TL;DR: Most ceramide creams on shelves are sold on the ceramide number alone. Man 1996 and Elias 2014 showed that ceramides without cholesterol and free fatty acids in roughly a 3:1:1 ratio can slow barrier repair, not speed it up. I went back through the original lamellar-body papers and worked out which drugstore creams actually hit the ratio, and which ones are ceramide theatre.
A reader sent me a photo of her bathroom counter and asked me to rank her ceramide creams. There were five of them. Two CeraVes, a Dr Jart Ceramidin, a Skin1004 Madagascar Centella, and a Sephora-house “ceramide complex” product. She had been using all five in rotation for six months and her cheeks were still rough, tight, and faintly red. Her conclusion was that ceramides did not work for her. My conclusion, after reading her ingredient lists, was that four of the five did not actually have a working barrier-repair formulation. They had ceramides on the label and nothing else from the lamellar matrix.
This is one of the most underexplained ideas in skincare. The stratum corneum lipid matrix is not a ceramide layer. It is a co-crystallised lattice of three lipid classes in a specific ratio, and the ratio matters more than the absolute amount of any one component. Man 1996 (PMID: 8618046) is the paper that nailed this down. Elias built on it for two decades after, including the 2014 review on atopic dermatitis (PMID: 24144735) that established the lipid-replacement framework still used in prescription barrier-repair products like EpiCeram. The 3:1:1 ratio of ceramides to cholesterol to free fatty acids is the headline. The detail is more interesting.
What the studies actually show
Man 1996 took tape-stripped barrier-damaged skin in a controlled mouse model and tested barrier recovery using mixtures of the three lipid classes in different ratios. The endpoint was transepidermal water loss recovery over 6 to 24 hours. The result that gets quoted is that a 3:1:1 ceramide-cholesterol-fatty acid mixture restored barrier function fastest. The result that does not get quoted is that mixtures missing one of the three classes were slower to recover than a no-treatment control. In other words, applying ceramides alone to a damaged barrier produced worse recovery than applying nothing. Man’s explanation was that the keratinocyte lamellar bodies cannot assemble a functional lipid lattice from one component. The cell has to make the other two on its own, which is metabolically slow, and during that slow assembly the partial lattice impedes the cell’s own lipid extrusion.
The Elias 2014 review extended this into the atopic dermatitis literature. Atopic skin has roughly 50 percent reduced total ceramides, particularly the ultra-long-chain ceramides EOS and EOH that anchor the lamellar matrix to the corneocyte cell envelope. The Elias paper makes the point that giving atopic patients short-chain ceramides alone can paradoxically worsen the matrix because the short chains do not anchor properly and they crowd out the cell’s own production. The clinical implication, which Elias was explicit about, is that physiological lipid mixtures should match both the ratio and the chain length distribution of native stratum corneum lipids.
Mojumdar 2014 (PMID: 25468346) confirmed the geometric reasoning. The long-periodicity phase of the stratum corneum is a 12-13 nanometre repeat unit of ceramide, cholesterol, and fatty acid arranged in a specific orientation. Cholesterol fills the gaps between ceramide tails. Fatty acids stabilise the lateral packing. Take any one out and the lattice loses its water-impermeable character. This is structural geometry, not marketing.
Spada 2018 (PMID: 30410378) ran a clinical trial of a physiological lipid cream against a standard moisturiser on 20 subjects with dry skin and measured corneometer hydration and TEWL over 4 weeks. The physiological lipid arm produced both faster acute hydration and slower nocturnal rebound dehydration. The standard moisturiser arm had a strong initial bump followed by a return to baseline within 6 hours.
What this means for the cream on your counter
The CeraVe Moisturizing Cream is the reference physiological lipid product at the drugstore tier. It contains ceramides 1, 3, and 6-II, cholesterol, and behenic acid (a fatty acid) in a ratio that internal CeraVe formulation literature describes as 3:1:1. I have not seen the lab assay, but the ingredients are listed in roughly the right proportions and the clinical performance matches a 3:1:1 mixture. The CeraVe PM and the Healing Ointment are similar. The Cetaphil Restoraderm and the EpiCeram prescription are also 3:1:1 designed products.
The Dr Jart Ceramidin Cream has ceramide NP at a high concentration but contains negligible free fatty acid and limited cholesterol. The Skin1004 Madagascar Centella Cream is primarily a centella product with a small ceramide addition for marketing. The Sephora-house ceramide complex products I have inspected tend to list ceramides high on the INCI but have no cholesterol and no fatty acid component, which by Man’s data means they may underperform a plain occlusive.
This does not mean these products are bad. The Skin1004 is a fine centella product. The Dr Jart Ceramidin will hydrate skin. They are not, however, doing the lamellar matrix repair work that Man and Elias described. They are doing surface hydration and occlusion, which is what a tub of petrolatum would do at a fraction of the price.
The honest summary is that almost no consumer is in a position to verify the lipid ratio of their cream. The INCI label does not disclose ratios. The marketing copy will say “ceramide complex” or “ceramide-rich” without specifying that there are no fatty acids. The one heuristic I trust is that if a cream contains all three lipid classes named on the ingredient list (any ceramide, any sterol or cholesterol, and any C16-C22 fatty acid such as behenic, stearic, or palmitic), it has a chance of being a working physiological mixture. If two of the three are absent, it is not.
The contrarian aside
I am sceptical of the precision of 3:1:1. Man’s data showed that 3:1:1 was the fastest recovery in his model. Other published data, including from the Elias group itself, shows that ratios from roughly 2:1:1 to 3:2:1 produce comparable recovery within the noise of the assay. The clinically meaningful claim is not “exactly 3:1:1” but “all three classes present, none dominant over 4x the others.” Most of the consumer literature has hardened this into a magic number, which I think misrepresents the underlying biology.
The other thing the studies do not address is delivery. Ceramides are wax-like at room temperature and have to be solubilised by the formulation chemist. A cream that lists ceramides on the INCI but uses a low-shear emulsion may have ceramide crystals that never reach the stratum corneum. This is invisible to the consumer. CeraVe uses a multivesicular emulsion technology (MVE) that they claim improves delivery, though the published data is mostly their own. I trust the clinical performance more than I trust the technology claim.
What I would tell my past self
I used to buy expensive ceramide creams. The Tatcha Indigo. The La Mer. The early Drunk Elephant Lala Retro. None of them outperformed a 15-dollar tub of CeraVe Moisturizing Cream on TEWL or corneometer hydration in any informal comparison I could run on my own skin. I should have figured this out earlier.
If your barrier is in real trouble, the CeraVe Moisturizing Cream twice a day for 4 weeks is the reference protocol I would build before considering anything else. La Roche-Posay Cicaplast Baume B5 and the Eucerin Aquaphor Healing Ointment are useful add-ons for occlusive sealing at night. The Avene Cicalfate and Bioderma Atoderm Intensive Baume are alternates at slightly higher price points. None of the 60-dollar-plus ceramide creams I have tested outperform these.
If your barrier is mostly intact and you want maintenance, almost any moisturiser will do. The 3:1:1 framework is for damaged barriers. For comfortable, calm, well-functioning skin, you are paying for marketing once you go above the drugstore tier. The barrier damage test gives you a sense of whether you are in repair mode or maintenance mode.
FAQ
Does the ceramide number on the label matter?
Less than the ratio. Ceramide 1, 3, and 6-II are the most common in physiological-lipid products and approximately match the stratum corneum profile, but a 3-ceramide product without cholesterol and fatty acids is not a barrier-repair formulation regardless of which ceramide subtypes are listed.
Can I just add cholesterol and fatty acids on top of my ceramide cream?
Not effectively. The lipid mixture has to be co-formulated for the corneocyte uptake mechanism to work. Layering them as separate products does not reproduce the Man 1996 protocol.
Is CeraVe really the same as expensive ceramide creams?
For barrier repair, in my experience and based on the published comparison data, yes. The Cetaphil Restoraderm and EpiCeram prescription are also physiological-lipid 3:1:1 designs. La Mer and most luxury ceramide products do not use this framework.
How long does barrier repair take?
Man’s data showed 24 to 72 hours for acute tape-stripped recovery in mice. In humans with mild barrier damage, 4 to 8 weeks of consistent physiological lipid use produces measurable TEWL improvement. Severely compromised barriers can take 12 to 16 weeks.
Will a ceramide cream help with eczema?
Yes, with caveats. The Elias 2014 review supports physiological lipid mixtures as adjuncts in atopic dermatitis. They do not replace prescription topical anti-inflammatories during a flare but they reduce flare frequency between episodes.
Related Elelaf tools
Sources
- Man MQ, Feingold KR, Thornfeldt CR, Elias PM. Optimization of physiological lipid mixtures for barrier repair. J Invest Dermatol. 1996;106(5):1096-1101. PMID: 8618046
- Elias PM. Lipid abnormalities and lipid-based repair strategies in atopic dermatitis. Biochim Biophys Acta. 2014;1841(3):323-330. PMID: 24144735
- Coderch L, Lopez O, de la Maza A, Parra JL. Ceramides and skin function. Am J Clin Dermatol. 2003;4(2):107-129. PMID: 12553851
- Mojumdar EH, Helder RW, Gooris GS, Bouwstra JA. Skin lipids: localization of ceramide and fatty acid in the unit cell of the long periodicity phase. Biophys J. 2014;107(11):2670-2679. PMID: 25468346
- Spada F, Barnes TM, Greive KA. Skin hydration is significantly increased by a cream formulated to mimic the skin’s own natural moisturizing systems. Clin Cosmet Investig Dermatol. 2018;11:491-497. PMID: 30410378