What changes when the skin barrier breaks down

Comparing the stratum corneum to a brick wall is accurate. The corneocytes are the bricks and the lipids filling the space between them are the mortar. The problem is the composition of that mortar.

Online and in product descriptions you will often see the sentence “ceramides, cholesterol and free fatty acids in a 1:1:1 ratio”. We looked for the original source.

We could not find a paper reporting this ratio measured directly in human stratum corneum. 1:1:1 (equimolar) was a convention for preparing artificial model membranes in the laboratory. The relevant papers state that they “prepared synthetic membranes at an equimolar ratio”.

The values actually measured in human skin are different.

In short, ceramides predominate by a wide margin, and the ratio of the three components changes depending on which depth you measure. There is no single value that would deserve the label “golden ratio”.

Natural moisturising factor

There is one more thing besides the lipids: the natural moisturising factor (NMF), which holds water inside the corneocytes. It is made up of about 40% free amino acids, 12% each of PCA and lactate, 7% urea and so on, and accounts for about 10% of corneocyte mass and 20 to 30% of the dry weight of the stratum corneum.

The source of this composition table, however, is secondary material cited in a review, and we could not verify the original paper.

Three things are actually measured, in clinic and in research.

• Transepidermal water loss (TEWL) — the amount of water vapour escaping through the skin. How much the barrier ‘leaks’
• Stratum corneum hydration — how much water the surface layer holds
• Skin surface pH — acidity

But stating a normal value for TEWL requires a great many conditions attached.

And there is a result that runs against the common assumption.

In the same meta-analysis, people over 65 had consistently lower TEWL than those aged 18 to 64. The explanation that “TEWL rises with age” conflicts with this data. In a study measuring 223 nursing home residents (mean age 83.6), forearm TEWL was 10.4, not especially high, while the prevalence of xerosis was 99.1%.

In other words, the problem with aged skin is less ‘leaking a lot’ than ‘low hydration, high pH and slow recovery’.

Nor is TEWL a universal indicator. The conclusion of the nursing home study above was that “the skin barrier parameters measured appear to be of limited diagnostic value”. In another study of 255 older people, TEWL and pH had almost no correlation with blood inflammatory markers, and only hydration showed a negative correlation.

“The normal pH of skin is 5.5” is one of the most common sentences in cosmetic copy. We looked for measured values.

The differences between sites are also large — forehead 4.4, upper eyelid 4.6, chin 5.6. It also varies by about 0.4 to 0.7 within a single day. Newborns start at around 6.0 and come down.

So skin pH cannot be stated as a single number. 5.5 is higher than the measured mean (4.7–4.9) and sits within the range of variation by site and time of day. It can serve as a criterion for choosing a product’s pH, but it is hard to call it a “normal value”.

What happens when pH rises

The mechanism here has been worked out.

• The enzymes that process lipids (β-glucocerebrosidase, acid sphingomyelinase) are active only in an acidic environment
• After the barrier was disrupted with acetone, exposure to a neutral or alkaline buffer delayed the onset of recovery. With an acidic buffer, recovery proceeded normally
• Alkalinisation induces desquamatory enzymes (kallikrein 5)
• The optimal growth pH of Staphylococcus aureus is 7.5

And even water or a mild cleanser raises pH immediately, and returning to normal can take up to six hours.

There is an experiment that damaged the barrier deliberately with a surfactant. 0.5% sodium lauryl sulphate was applied under occlusion to the backs of 30 healthy women for 24 hours, and measurements were taken 24 hours after removal.

What stands out is that the microbiome changes along with everything else. Barrier damage is not only a matter of water escaping; it is also a matter of the composition of the organisms living on the skin changing.

About the next stage — inflammatory mediators being released and immune cells being recruited — there is something we should say honestly.

We could not find a study measuring this cascade directly in humans. In a human skin equivalent (an in vitro tissue) with filaggrin artificially reduced, a substance called TSLP increased 1.6-fold and T-cell migration rose sharply, but that is an in vitro experiment. Even the leading review in this field did not cite a primary study showing that “tape stripping or scratching induces TSLP in humans”.

So this article does not treat “a broken barrier causes inflammation” as an established human fact. The direction looks that way, but the data measured in people is lacking.

Tight junctions — half is the threshold

Beneath the stratum corneum there is a protein (claudin-1) that seals the space between cells. A 2020 study comparing 13 atopic patients with 13 healthy subjects produced an interesting result.

In healthy skin this protein was distributed over a very wide range, 46 to 100% of maximum staining intensity, and yet barrier function was normal. Then from about 50% downwards the barrier deteriorated sharply.

In other words, a small reduction in one protein does not break the barrier down. There is a threshold.

This is the most solid data in the article. It is also the part that differs most from the common assumption.

The stratum corneum of 30 older men (mean age 67) and a young male group (mean age 27) was removed completely, “until the skin was glistening”. That took on average 48 strips in the older group and 45 in the young group. Recovery was then measured at 2 · 6 · 18 · 30 · 72 hours.

Two things can be read from this.

First, even stripped away completely, the stratum corneum comes back within three days. The skin’s capacity to recover is stronger than is commonly assumed.

Second, with age recovery does not fail; it starts later. The older group arrived at the same place as the young group by 72 hours. The peak of the response, though, shifted from 6 hours to 30 hours — about a day.

So we could not find the evidence for the sentence “barrier recovery takes four weeks”. Where it was removed completely by physical means, it normalised within 72 hours. Chemical damage such as from a surfactant may take longer, but the data we could verify tracking that recovery curve over time in humans went only as far as the 24-hour point.

We checked “a 28-day turnover” too

Epidermal turnover has actually been measured. The method was to create pigmentation with ultraviolet light and measure the time until it disappeared.

In 6 Japanese men (mean age 37.3) it was 36.2 ± 6.2 days. We could not find a primary source for 28 days. With a sample of only six this figure is not definitive either, but at least it cannot be confirmed that the value of 28 days came from measurement.

Two rows are not what one would expect.

The belief that acids (AHAs) break down the barrier — when 4% glycolic acid was applied twice a day for three weeks, TEWL did not change. Desmosome breakdown was confined to the outermost layers. This is of course data for one concentration over one period, and we could not find quantitative human data on physical, abrasive exfoliation. Concentration · pH · frequency · whether there is friction all have to be looked at separately.

The belief that a dry environment ruins the barrier — workers spending 12 hours a day in an extreme environment of 1.5% relative humidity actually had lower TEWL, showing an adaptive fall within two weeks.

The two things that are most clearly bad are hot water and surfactants. But look carefully at the conditions in the hot water data — that was immersion in 41.3°C water for 10 minutes. It is not the same as a short everyday shower.

This part produced an unexpected result.

That last row contains an important distinction. Hydration and the barrier are separate axes. A moisturiser holds water in place of the natural moisturising factor, but that does not mean it has repaired the barrier structure itself.

So we try not to state flatly that “applying a ceramide cream regenerates the barrier”. The largest systematic review (3,262 patients) classified urea and glycerol as the better documented, and even the best randomised trial of ceramides was on a scale of 34 subjects · 2 weeks and found no difference between groups on severity.

That does not mean ceramides are bad. It means the strength of the evidence is different.

There is a 2019 study that measured 361 subjects across four cities in Korea and China under the same conditions, including 88 Korean women in Suwon.

In other words the claim that “Korean skin is unusually weak” conflicts with this data. In this comparison at least, the opposite was true.

We also checked the frequently quoted claim that “more than half of Koreans have sensitive skin”. In a 2019 survey administering a 64-item questionnaire to 1,000 Korean men (aged 20–60), 56.1% were the sensitive type and 43.9% the resistant type. The sensitive type predominated in every age band.

This is a self-report questionnaire, however, and not a prevalence based on instrumental measurement or clinical diagnosis. Another survey carried out in China similarly found self-reported “very/somewhat sensitive” at 55 to 60%.

On the seasons

There is a study that measured 89 Korean women every month for 13 months. Temperatures ranged from −1.7°C in January to 26.5°C in August, and relative humidity from 46% in February to 75% in July and August. Scaliness showed a negative correlation with temperature and humidity.

But this study did not measure TEWL. So we could not find data supporting a statement such as “Korean TEWL rises by such-and-such a percentage in winter”.

Taking the data above together, this is what settles out in actual practice.

• We look at hydration and the barrier separately — as the glycerol data shows, hydration can improve while TEWL stays where it was. It is better not to lump everything together as “dehydrated underneath”
• After treatments we advise according to the recovery clock — 72 hours is the benchmark for physical damage, and if you are older you need an extra day at the start. Knowing about that one day is different from not knowing
• The first thing we ask people to cut down on is hot water and washing — in the measurements these are the two most clearly harmful. Even water alone raises pH, and returning to normal can take up to six hours
• We look at use rather than ingredients — what the largest systematic review pointed to was urea and glycerol. More than which ingredient is in it, whether you apply it often enough is in practice the bigger variable

And you need not arrive at the conclusion that your skin is too weak for treatments. What the data above shows is rather that the skin’s capacity to recover is considerable. What matters is not adding the next insult while recovery is still under way.

Reduced to a sentence — the skin barrier recovers better than you would think, and is explained with wrong numbers more often than you would think. More than what you apply, not adding the next insult before recovery has finished is what actually makes the difference.

Written and reviewed by Lee Chi-Hak, MD, medical director of Miso Clinic in Daegu, South Korea. Every study cited above is given together with its design, its size and the limitations the authors themselves recorded, and where we could not find data, we have said that we could not find any.