Comedogenic ratings: where the 0-5 scale came from and why it cannot predict your skin

The 0-5 comedogenic scale comes from rabbit ear tests on neat ingredients. What it really measures, why the online charts disagree, and what to do instead.

Comedogenic ratings are the most confidently quoted numbers in skincare and among the least well supported. A single digit attaches to an oil, travels across a thousand pages, and ends up deciding what someone puts on their face. The scale is not fraudulent and it did come from real laboratory work. It simply does not carry the information people use it for.

Where the scale actually came from

The provenance is traceable, which is more than can be said for the charts. Almost nobody who quotes a number has followed it back to the paper it came from.

The rabbit external ear was adopted because its follicles plug quickly and reproducibly. A material was applied repeatedly over roughly two weeks, the tissue was examined, and the degree of follicular keratosis was graded. That grading is the origin of the 0 to 5 scale, where 0 means no observed plugging and 5 means severe.

None of this was bad science for its purpose. The assay was designed to screen raw materials so formulators could avoid the worst offenders, at a time when several widely used esters were causing real problems. It was never designed to rank oils for consumers, and it was never validated as a predictor for an individual person.

What the assay actually measured

Four things are true of the classic tests, and all four limit what the numbers can tell you.

• Materials were usually applied neat or at high concentration. A finished balm may contain the same material at 5 or 10 percent.
• Application was repeated and often occluded, on a site chosen precisely because it plugs easily. The site is the rabbit ear canal and the inner surface of the pinna, not facial skin of any species, and the assay was intentionally over-sensitive so that it would not miss anything.
• The endpoint was histological follicular keratosis, not a visible spot. Microscopic plugging in an animal model is a proxy for a clinical outcome, not the outcome itself.
• Sample identity was rarely pinned down. "Coconut oil" is not one material. Refined, virgin, fractionated and hydrogenated versions differ substantially, as do batches from different origins.

Five critiques, each with a source

The objections below are not internet scepticism. Every one of them is in the peer-reviewed literature, and two come from the people who built the assay in the first place.

1. The rabbit is more sensitive than the human, by design. The assay's own authors wrote that "the rabbit model is more sensitive than the human", and that substances weakly comedogenic in the rabbit "are probably safe for human use with the possible exception of acne-prone persons" (Mills and Kligman, Archives of Dermatology 1982, PMID 7138047). A high rabbit score was intended as a flag for further work, not as a verdict.
2. Correlation with human experience was absent. Frank, Journal of the American Academy of Dermatology 1982;6:373 (PMID 6461674), wrote of "the absence of correlation with experience in the human". Fulton himself conceded in 1984 that the rabbit ear "is not an ideal animal model but is the best we have", which is a claim about what was available, not about what it predicts.
3. Ingredients are not products. Draelos and DiNardo tested finished formulations on subjects with prominent follicular orifices and concluded that "finished products using comedogenic ingredients are not necessarily comedogenic" (Journal of the American Academy of Dermatology 2006;54:507, PMID 16488305). That single sentence undoes most of what ingredient charts are used for. It has been in print for twenty years and has changed almost nothing.
4. There is no standard protocol and no oversight. A 2025 review in JAAD Reviews (Starzyk and colleagues, 6:78-83) puts it directly: "a major concern is the lack of standardized testing and regulatory oversight, allowing companies to freely label products as 'noncomedogenic' regardless of their actual potential to cause acne".
5. The claim is not regulated. The US Food and Drug Administration maintains no list of approved cosmetic claims and requires no substantiation before one is printed. Its published position on the analogous term is that "there are no Federal standards or definitions that govern the use of the term 'hypoallergenic'". Be precise about this: there is an FDA page on hypoallergenic, and no FDA page naming non-comedogenic, so the accurate statement is that non-comedogenic falls under the same general regime rather than that the agency has ruled on it specifically.

Why the online charts disagree

Compare three comedogenic charts and you will find the same oil rated differently on each, often by two points. There are structural reasons for this, not just sloppiness.

Charts merge rabbit and human data with different endpoints and concentrations, then present them on one scale as if they were commensurable. Many entries have no traceable source at all and appear to have been inferred from fatty acid composition or copied with a small edit. Materials are named at the wrong level of specificity, so a single row covers a family with wide variation. And a number, once written down, gets repeated with more confidence than it was originally offered with.

The chart, with the caveat attached

Below are the ratings usually quoted for balm materials, alongside how each behaves on skin. The right-hand columns are the ones to formulate with.

Look at the last two rows together. Isopropyl myristate is one of the highest rated materials on the scale and one of the driest-feeling on skin, while cocoa butter is rated the same as coconut oil despite behaving nothing like it. Whatever the scale measures, it is not greasiness, occlusivity or anything a user perceives. The properties you can feel are set out in balm texture science, and the material-by-material comparison sits in carrier oils compared with the numbers in the oil comparison table.

Beeswax is untested, not low

One row says "untested" rather than a number. It is the clearest example on the page of a rating that was never measured.

Beeswax is usually listed at 0 to 2. Follow that back and there is nothing at the end of it. The Cosmetic Ingredient Review re-review packet published in 2025, covering beeswax and the plant waxes, does not contain the words "comedo", "comedogenic" or "acne" anywhere, including in the earlier 1984 and 2005 reports it reproduces in full. A literature search returns no comedogenicity study of beeswax either. It is not that the assay was run and beeswax passed. It appears never to have been run.

What the safety record does contain is a different set of endpoints. Beeswax at 100 percent was non-irritating in a reconstructed human skin assay, and non-sensitising in a murine local lymph node assay at concentrations up to 10 percent, and the panel's conclusion is that these waxes are "safe as used in cosmetics under present practices of concentration and use". Those are real findings and they matter. None of them is a measurement of follicular plugging. Non-irritating and non-sensitising are simply different properties from non-comedogenic.

So the honest entry for beeswax is a blank, and this page prints a blank. Generalise that habit: for a lot of naturally derived materials the truthful cell is "untested", and a chart with no blanks anywhere in it is telling you something about the chart rather than about the ingredients.

Coconut oil, the case everyone argues about

Coconut oil generates more disagreement than the rest of the chart put together. It is commonly rated 4. Plenty of people use it on the face with no trouble, plenty break out within a fortnight, and both experiences are real. The rating explains neither, because it was produced under conditions that resemble nobody's routine: near-neat material, repeated occluded application, on a site selected for sensitivity. Context is more useful. Coconut oil at 8 percent of a lip balm is a different proposition from coconut oil on the cheeks nightly, and fractionated coconut oil is a different material again, being the C8 to C10 fraction with the lauric acid removed. The formulation detail is in coconut oil in balms.

What matters more than the number

If you want to predict whether a product will suit you, these variables all outrank the rating.

What to do instead

1. Rank by feel, not by chart. If you are congestion-prone, favour materials that leave little residue: jojoba, squalane, light esters. This is a heuristic, not a rule, but it is at least based on something you can perceive.
2. Change one product at a time. Nothing can be learned from a routine that changed in three places in the same week.
3. Run a two-week half-face trial. New product on one side only, everything else identical. Two weeks, because comedones take time to become visible. Photograph both sides in the same light at the start and the end.
4. Use a forearm patch only for irritation. Two days detects stinging and redness. It does not detect follicular congestion: forearm skin is not facial skin.
5. Consider the format. A cleansing balm emulsified and rinsed away in ninety seconds carries much less risk than the same oils left on overnight.

The petrolatum question

Petrolatum is the standing counter-example. It is strongly occlusive and generally rated low for comedogenicity, while several light, dry esters rate high. If occlusivity caused comedogenicity, that pattern would be reversed. See petrolatum and mineral oil, the wider argument at natural versus petroleum, and the definitions in occlusive, emollient and humectant.

How to read a product claim

Given all of the above, treat the word on the front of the pack as a statement of intent. Some manufacturers do commission human comedogenicity testing and can produce the report if asked. Most select ingredients with low quoted ratings and print the word. That is not dishonest, but it should not be read as evidence. An ingredient list you can interpret tells you far more than the claim above it, and reading a balm label covers how.

The judgement

Use comedogenic ratings the way you would use a rumour from a source who was once reliable: as a hint about where to look, never as a conclusion. If a material is rated 4 or 5, that is a reason to introduce it carefully and watch, not a reason to avoid it forever. If a material is rated 0, that is not permission to slather it on and blame something else when your skin objects. Your own two-week half-face trial, run properly on one variable, is worth more than every chart on the internet, because it is the only test performed on the skin you actually have.

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