The Endocrine Society published a review in September 2025 that was written for endocrinologists and diabetologists and got almost no consumer press attention. It is titled “GLP-1RA and the possible skin aging.” It is five and a half pages long. It is the most important dermatology paper published on the Ozempic phenomenon so far, and almost nobody in the aesthetics world has read it.
The framing you have seen in the mainstream press — and the framing every aesthetic practice in America has been running with — goes like this. GLP-1 drugs cause rapid weight loss. Rapid weight loss depletes facial fat pads. Depleted fat pads cause the skin to sag. Solution: filler. This is the mechanism story the injectable industry is telling, because it is the story that maps most cleanly onto the product they are selling. It is also the story that lets patients believe the drug is doing something to the fat rather than something to their skin.
The Paschou et al. paper says: no, the drug is doing something to your skin. And the something is happening at the level of the stem cells and fibroblasts embedded in your dermis, which express GLP-1 receptors directly, which nobody was talking about until this paper came out. The volume-loss mechanism is real, and it is not the whole story. It may not even be the main story.
What the drug is actually doing to your face
Start with the volume numbers, because they are the ones that make the mechanism story urgent. A quantitative 3D imaging study of patients on semaglutide 2.4 mg weekly (the Wegovy obesity dose), published in PMC in July 2026 (PMC13385476), measured facial fat pad volumes before and after treatment using stereoscopic photogrammetry. Average reductions were 41.8% in the temporal fat pads and 69.9% in the cheek fat pads. Total body fat reduction in the same cohort: 9.2%.
Read that again. Cheek fat pad volume dropped almost seven times faster than total body fat. Temporal fat pad volume dropped more than four times faster. Whatever is happening in the facial fat compartments is not proportional to overall weight loss. It is preferential. The face is being emptied of specific fat depots at a rate the rest of the body is not experiencing. There is no version of this data in which the rapid-weight-loss framing alone accounts for what patients are seeing in the mirror. Something else is happening.
The Paschou paper is the something else. It is a review of the underlying cell biology that makes the case that GLP-1 receptor agonists have direct action on skin cells — not just on adipose tissue in the body cavity, not just on pancreatic beta cells, not just on hypothalamic neurons where the appetite regulation happens. Directly on the cells inside your dermis. Adipose-derived stem cells (ADSCs) and dermal fibroblasts both express GLP-1 receptors on their surfaces. When you inject a GLP-1 agonist, those receptors get hit alongside every other GLP-1 receptor in the body.
The ADSC-DWAT-estrogen axis nobody is talking about
Here is where the paper gets genuinely surprising, because it makes an endocrinological argument that has not entered the aesthetic conversation at all yet.
The stem cell population inside your dermis and subcutaneous fat — adipose-derived stem cells — is not just structural filler. It is metabolically active. ADSCs produce a set of protective cytokines that support surrounding fibroblasts, and they consume glucose to generate the ATP they need to do that work. When GLP-1 receptors on the ADSC surface are stimulated, two things happen in sequence, per Paschou et al. First, glucose uptake into the ADSC drops. This cuts ATP production, which pushes some cells toward apoptosis and impairs the survivors. Second, the ADSCs’ production of protective cytokines drops. Without those cytokines, the local environment shifts toward reactive oxygen species (ROS) production, which then causes oxidative damage to the fibroblasts that live next door in the dermis. Damaged fibroblasts produce less collagen. Skin ages faster.
That is one arm of the mechanism. The other arm is more interesting. Dermal white adipose tissue (DWAT) — the layer of fat that sits inside the dermis rather than beneath it — is one of the sites of local estrogen synthesis in the body. Not the ovaries, not the adrenal glands, but the fat pad in your face itself. This estrogen produced locally in the dermis is a direct stimulator of fibroblasts. Fibroblasts stimulated by local estrogen produce collagen. This is one of the reasons post-menopausal skin loses collagen so dramatically — systemic estrogen drops, and the local DWAT-derived estrogen that had been compensating also declines.
Paschou et al. state the key mechanism directly: “the stimulation of GLP-1R on ADSCs reduces indirectly the production of estrogens from dermal white adipose tissues (DWAT), which reduces stimulation of fibroblasts to produce collagen.” The GLP-1 drug is, in effect, chemically depriving the fibroblasts in your face of the local estrogen signal that was telling them to keep making collagen. The volume loss is one problem. The collagen-synthesis suppression at the molecular level is a parallel problem. Fixing the volume with filler does not fix the collagen problem. Filler compensates for the appearance. It does not restore what the drug is disrupting at the cellular level.
This is why the phrase “Ozempic face” is inadequate to what is happening. What is happening is that GLP-1 drugs are inducing a locally menopausal state in the dermis of your face. Then adding the fat pad depletion on top of that.
The hair loss numbers are also worse than the drug companies want you to know
The FDA has been tracking this. FAERS pharmacovigilance data covering 2022–2023, published in a June 2026 Springer analysis, shows a reporting odds ratio for alopecia of 2.46 (95% CI 2.14–2.83) for semaglutide and 1.73 (95% CI 1.42–2.09) for tirzepatide against the full FAERS comparator background. Reporting odds ratios above 1.0 with confidence intervals excluding 1.0 are treated as pharmacovigilance signals. These are strong signals in both cases.
The registrational trial data is consistent. STEP 1 (semaglutide 2.4 mg): alopecia in 3.0% of the drug arm versus 0.9% of placebo — a roughly threefold rate. STEP 3: 5.9% alopecia on semaglutide. SURMOUNT-1 (tirzepatide 10 mg): 4.2–5.7% dose-dependent alopecia versus 0.9% placebo, up to a sixfold rate. In the dose-response subgroup analysis, patients who lost more than 20% of body weight had a 5.3% alopecia rate, versus 2.5% in patients losing less than 20%. The signal scales with weight loss, which is consistent with a stress-induced telogen effluvium mechanism, but the mechanism is not the entire story either. The FAERS data controls for that: semaglutide has a higher pharmacovigilance signal (ROR 2.46) than tirzepatide (ROR 1.73) despite tirzepatide producing more weight loss on average, which suggests some drug-specific component beyond weight loss alone. The October 2025 review in Skin Appendage Disorders (Burke, Sa, Alvarez Cespedes, and Tosti at the Frost Department of Dermatology, University of Miami; DOI 10.1159/000544023) frames this as class-effect telogen effluvium with a real but incompletely characterized drug contribution.
The regrowth story is somewhat reassuring: telogen effluvium is typically reversible within 6 to 12 months once the triggering stressor stabilizes, and patients who maintain protein intake at 1.2 g/kg body weight or higher during the weight loss phase report meaningfully lower incidence. But the fact that this is a class-labeled adverse event that emerges at 6–12 months of use, dose-dependent, appearing in one out of every twenty patients in the higher-dose arm, is not a footnote. It is a real-life adverse event that a lot of patients are living through in silence because their weight loss is going well.
The immune-mediated skin reactions on the label
The Burke et al. review is the current best summary of the full dermatologic side effect profile of the class. Beyond volume loss and alopecia, GLP-1 receptor agonists are associated with:
Injection-site reactions at the subcutaneous injection site — nodules, erythema, and less commonly injection-site discharge. FAERS data flags semaglutide specifically for injection site discharge with a proportional reporting ratio of 44.9 (n=232 cases) and exenatide for injection site nodules with a PRR of 95.2 (n=1,925 cases). These are class-related rather than semaglutide-specific but track with the injectable delivery route.
Immune-mediated responses: hypersensitivity, urticaria, and — the one that gets attention because of its severity — bullous pemphigoid, an autoimmune blistering disease. Case reports have accumulated across the class, though the incidence is low. The mechanism is not fully characterized but likely involves the drug’s known effects on immune signaling.
Positive dermatologic effects also exist in the emerging evidence: potential enhancement of wound healing, and possible therapeutic benefit in psoriasis (via anti-inflammatory action). The class is not universally bad for the skin. But the bad news is currently outpacing the good news at the level of routine clinical presentation.
A 2026 comprehensive literature review covering 40 studies of GLP-1-associated dermatologic changes (published in dermatologynews.net summarizing the pooled literature) identifies the following as documented and consistent: loss of collagen and elastin fibers (histologically and via elasticity measurements), reduced skin hydration, facial volume depletion, dullness, laxity, and increased visibility of pre-existing wrinkles. Not one aesthetic side effect. A cluster of them, appearing together, on a timeline that begins earlier than the visible weight loss.
What the smart practices are actually doing about it
The dermatology community is moving faster on this than the endocrinology or obesity medicine communities. The LiVDerm clinical dermatology explainer from August 2025 lays out the framework the leading practices are converging on:
Pre-treatment dermatology consultation, before the first injection. This is the meaningful shift. The current standard of care in a lot of endocrinology practices is to prescribe the GLP-1, monitor the metabolic response, and refer to a dermatologist only when the patient shows up unhappy about their face. The emerging standard, driven by dermatology, is to do the baseline consult before treatment starts. Standardized photography, baseline volumetric assessment, discussion of what to expect over the first six months, and a plan for how to intervene if changes are more aggressive than acceptable.
Microdosing as a volume-preservation strategy. Slower rate of weight loss allows the skin more time to remodel. Some practices are titrating GLP-1 doses lower than the standard escalation ladder specifically to blunt the facial volume loss without sacrificing the metabolic benefit. This is off-label and not FDA-endorsed, but it is being deployed in aesthetically motivated patients who want the drug but do not want the face.
Volumetric intervention timed to the loss curve rather than after it. Sculptra, Radiesse, and hyaluronic acid fillers are being deployed proactively during the weight loss phase rather than reactively after volume depletion is complete. Sculptra in particular, because it stimulates new collagen production over months rather than filling immediately, is being used as a countermeasure to the collagen suppression mechanism itself, not just the volume loss.
Nutritional and topical protocols for the collagen problem. Adequate protein intake (1.2 g/kg minimum), oral collagen peptides (5–10 g daily), retinoid maintenance therapy, and increasingly, topical or systemic support for the DWAT-estrogen-fibroblast axis. This is where the therapeutic frontier gets speculative. If the mechanism is a local estrogen deficit in DWAT, then topical estrogen (estradiol cream, applied to the face at low doses) becomes a mechanistic candidate. Some off-label prescribing of topical estradiol for GLP-1 patients is already happening in menopause-aware dermatology practices. The evidence base is thin. The mechanism is defensible.
What this means for anyone considering GLP-1 therapy
Do not not-take the drug because of Ozempic face. If you have a metabolic indication, or a serious weight-management indication, or a cardiovascular risk profile that GLP-1 therapy addresses, the systemic benefits are real and well-documented and dwarf the aesthetic concerns. Skintelect is not in the business of talking anyone out of an evidence-backed treatment for a real medical condition.
But do go in with your eyes open. Understand that the drug is not just melting fat — it is disrupting a specific dermal-endocrine axis that supports collagen synthesis in your face. Understand that the visible changes you may see over the first six months are not merely a passive consequence of weight loss, but an active molecular effect of the drug on the stem cells and fibroblasts inside your skin. Understand that these effects can be partially compensated with pre-treatment planning, dose modulation, nutritional intake, and timely volumetric intervention — but not fully reversed, at least not with anything currently in clinical practice.
And understand that the dermatology consultation for someone starting Semaglutide, Wegovy, Ozempic, tirzepatide, Mounjaro, or Zepbound is now a legitimate pre-treatment step, not a cosmetic afterthought. If your prescribing physician has not raised it, raise it yourself.
The molecular biology paper of 2025 that almost no consumer press covered is going to be, in retrospect, the paper that changed how this class of drug is prescribed. It is also the paper that answered the question your face has been asking for three years.
The rapid weight loss framing was never wrong. It was incomplete. The molecular mechanism is under the fat pad, not on top of it. Treat accordingly.