Ceramides are the largest group of lipids in the mortar that seals the outer layer of skin, and eczema-prone skin carries a measurably different ceramide profile rather than simply less of everything. The difference is one of chain length: more very short chains, fewer of the very long ones. Emollients that supply ceramides alongside cholesterol and free fatty acids are built on the barrier-repair work below. The evidence is worth reading precisely, because it is far stronger on the mechanism, and on emollients in general, than on any single product.
What the lipid matrix is made of
The stratum corneum is the outermost layer of skin: flattened cells set in a lipid mixture, the mortar between the bricks. Three lipid classes make it up. Ceramides, cholesterol and free fatty acids are "the main lipid classes in SC" (Janssens et al, 2012).
Ceramide is a family, not a molecule. Janssens and colleagues measured 12 subclasses in human stratum corneum, and the letters on a cosmetic label describe their chemistry. The first letter is the attached fatty acid: N for non-hydroxy, A for alpha-hydroxy, EO for an esterified omega-hydroxy chain. The second is the sphingoid base: P for phytosphingosine, S for sphingosine, H for 6-hydroxysphingosine. Ceramide NP is a non-hydroxy fatty acid on a phytosphingosine base. The EO group carries the longest chains in the matrix, and its loss is what tracks with disordered lipid packing.
What changes in eczema skin
Janssens and colleagues sampled non-lesional skin, meaning skin that looked unaffected, in people with atopic eczema. "The level of ceramides with an extreme short chain length is drastically increased", specifically 34-carbon species across the AS, AH, NS and NH subclasses, while the very long chain ceramides of the EO subclass are "significantly reduced". Average ceramide chain length fell by 0.64 carbon atoms. That sounds trivial. The correlation between transepidermal water loss and those chain-length measures was 0.77, and disease severity correlated at 0.56.
Two details change how you read this. The shift was present in skin showing no visible eczema, so the defect is not confined to flaring skin. And it was independent of filaggrin mutations, the best understood genetic cause of a leaky skin barrier, which makes the lipid defect a second mechanism rather than a consequence of the first. The authors read the pattern as reduced elongation of the acyl chains. The skin is making ceramides, but not finishing them.
Why the ratio matters more than any one lipid
The key experiment is nearly thirty years old. Man and colleagues applied lipid mixtures to skin whose barrier had been damaged, then tracked recovery by transepidermal water loss (Man et al, 1996). Applying "one or two of these lipids to perturbed skin delays barrier recovery": an incomplete mixture is worse than nothing. "Only equimolar mixtures allow normal recovery." Repair then gets faster still "as the ratio of any of these ingredients is increased up to 3-fold".
That is where the 3:1:1 figure on product pages comes from: a repair ratio for damaged skin, with equimolar as the baseline that allowed normal recovery, and the dominant lipid can be any of the three. The work was done in murine skin with preliminary confirmation in damaged human skin. Elias, Man and Darmstadt restated the practical point in 2022: "incomplete or unbalanced mixtures aggravate underlying barrier defects". Some substitutions hold. Glycosylceramides and sphingomyelin stand in for ceramides and triglycerides for fatty acids, while phospholipids and cholesterol esters do not.
What the trials actually show
The Cochrane review of emollients and moisturisers for eczema pooled 77 studies covering 6603 participants over a mean of 6.7 weeks (van Zuuren et al, 2017). There were fewer flares with moisturisers (risk ratio 0.40, two studies, 87 participants) and less topical corticosteroid used (mean difference of 9.30 g less, two studies, 222 participants). On the choice between products the same review is blunt: "We did not find reliable evidence that one moisturiser is better than another."
Using an emollient regularly has solid trial support. A ceramide formulation outperforming a plain one does not, at least not at the standard Cochrane applies. The mechanism argues for the physiological mixture; the head-to-head trials are not yet there. That is a reason to choose on texture and tolerability, and to treat a large price premium sceptically.
How to read an ingredient list
Run three checks. First, look for named subclasses: ceramide NP, AP, EOP, EOS, NS or AS. A bare "ceramide" with no letters tells you nothing about which one is in the tub. Second, look for cholesterol. A ceramide-only product, or ceramides with a fatty acid and no cholesterol, is exactly the kind of incomplete mixture that delayed repair in Man's work. Third, look for a fatty acid source. Linoleic acid was one of the three that worked in that study, alongside palmitic and stearic acid.
Phytosphingosine and sphingosine are precursors the skin converts, so they are a reasonable partial substitute. Pseudo-ceramides such as cetyl-PG hydroxyethyl palmitamide imitate the structure without carrying the evidence attached to the named subclasses. A position low in the ingredient list is normal for these lipids and is not by itself a warning sign. Once the ingredient list passes, the format of the tub you take home is a separate decision, and choosing an emollient covers it.
What keeps stripping them out
Barrier lipids are removed by what you wash with, not only by the condition. Hot water and surfactants dissolve them. Chlorinated pool water adds an alkaline pH that works against the skin's acidic surface, which is covered in more detail in swimming with eczema. This is why timing beats quantity. The NHS advises that emollients "are best applied after washing your hands, taking a bath or showering" and can be used "at least 3 or 4 times a day". Replacing lipids in the minutes after they have been stripped is the whole point of that routine.
What ceramides do not do
Ceramides restore structure. They do not suppress the immune response that drives a flare, which is what topical steroids act on. During an active flare a ceramide emollient sits alongside prescribed treatment rather than replacing it. Applied lipids also fill a gap in the matrix rather than changing what the skin manufactures, which is the defect the chain-length data describes.
References
- Man MQ, Feingold KR, Thornfeldt CR, Elias PM. Optimization of physiological lipid mixtures for barrier repair. Journal of Investigative Dermatology, 1996, 106:1096-1101. https://hero.epa.gov/reference/7012768/
- Elias PM, Man MQ, Darmstadt GL. Optimised emollient mixture for skin barrier repair: applications to global child health. Journal of Global Health, 2022. https://jogh.org/2022/jogh-12-03019
- Janssens M, van Smeden J, Gooris GS, et al. Increase in short-chain ceramides correlates with an altered lipid organization and decreased barrier function in atopic eczema patients. Journal of Lipid Research, 2012, 53:2755-2766. https://pmc.ncbi.nlm.nih.gov/articles/PMC3494247/
- van Zuuren EJ, Fedorowicz Z, Christensen R, Lavrijsen A, Arents BWM. Emollients and moisturisers for eczema. Cochrane Database of Systematic Reviews, 2017. https://pmc.ncbi.nlm.nih.gov/articles/PMC6464068/
- NHS. Emollients. https://www.nhs.uk/conditions/emollients/