. That is a fact about what the republishers say, and it is worth knowing, but it is not independent corroboration of anything, and this page used to lean on it as though it were.

Others in the same family

Hydrocarbons 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 hydrocarbons as a family ›

• Hydrogenated Polyisobutene Clear Low
• Petrolatum Clear High
• Squalane Clear Moderate

Last reviewed 2026-08-20 · How we decide

‹ Check a full ingredient list