Clear Skin, Relentless Itch: When a Biologic's Own Numbers Stop Explaining the Symptom
By every visible measure the drug is working. The one thing it hasn't touched is the one thing she actually came in complaining about again.
She has kept a garden every summer for as long as she can remember, and this is the first year in nearly a decade her arms and legs have looked clear enough, by six months into dupilumab, that she's stopped covering them while she works outside. Her EASI score has dropped from 24 to 3 — close to complete skin clearance by any visual measure — and yet she is back in clinic because the itch that brought her in originally hasn't meaningfully improved, still rating a 7 out of 10 most nights and still the thing keeping her up.
A careful review turns up nothing obvious: no new soaps or garden exposures, no visible sign of infection, and her injection technique and dosing schedule were both confirmed correct at her last visit. What's left is a real, described phenomenon in dupilumab's own trial data — a measurable dissociation, in a meaningful subset of otherwise-responding patients, between how much the skin clears and how much the itch actually improves, consistent with IL-31 continuing to drive pruritus through a pathway only partly overlapping with the one dupilumab blocks.
She has no other chronic medical conditions, no history of anxiety or sleep disorder predating her eczema, and no new stressors she can identify that might otherwise explain a change in symptom perception rather than a true persistence of itch. She did mention, when asked directly, that she has started keeping garden gloves on longer than usual after finishing outside specifically to avoid scratching in front of her grandchildren, who visit most weekends — a small behavioral adaptation that, six months in, she had simply started to accept as her new normal rather than something still worth raising at a follow-up visit.
A drug that cleared the skin without touching the itch
Before we add a second biologic, I want to make sure we're not missing something simpler. Persistent itch with clear-looking skin can still be plain xerosis, a new contact exposure she hasn't connected to her gardening, or a technique issue we haven't caught. Adding nemolizumab is a real cost and commitment, and it should follow ruling those out, not precede it.
I agree the simple explanations deserve a real check, and it sounds like most of them already have been — no new exposures, no infection, confirmed technique and adherence. What's left fits a described pattern in dupilumab's own data: itch and EASI improvement don't always move together, and IL-31 is a specific, separate driver of pruritus that isn't fully blocked by the IL-4/IL-13 pathway dupilumab targets.
This isn't a diagnosis I'm reaching for because we've run out of other ideas — it's a mechanism with its own trial evidence in atopic dermatitis specifically — ARCADIA 1 and ARCADIA 2, not the OLYMPIA prurigo trials people reach for — showing real itch reduction when it's specifically targeted.
That's fair on the workup — it really has been thorough, not just assumed clean. Where I'd still slow us down is the pairing. ARCADIA gave nemolizumab with topical steroids, in patients not on a second biologic; nobody has trialed it on top of dupilumab, so we'd be running her as an n of one on the combination, not on either drug.
Which I can live with, provided we say it to her in those words and provided we structure it as a test rather than an addition. Her EASI is 3. If IL-31 is genuinely carrying the itch, nemolizumab alone should hold the skin and drop the pruritus score, and we'd learn something. If we stack both and she improves, we won't know which drug did it, and she'll be on two biologics indefinitely because neither of us will be willing to be the one who stops the wrong one.
Agreed: continue dupilumab, add nemolizumab specifically for the persistent itch, with a follow-up in six weeks to reassess pruritus NRS against her baseline of 7. Agreed also, and written into the plan rather than left as an understanding: the combination is unstudied, she is to be told so in those words, and six weeks is a decision point — if the NRS has not moved, the pairing stops rather than continuing on the reasoning that it might yet work.
Not agreed, and left standing rather than smoothed over: whether nemolizumab should have replaced dupilumab rather than joined it. The allergist's point that a stacked regimen makes the result uninterpretable was accepted as correct and overruled on practical grounds — nobody was willing to trade an EASI of 3 for a cleaner experiment — which is a different thing from agreeing it was the better design.