"Ozempic face" is one of the most searched phrases attached to any drug in the class — about a quarter of a million searches a month in the United States. Nothing in any label corresponds to it, and nothing in the randomised evidence measures it. What exists is a literature, and the useful thing to do with a literature is count it.
Everything below was read on 23 September 2026: the PubMed record, each abstract cited, and the current FDA labels for Wegovy (Novo Nordisk, effective 18 June 2026, via DailyMed), Zepbound, Mounjaro and Foundayo (via openFDA). This page does not show faces, and it will not; photographs of faces are the kind of evidence this subject has too much of.
What the phrase describes
The descriptions in the literature agree closely: hollow temples and cheeks, sunken eyes, deeper folds from nose to mouth, looser skin along the jaw and neck — the appearance of ageing, arriving over months. A 2025 systematic review in the plastic-surgery literature (Daneshgaran, Shauly and Gould, PMID 40626110) summarised 23 articles as agreeing that GLP-1 drugs "cause morphological changes resembling advanced aging". What those articles describe are changes seen after weight loss that the drug produced.
That is the unresolved question in one sentence. Facial fat is fat. People who lose a lot of weight by any route lose some of it from the face, and older skin recovers less. The claim that makes "Ozempic face" more than "weight-loss face" is that the drug does something extra — by speed, by pattern, or by some direct effect. That claim is what the evidence below has to carry.
The labels: no entry
The four current weight-management and diabetes labels read for this page — Wegovy, Zepbound, Mounjaro and Foundayo — contain no adverse reaction for facial fat loss, facial ageing or lipoatrophy. The word "lipoatrophy" does not appear in any of them, and none lists a facial reaction. The labels do count other consequences of weight loss that were measured in trials: hair loss, which they attribute to weight reduction (our hair loss page), and, on the Wegovy label, more hip and pelvis fractures in women and older patients in the heart trial (our Wegovy side effects page). No clinical-studies section in the four labels reports facial volume as an endpoint, so there is nothing for a label to count.
The literature, counted
A PubMed search for the exact phrase "ozempic face", run on 23 September 2026, returns 32 records dated 2023 to 2026. Grouped by what they are:
| Type | Count | What it can show |
|---|---|---|
| Letters and commentaries | 6 | Opinion and case impressions |
| Narrative reviews | 8 | Summaries of the above |
| Systematic or literature reviews | 2 | Summaries, graded for bias |
| Search-interest (Google Trends) studies | 2 | That people search for it |
| Consensus statement | 1 | What a panel agreed |
| Case report | 1 | One treatment in one patient |
| Questionnaire development | 1 | How to measure it in future |
| Retrospective imaging study | 1 | A measured change in faces |
| Retrospective surgery-rate study (preprint) | 1 | A procedure rate, not a face |
| Other articles, and one erratum | 9 | Practice articles, a treatment series and a prevention model, citing the term |
None is a randomised trial. One measured a face. The ratio of commentary to measurement is the finding: two studies establish that the public is searching for the phrase, and one establishes what happens to the tissue.
The one measurement
Sharma and colleagues at Vanderbilt (Otolaryngology–Head and Neck Surgery, 2025, PMID 40407186) searched their hospital's records for patients who had been prescribed a GLP-1 drug and happened to have CT or MRI scans of the head and neck both before and after. They found 20. Median age 54, mean time on the drug 321 days, mean weight loss 11.0 kg.
- Median midface volume fell 9% (interquartile range 3% to 14%).
- Superficial fat fell 11%; deep fat 7%, with a range from −20% to +15%.
- Superficial loss tracked weight loss (rho 0.59, p = 0.006); deep loss did not.
- Their regression: roughly 7% of midface volume per 10 kg of weight lost.
Three limits follow from the design, and the authors call their study "one of the first quantitative assessment[s]". Twenty people is a small group. Scans ordered for other medical reasons are not a random sample of anyone. And there was no comparison group of people who lost the same weight another way — so the study measures how much facial volume goes with GLP-1 weight loss, but cannot say whether that is more than goes with any weight loss.
Read plainly, the finding is this: the face lost volume roughly in step with the body, and mostly from the layer nearest the skin. That is what the phrase describes. It is not evidence of a drug effect beyond weight loss, and the paper does not claim one.
The surgery-rate preprint
The most-cited attempt to separate drug from weight is a 2026 preprint from Cleveland Clinic (Almeida et al., Research Square, PMID 42245794). It has not been peer reviewed. Using a large electronic-records network, it matched women with diabetes and hypertension who started a GLP-1 drug to similar women who did not, and counted upper-eyelid surgery (blepharoplasty) over five years.
| GLP-1 users | Matched controls | |
|---|---|---|
| Women in each group | 23,349 | 23,349 |
| Had upper-eyelid surgery | 96 (0.41%) | 36 (0.15%) |
| Hazard ratio | 2.22 (95% CI 1.51–3.26) | — |
A secondary analysis compared women who lost weight without a GLP-1 drug to women whose weight was stable, and reported a hazard ratio of 1.31, 95% confidence interval 0.89 to 1.93, p = 0.56. From this the authors conclude that "weight loss alone did not demonstrate the same association".
Two things in that comparison do not hold as stated:
- The three numbers disagree. A hazard ratio of 1.31 with an interval of 0.89 to 1.93 corresponds to a p-value of about 0.17, not 0.56. Our arithmetic: the interval's width on the log scale gives a standard error of about 0.20, and 1.31 sits about 1.37 standard errors from 1. One of the three reported figures is wrong, and the abstract does not say which.
- The intervals overlap. The drug group's interval runs from 1.51 to 3.26; the weight-loss group's from 0.89 to 1.93. They share 1.51 to 1.93. Showing that one estimate is significant and the other is not is a different thing from showing the two differ, and the preprint does not report a test of the difference.
Even taken at face value, eyelid surgery is a decision, not a measurement. People who can obtain GLP-1 drugs may also be more able, and more inclined, to seek cosmetic procedures. A surgery rate measures who had surgery.
Who writes this literature
It is fair to state, because it is printed on the papers, that much of this record is written by people who treat the condition it describes. The 2025 international consensus statement on managing the aesthetic needs of patients losing weight on prescription drugs (Nikolis et al., Journal of Cosmetic Dermatology, PMID 40135477) lists an employee of Galderma, a manufacturer of dermal fillers, among its thirteen authors, and ten of the other twelve declare speaker, consultant, trainer or investigator ties to Galderma. The review that proposes a prevention algorithm (PMID 42260145) describes it as "anatomy-driven" — a model, not a tested intervention. And the one case report (PMID 41127050) describes a filler product preserving volume in GLP-1 users.
None of that makes the observations wrong. Surgeons see faces and are the people most likely to notice a change. It does mean the literature was assembled largely by the profession that sells the remedy, and that the measured part of it is one twenty-patient study.
What the searches measure
Two of the 32 records are Google Trends analyses (PMID 41255744 and 41521763). They find that searches for "Ozempic face" rose with GLP-1 prescribing and moved with searches for fillers and plastic surgeons. That is evidence about attention. It is often cited as if it were evidence about incidence, and it is not.
What is not known
- Whether GLP-1 drugs change the face more than equal weight loss by other means. No study with a proper comparison group has measured it.
- How common it is. No study reports a rate of visible facial change among people taking these drugs.
- Whether speed matters. It is argued in several reviews; it has not been measured.
- What happens after stopping. Most people regain weight after stopping, as the trials show; whether facial volume returns has not been measured.
- Whether anything prevents it. Protein, resistance training, slower titration and early cosmetic treatment are all proposed. None has been tested for facial outcomes.
Anyone concerned about changes in their appearance on one of these drugs should raise it with their prescriber, and anyone considering a cosmetic procedure should know that the evidence for doing so pre-emptively is a consensus of practitioners, not a trial. The body-composition side of the same question — how much of the weight lost is muscle — is on our muscle loss page, and the photographs that circulate as evidence of results are examined on our before and after page.
Sources
- PubMed search, exact phrase "ozempic face", run 2026-09-23 via the NCBI E-utilities interface: 32 records. Publication types as indexed, grouped by us after reading each title and abstract.
- Sharma RK, Vittetoe KL, Barna AJ, et al. Radiographic Midfacial Volume Changes in Patients on GLP-1 Agonists. Otolaryngol Head Neck Surg. 2025;173(2):360-366. PubMed 40407186
- Almeida VFA, Ha J, Duraes F, et al. Use of GLP-1 Receptor Agonists is Associated with an Increased Rate of Upper Eyelid Blepharoplasty. Research Square [preprint], 2026-05-26. PubMed 42245794
- Daneshgaran G, Shauly O, Gould DJ. "Ozempic Face" in Plastic Surgery: A Systematic Review. Aesthet Surg J Open Forum. 2025;7:ojaf056. PubMed 40626110
- Nikolis A, Enright KM, Fabi SG, et al. Consensus Statements on Managing Aesthetic Needs in Prescription Medication-Driven Weight Loss Patients. J Cosmet Dermatol. 2025;24(4):e70094. PubMed 40135477 — conflict-of-interest statement read.
- Klassen AF, Gallo L, Dayan S, et al. Measuring Facial Appearance in GLP-1 Agonist Use with the FACE-Q Aesthetics Item Library. Aesthet Surg J. 2026. PubMed 42140744
- Carboni A, Woessner S, Martini O, et al. Natural Weight Loss or "Ozempic Face". J Drugs Dermatol. 2024;23(1):1367-1368. PubMed 38206146
- WEGOVY prescribing information (Novo Nordisk, SPL effective 2026-06-18, DailyMed); ZEPBOUND (2026-08-28), MOUNJARO (2026-07-29) and FOUNDAYO (2026-07-29) prescribing information via openFDA — all read 2026-09-23 and searched for facial adverse reactions.