Can Ozempic Cause Wernicke Encephalopathy? 15 Cases

A study published in Clinical Nutrition identified 15 reported cases of Wernicke encephalopathy in people taking GLP-1 receptor agonists, 14 of them involving semaglutide or tirzepatide. The authors’ ask is increased clinical awareness, not alarm. The signal the paper describes is nutritional: on a GLP-1, appetite falls, intake falls, and the vitamin B1 a body needs does not fall with it.

What Wernicke encephalopathy is

Wernicke encephalopathy is a neurological emergency caused by a severe deficiency of thiamine, also known as vitamin B1. The body needs B1 to turn carbohydrate into energy, and the brain is hit early when the supply runs short. Severe deficiency can produce confusion, trouble walking, balance problems and vision changes, and an untreated case can end in permanent neurological damage. The reason it is easy to miss is that the classic picture rarely arrives whole: in the 15 cases below the full clinical triad of symptoms appeared in only 2 patients, and the authors’ own summary notes that the full triad did not develop in most cases.

What the study found

The paper is a pharmacovigilance study and literature review published in Clinical Nutrition (February 2026; online 2 January 2026). It identified 15 cases of GLP-1 RA-associated Wernicke encephalopathy: 13 from FAERS, the FDA’s Adverse Event Reporting System; 1 from published literature; and 1 from the authors’ own medical centre. Most cases occurred with semaglutide (n = 8/15) or tirzepatide (6/15) — 14 of the 15 — and were reported in 2023-2024 (14/15). Semaglutide is the medicine sold as Ozempic and Wegovy; tirzepatide is sold as Mounjaro and Zepbound.

Source: the abstract of the Clinical Nutrition paper, PubMed 41534460, read September 15, 2026.

Most patients — 13 of the 15 — reported gastrointestinal manifestations of either weight loss, vomiting, loss of appetite or malnutrition, which is the finding that carries this page’s nutrition frame. Long-term neurological sequelae were noted in 7 of 11 patients with follow-up data. On the statistical side, the paper’s disproportionality analysis showed increased reporting of Wernicke encephalopathy with GLP-1 receptor agonists compared with other medications: a reporting odds ratio of 2.35 (95% CI 1.38-4.01) and an IC025 of 0.29.

What a reporting odds ratio of 2.35 measures, and what it does not

The reporting odds ratio is a measure of how the reports sit in a database, not a multiplier on anyone’s personal risk. It compares the share of Wernicke encephalopathy reports among GLP-1 reports with the share among reports for other medicines, inside a voluntary reporting system. It is not an incidence, not a rate and not a relative risk.

Who is at risk

The conditions the authors single out are persistent vomiting, rapid weight loss and poor intake, and their closing sentence names “individuals experiencing severe gastrointestinal symptoms” as the group warranting attention. The mechanism underneath is simple arithmetic about nutrition. Karthik Achari, the Chicago clinician and founder of PepMD quoted by Fox News, put it in one line: “The drug shrinks your appetite. It doesn’t shrink your body’s requirements.” Erin Palinski-Wade, a New Jersey nutritionist also quoted by Fox News, made the same point about the food itself: “GLP-1 medications can dramatically reduce appetite, but that also means people may unintentionally under-eat the very things their bodies need most, especially protein, fiber and fluids.” KFF Health News summarised the same mechanism in its September 14, 2026 morning briefing: “while the drug shrinks your appetite, your body’s requirements remain the same”.

That asymmetry — appetite down, requirement unchanged — is independently visible outside the Clinical Nutrition dataset. A French pharmacovigilance centre, publishing in the European Journal of Clinical Nutrition in November 2025, analysed the World Health Organization’s VigiBase, a different database from FAERS, and described a context of nausea, vomiting or reduced food intake in 68% of cases, with weight loss ranging from 3.5 to 13.3 kg per month over 3 to 6 months. Two independent spontaneous-report databases producing the same surrounding picture is a stronger signal than either database alone, and what both are describing is the nutritional context rather than the medicines themselves.

Put plainly: appetite goes down, intake goes down, and the amount of thiamine taken in goes down with it, while the body’s requirement for thiamine does not move. In people who are already vomiting, or who have lost weight rapidly, that gap is the mechanism the reports point at.

What to do

Nausea and vomiting on semaglutide: when to worry

If you take a GLP-1 medicine and you have persistent vomiting, a rapid weight-loss phase, or a stretch of very poor intake, the question this paper puts in front of clinicians is whether thiamine deficiency is being considered. Its concluding sentence is that early detection may prevent neurological sequelae. Thiamine status is a clinical judgement, made from a person’s history, symptoms and test results, which is why the step this page can point to is a conversation with the clinician managing your care rather than anything you act on from a news article. The paper says nothing about how a medicine is managed, and neither does this page.

What it does not show

Association, not causation

Association is the strongest relation this paper reports. Its own stated aim was to evaluate whether treatment with GLP-1 receptor agonists “is associated with” the occurrence of Wernicke encephalopathy, and a spontaneous-report database records what someone reported rather than establishing why an event happened. Four limits follow directly from the design:

• 15 cases is a small number. The authors call Wernicke encephalopathy “a potentially rare but severe adverse event”. Fifteen is the study’s total across three sources, and it is not the FAERS total either — 13 of the 15 are FAERS reports.
• The reports cluster in a window. 14 of the 15 fall in 2023-2024, a window in which both the use of these medicines and the volume of reports about them rose sharply. A reporting pattern can shift for reasons that have nothing to do with how often an event occurs.
• The statistics measure reporting, not risk. A reporting odds ratio of 2.35 and an IC025 of 0.29 describe disproportionality inside a database. They cannot be used to estimate any individual’s chance of developing the condition.
• One case is unaccounted for in the abstract. Eight semaglutide reports plus six tirzepatide reports is 14 of 15, and the abstract does not say which medicine the remaining case involved.

The authors’ own conclusion is measured: “WE is a potentially rare but severe adverse event of GLP-1 RA treatment, mainly with semaglutide or tirzepatide. As early detection may prevent neurological sequelae, increased clinical awareness is warranted, especially in individuals experiencing severe gastrointestinal symptoms.” They stop short of claiming the medicines account for the cases, and every outlet in the September wave struck the same note: Fox News reported that the study relies largely on reports submitted to the FDA, and The Independent, carried on Yahoo Australia, reported that the researchers described those reports as insufficient for proof of cause.

A separate signal, in the same week. In the same week, a separate meta-analysis of randomised trials (Diabetes Obes Metab, published online 9 September 2026, DOI 10.1111/dom.71331) examined dysaesthesia risk with GLP-1 medicines — a different question, a different study design and a different dataset, and none of the figures on this page should be read across to it. Our coverage of the reported-symptom side effects sits in a different lane again: what users report, and what that reporting can and cannot establish.

Related reading: Underreported GLP-1 side effects: the two symptom classes the Penn study flagged | Semaglutide discontinuation rates 2026: who stops and what’s next | Ozempic hair loss: what the 2026 NYU and Penn studies found | GLP-1 vision loss: NAION lawsuits and what US labels still omit

Sources

1. PubMed 41534460 — Lev D, Leibowitz A, Lang A, et al., Clinical Nutrition 2026 Feb;57:106571 — the study’s abstract, read September 15, 2026, and the source of every figure on this page.
2. DOI 10.1016/j.clnu.2025.106571 — publisher record for the same paper (February 2026 in print, online 2 January 2026).
3. Fox News — Kelly McGreal, September 13, 2026 — the report that opened the September coverage wave, and the source of the quotes from Karthik Achari and Erin Palinski-Wade.
4. KFF Health News — Morning Briefing, September 14, 2026 — the summary of the Fox report and the appetite-versus-requirements framing.
5. The Hill — Josie Fischels, September 14, 2026 — read via its NewsNation syndication on Yahoo Health; thehill.com was rate-limiting requests and is not cited here.
6. The Independent — Julia Musto, via Yahoo Australia, September 15, 2026 — a September 15 re-carry, not a copy of The Hill’s piece.
7. Gras C, De Wit V, Oussedik N, et al., European Journal of Clinical Nutrition 2025 Nov;79(11):1160-1163 (PMID 40908328) — the VigiBase analysis behind the 68% figure and the weight-loss range.
8. Bidesie J, Oudman E., Obesity (Silver Spring) 2026 Sep;34 Suppl 2:64-69 (PMID 42399213) — the independent systematic review cited for the gastrointestinal context.
9. Goldenberg RM, Ibrahim C, Trinacty J., Diabetes, Obesity and Metabolism, published online September 9, 2026 (DOI 10.1111/dom.71331) — the separate, unrelated same-week signal named above.
10. PeptidesBeat — Underreported GLP-1 Side Effects: What the Study Found — our coverage of the reported-symptom class of GLP-1 side effects.
11. PeptidesBeat — editorial policy — our sourcing and preprint-labelling rules.

Corrections. This page is compiled from primary sources, each dated where it appears. One dating point is worth stating rather than annotating later: the Clinical Nutrition paper is a February 2026 print publication, online 2 January 2026, and much of the September 2026 coverage that prompted this page dates the paper to September. Its own bibliographic record does not support that, so the September wave is cited here as coverage and not as a publication date. If a figure changes after publication we annotate the change and its date here rather than editing it silently. Last verified September 15, 2026.

PeptidesBeat is an independent editorial publication covering peptide policy, research, and industry developments. We do not sell peptides, recommend dosing, or provide medical advice. All content is informational. Peptides referenced may be subject to FDA restrictions; consult a licensed healthcare professional for any therapeutic question.