The Standard Gene Panel Looks for Four Mutations. Sequencing the Whole Gene Found Nine.

If you have read anything about the science of eczema in the last fifteen years, you have read about filaggrin. It is the protein that helps hold the top layer of skin together and keeps water in, the gene for it is FLG, and a broken copy is the strongest single genetic risk factor anybody has found for atopic dermatitis.

What gets said next, usually, is that filaggrin mutations are rare in people of African ancestry, and that this means filaggrin matters less for them.

The first half is true of the mutations that were looked for. The second half turns out to be wrong, and the reason it is wrong is a methods problem, not a biology problem.

What the standard test actually tests

Most filaggrin genotyping in research and in clinics screens for a short list of loss-of-function variants — R501X, 2282del4, R2447X, S3247X and a few relatives. That list exists because those are the variants that were found first, in European cohorts, and they are common enough there to be worth a cheap targeted assay.

A US longitudinal cohort of 857 children with eczema, followed for 3,684 person-years, ran exactly that panel. One or more of those mutations turned up in 27.5 percent of the white children and 5.8 percent of the African American children.

Almost five times less common. That is the number that became the received wisdom.

But the same paper reported something the received wisdom dropped. Among children who did carry one, the effect on how persistent their eczema was came out almost the same in both groups — odds ratio 0.42 of ever being symptom-free in white children, 0.53 in African American children, with a p value of .62 for the difference between them.

So the mutation was rarer, and just as consequential when present. Nobody had shown that filaggrin mattered less. They had shown that those four variants were rarer.

Then somebody sequenced the whole gene

In 2018 a group did the obvious thing that nobody had done: instead of asking whether African American children carried the European variants, they sequenced FLG exon 3 properly in 262 African American children with eczema.

They found nine unique loss-of-function variants. Overall minor variant frequency 6.30 percent. The three commonest were p.R501X at 1.72 percent, p.S3316X at 1.34 percent, and p.R826X at 0.95 percent.

Look at those numbers next to each other. In Europeans, a handful of variants account for most filaggrin loss of function, which is why a targeted panel works. Here, no variant predominates — the burden is spread across nine of them, most of which no standard assay is looking for.

And the children who carried one were less likely to become symptom free over an average of eight years of follow-up, odds ratio 0.36.

The authors’ closing sentence is the whole point, stated flatly: properly determining filaggrin loss-of-function status requires advanced sequencing techniques.

Which means every study that genotyped a mixed cohort with the four-variant panel and reported that filaggrin does not explain eczema in Black patients was reporting the sensitivity of its own assay.

And a second gene nobody was testing at all

Four years before that, the same group had run whole-exome sequencing on 60 African American children with eczema and looked at filaggrin-2, a separate gene, adjacent in function and almost never genotyped.

They found a premature stop codon and a large deletion, and then tested the specific variants against how persistent the disease was. Children with S2377X had an odds ratio of 0.44 for being free of symptoms; those with H1249R, 0.23.

That was the first identified genetic variation in a skin barrier protein in people of African ancestry with eczema. 2014.

Sixty children in the discovery set is small and the authors say so. But it establishes the shape of the problem: there was a barrier-gene story here the entire time, and it was invisible because the field was running a European lookup table against non-European genomes.

The immune half diverges too

Barrier genetics is one axis. The inflammation running through the barrier is another, and it does not look the same either.

The reviews of this are consistent on the broad strokes. Everybody with atopic dermatitis has strong Th2 activation — that is what makes it atopic dermatitis, and it is why the drugs that block Th2 signaling work across the board. Beyond that, the polarization differs. Asian patients show stronger Th17 and Th22 activation. Black patients show the highest serum IgE of the groups compared, with Th2 and Th22 skewing and comparatively little Th1 or Th17 involvement.

I would treat these as directional rather than settled. They come from immunophenotyping studies with modest numbers, and race is standing in for ancestry, environment and a lot else, which is the substitution this whole section keeps running into. But they are not nothing, because they map onto drug mechanisms. A drug aimed squarely at IL-17 and a drug aimed at IL-13 are not interchangeable if the dominant axis differs between patients, and almost nobody has powered a trial to find out.

What the clinic does with all of this, which is less than you would hope

Here is where the genetics stops being academic.

TARGET-DERM AD is an observational study running across 43 US academic and community centers. In 1,928 patients, moderate-to-severe disease was commonest in non-Hispanic Asian participants at 63 percent, then non-Hispanic Black at 61, then Hispanic at 49, then non-Hispanic white at 48.

The patients with the most severe disease, in other words, were the non-white ones.

The odds of being given an advanced systemic therapy were 0.71 for non-Hispanic Black patients and 0.66 for Hispanic patients compared with non-Hispanic white patients. Both significant at p below .01.

More severe disease, less of the good treatment. This is the identical pattern documented in acne and isotretinoin, in a different condition, with a different drug class, in the same country.

And where the modern drugs have been examined by racial subgroup, they work. A post-hoc analysis of four tralokinumab phase 3 trials — 1,876 patients pooled, of whom 134 were Black and 407 Asian — found improvement against placebo across every subgroup. The Black subgroup is small and the analysis is post hoc, so it cannot rule out a difference. It certainly does not support withholding the drug.

Two findings that pull against each other

Two things, and they pull in opposite directions, which is usually a sign the evidence is real.

The first is that eczema in richly pigmented skin is not simply eczema with different-colored inflammation. The barrier genetics differ, the immune polarization differs, the presentation differs — more follicular, more papular, more likely to leave pigment change behind — and the severity instruments were built around a redness you may not show.

The second is that none of that difference is a reason for less treatment, and the access data says less treatment is what is happening. Every mechanism listed above argues for earlier and harder control, not for watching and waiting.

If you are the parent of a child whose eczema is being managed as mild because it does not look inflamed, the number worth having in your head is the one from that 43-center study: 61 percent of the Black patients in it had moderate-to-severe disease, and their odds of being on the drugs for it were 29 percent lower.

This article is educational and is not a substitute for care from someone who can look at your skin. Anything new, changing, painful or spreading should be assessed rather than researched.