Most questions in this category have an answer somewhere, even if it is buried in a label. This one does not — and the reason it does not is more useful than a false answer would be.
Four separate FDA labels address what happens when someone taking one of these drugs goes under general anaesthesia. All four describe the same risk. All four then state, in the same sentence, that there is not enough evidence to say what should be done about it. Meanwhile professional societies in different countries have issued guidance that does not agree, and the largest recent analysis of actual outcomes points in a third direction again.
Labels below were read on 13 September 2026 from DailyMed, the National Library of Medicine's index of current FDA labelling. Nothing here is medical advice. Our editorial standards set out where this site stops.
What the labels say, word for word
Ozempic, Wegovy, Mounjaro and Zepbound each carry a warnings section headed "Pulmonary Aspiration During General Anesthesia or Deep Sedation". The wording is near-identical across all four, changing only the drug name. Each states that the drug delays gastric emptying, and then:
There have been rare postmarketing reports of pulmonary aspiration in patients receiving GLP-1 receptor agonists undergoing elective surgeries or procedures requiring general anesthesia or deep sedation who had residual gastric contents despite reported adherence to preoperative fasting recommendations.
And immediately after it, the sentence that defines the whole subject:
Available data are insufficient to inform recommendations to mitigate the risk of pulmonary aspiration during general anesthesia or deep sedation in patients taking [the drug], including whether modifying preoperative fasting recommendations or temporarily discontinuing [the drug] could reduce the incidence of residual gastric contents.
Read that carefully. The regulator names the two candidate mitigations everyone reaches for — change the fasting rules, or stop the drug — and says there is not enough evidence to know whether either works.
The one action the labels do specify is directed at prescribers: instruct patients to inform their healthcare providers of any planned surgeries or procedures.
This is an unusually explicit refusal. Labels routinely omit guidance on a question; it is much rarer for one to name the obvious response and record that the evidence for it is missing. Every claim made downstream by anybody else is being made in that gap.
Three positions, and they do not agree
| Source | Date | Position |
|---|---|---|
| FDA labelling, all four products | Current at 13 Sept 2026 | Risk documented; data insufficient to recommend fasting changes or stopping |
| American Society of Anesthesiologists guidance | 2023 | Hold the drug before surgery |
| Korean Society of Anesthesiologists consensus | 30 July 2026 | Continue in most patients, with individualised risk assessment, consideration of a 24-hour clear liquid diet, and gastric ultrasound where delayed emptying is suspected |
The Korean statement is the most recent of the three and is candid about its own foundations: it records that the current evidence base remains insufficient to formally support evidence-graded clinical practice guidelines, and that its recommendations were developed through expert discussion and review of a literature in which robust evidence is lacking. It is a consensus document that says so on its face.
One caveat on the ASA line, and we state it rather than paper over it. ASA's own 2023 page was not reachable at the time of writing — the URL returns a not-found error. The characterisation above is taken from how a 2026 paper in Anesthesiology, ASA's own journal, describes it in its methods: the study period was chosen to end "before the 2023 American Society of Anesthesiologists guidance to hold GLP-1 RAs preoperatively." That is a strong secondary source for the substance of the guidance and it is still a secondary source, and our standards require saying which is which.
That there is international variability is not our inference. It is the stated subject of a 2026 paper in the British Journal of Anaesthesia, whose title is "Perioperative management of glucagon-like peptide-1 receptor agonists: international guidance and variability in recommendations."
The outcome data, and why its date was chosen
The most informative recent evidence is observational, and its design contains a clever detail.
A study published in Anesthesiology on 12 August 2026 drew electronic health record data from June 2013 to June 2023 — a window ending deliberately before the 2023 ASA hold guidance, so that it captures a period when patients taking these drugs were generally not being held before surgery. Adults with type 2 diabetes undergoing anaesthesia were propensity-matched one to one against patients on metformin, on SGLT2 inhibitors, or on DPP-4 inhibitors, with no concomitant use of the comparator.
| Comparison | Outcome | GLP-1 | Comparator | Risk ratio |
|---|---|---|---|---|
| vs metformin | 14-day mortality | 0.98% | 2.20% | 0.44 (0.31–0.64) |
| vs DPP-4 inhibitors | 14-day mortality | 1.84% | 2.89% | 0.63 (0.45–0.89) |
| vs DPP-4 inhibitors | Bacterial pneumonia | 0.85% | 1.80% | 0.47 (0.29–0.76) |
| vs SGLT2 inhibitors | Major adverse cardiovascular events | 3.86% | 5.03% | 0.77 (0.60–0.98), not significant after adjustment |
And the finding most directly relevant to the warning: no increased aspiration risk against any comparator class.
The limits are real and the authors state them. This is observational, not randomised. Patients prescribed a GLP-1 drug differ from patients prescribed metformin in ways propensity matching narrows but cannot eliminate, and a mortality difference of that size in a 14-day window is more plausibly telling us something about who receives which drug than about what happens during an anaesthetic. Most importantly, it is not a trial of holding versus continuing. It shows what happened in a period when holding was not standard; it does not show that holding would have made things worse. The authors call for prospective studies, which is the correct conclusion and also an admission that the question at the centre of this page remains open.
Why the disagreement exists at all
The mechanism is not in dispute. These drugs delay gastric emptying — every one of the four labels states it plainly and cross-references its clinical pharmacology section. Preoperative fasting rules exist to produce an empty stomach before anaesthesia. A drug that slows emptying makes the standard fasting interval a weaker guarantee.
What is in dispute is what follows. Stopping a drug carries its own costs — glycaemic control in diabetes, and for a weekly injection an interruption that is measured in weeks rather than hours. Against that sits an event the labels describe as rare and the largest matched analysis did not detect. Two societies weighing the same thin evidence have landed on opposite defaults, and the Korean statement's answer — assess the individual stomach with ultrasound rather than infer from the prescription list — is an attempt to stop guessing on either side.
What this page does not answer
It does not say whether anyone should stop, hold, continue, or alter anything before a procedure, and it contains no dose and no schedule. It cannot, and no honest page can: the drug labels decline to take a position, the professional guidance disagrees across borders, and the trial that would settle it has not been run.
The one point every source agrees on is the point the labels make: the anaesthetic and surgical team need to know the drug is being taken. Local protocol decides the rest, and local protocol is exactly what varies.
Related reading on this site: the gastric-emptying effect behind the warning also drives the commonest adverse reactions, covered on GLP-1 side effects, and the gallbladder events that follow a different pattern across the three labels are on the gallbladder page.
Sources
- DailyMed, OZEMPIC (semaglutide) injection (SPL version 20), WEGOVY (semaglutide) injection / tablet (version 19), MOUNJARO (tirzepatide) injection (version 40) and ZEPBOUND (tirzepatide) injection (version 40) — all read 13 September 2026. Section 5.10 in the two Novo Nordisk labels and section 5.9 in the two Lilly labels, "Pulmonary Aspiration During General Anesthesia or Deep Sedation", for every quotation above.
- Bang YJ, Choi JH, Min JJ, Park SY, Choi WK, Hwang JY, Koo BN. "Korean consensus-based recommendations for preanesthetic gastric ultrasound and anesthetic considerations for GLP-1 receptor agonist-based therapy: a consensus statement from Clinical Practice Guideline Committee of the Korean Society of Anesthesiologists." Korean Journal of Anesthesiology, published online 30 July 2026. doi:10.4097/kja.26655. Abstract read 13 September 2026.
- Choi UE, Nicholson RC, Messinger C, Oprea AD, Sweitzer BJ, Hobai IA, Connors JM, Hepner DL. "Preoperative Glucagon-Like Peptide-1 Receptor Agonists and Postoperative Outcomes: An Observational Analysis." Anesthesiology, published online 12 August 2026. doi:10.1097/ALN.0000000000006326. Abstract read 13 September 2026, and the source of this page's characterisation of the 2023 ASA guidance.
- Lam C, Thiruvenkatarajan V, Nanjappa N, Endlich Y, Lin DY. "Perioperative management of glucagon-like peptide-1 receptor agonists: international guidance and variability in recommendations." British Journal of Anaesthesia 2026;136(5):1682–1685. doi:10.1016/j.bja.2026.01.028. Cited for its subject; the full text was not opened for this page and no finding is attributed to it beyond its stated topic.
- The American Society of Anesthesiologists' own 2023 guidance page did not resolve when requested on 13 September 2026.
Guidance in this area is moving quickly and varies by country and by hospital. Every position above is dated and attributed to the document that states it.