saw another post this week from someone whose chronic eczema cleared on tirz, and the inflammation angle keeps coming up in this sub, so worth sorting what we actually have from what we’re inferring. the published mechanism story is reasonable but thin. GLP-1 receptors are expressed on a bunch of immune cells, there’s preclinical work showing GLP-1 agonism dampens NF-kB signaling and shifts macrophage polarization, and CRP drops in the registration trials are real and not fully accounted for by weight loss alone. SUSTAIN and STEP both showed CRP reductions that survived adjustment for body weight change, which is the part i find most interesting. the part that’s mostly absent from the literature: dermatologic outcomes as a prespecified endpoint. eczema, psoriasis, hidradenitis, lip dermatitis, none of it was a primary or secondary outcome in any phase 3 i’ve read. so what we have is anecdote, a few case series, and one or two small observational pulls from EHR data that are confounded six ways. the GI side effect profile alone changes diet pretty dramatically, and food triggers are a huge confounder in atopic dermatitis. you can’t separate “tirz dampened my Th2 response” from “i stopped eating the four things that flare me because nothing tastes good at week three.” that doesn’t mean the effect isn’t real. it probably is, in some people, for some inflammatory phenotypes. it means the effect size is real but it’s the kind of real that disappears if you squint at the funnel plot. the thing i’d actually want to see: a prospective cohort with paired CRP, weight change, and dietary recall at 12 and 24 weeks, stratified by baseline inflammatory marker tier. nobody’s running it, because the trial wouldn’t pay for itself.
the CRP-survives-adjustment-for-weight piece in SUSTAIN and STEP is the part that always pulls me up short too, because it’s one of the few signals in this whole conversation that isn’t downstream of either weight loss or appetite suppression. but i’d push the dietary confounder one step further: in T2D specifically, postprandial glucose excursions are themselves a meaningful inflammatory driver, and tirz flattens those independent of what you’re actually eating. so even when the diet is held roughly constant, the glycemic variability piece moves, and that’s a separate input into the inflammatory axis that the “stopped eating the four things that flare me” framing doesn’t fully capture. the stratification by baseline inflammatory marker tier is the bit i’d most want to see, because the responder/non-responder split in the eczema anecdotes is almost certainly hiding in there. people with elevated baseline hsCRP and a Th2-dominant phenotype probably respond differently from people whose baseline inflammation is normal range and whose eczema is more barrier-driven. bundling those two groups into one cohort is exactly how an effect size gets blurred into noise.
the glycemic variability point is the one i’d been hand-waving past and shouldn’t have. CGM cohorts have linked MAGE and time-in-range to hsCRP independent of mean glucose, which is exactly the kind of signal that would survive adjustment for weight and still leave headroom for an inflammatory effect that isn’t really pharmacology so much as smoother curves under the hood. the barrier-driven vs Th2-dominant split you flagged is where i suspect the eczema responder story actually lives, if anyone ever bothers to phenotype properly. the handful of patients i’ve seen self-track this carefully tend to use whatever logging tool tolerates a daily entry without friction (the CareClinic watch complication is one i’ve poked at for that reason), but the real bottleneck is paired bloods often enough that the variability piece stops being noise.
MAGE and time-in-range tracking with hsCRP independent of mean glucose is the cleanest version of this argument I’ve seen anyone make in this thread, and it’s the part of the inflammation story that doesn’t require any direct receptor handwaving. my own CGM 14-day average dropped from 148 to 102 between weeks 4 and 12 on Mounjaro, but the MAGE drop was steeper and earlier than the mean, which is consistent with what those CGM cohorts show. if hsCRP is partially driven by oxidative load from excursions rather than steady-state hyperglycemia, then “smoother curves under the hood” is doing the work that pharmacology is getting credit for, and the eczema responders who are also T2D or pre-diabetic might be picking up an effect that has nothing to do with Th2 at all. the barrier-driven vs Th2-dominant split is exactly where the phenotyping would have to happen, and afaik nobody’s stratified responders that way in any of the case series i’ve come across. filaggrin status alone would split that cohort meaningfully, and it’s a single SNP panel, not exotic. the fact that nobody’s done it is more about funding than feasibility. the paired bloods bottleneck is real though. asking patients to log hsCRP at 12 and 24 weeks with dietary recall and CGM exports is a big ask outside of a funded cohort, and self-tracked CRP is rare enough that you’d be working with maybe two or three serial draws per person if you’re lucky. the variability piece needs at least quarterly bloods to stop being noise, and most people just don’t get that drawn unless something else is flaring. ymmv but the patients i’ve seen who do track that carefully are almost always T2D with a reason to be at the lab anyway, which biases the cohort before you even start.