Evidence review · Ozempic, Wegovy and Rybelsus (semaglutide), Zepbound and Mounjaro (tirzepatide) · new ANESTHESIOLOGY 2026 records study, randomized trial, ultrasound studies, guidelines, ClinicalTrials.gov review · Sources checked October 19, 2026

Published October 19, 2026 · Updated October 19, 2026

Should you stop Ozempic, Wegovy, Zepbound or Mounjaro before surgery? U.S. guidance now says most people can keep taking it, but the one randomized test found food left in the stomach far more often when people didn’t skip a dose.

On October 18, anesthesiologists presented a study of about 2 million surgeries: people already on GLP-1 drugs had 11% lower odds of heart complications and 8% lower odds of infections afterward.[1] In plain numbers that is 1.0% vs 1.1% and 4.9% vs 5.3%, and the study didn’t establish whether anyone held a dose.[2][3] The worry it doesn’t settle is a full stomach under anesthesia. These drugs slow stomach emptying, and their labels now carry an aspiration warning.[17][18] The only randomized test of skipping a dose found food blocking the exam in 25% of people who kept taking the drug and 3% of those who held one dose, though nobody aspirated.[11] U.S. guidance now says most people can stay on their drug, while European guidance says stop weekly shots for one to two weeks.[8][10] We read the ClinicalTrials.gov records. Of 40 registered studies on these drugs around procedures, only that one trial has posted results, and none tests holding versus continuing before surgery under general anesthesia.[20]

Why this matters now

The American Society of Anesthesiologists put out the new study at its annual meeting in San Diego on October 18, and Newsweek and the KFF Health News morning briefing carried it on October 18 and 19.[1][2][4] The same meeting held sessions on GLP-1 drugs and on fasting and aspiration.[5] On X, posts tying these drugs to surgery or anesthesia ran at 18 to 38 a day last week.[6] For anyone on a weekly shot with an operation coming, the question is practical: take this week’s dose or skip it?

The short answer

• Don’t decide alone. Tell your surgeon and anesthesia team which drug you take, the dose and the date of your last dose, as early as you can. The Zepbound label tells patients to let their providers know about any planned surgery, and Newsweek’s outside expert gave the same advice on drug, dose and last dose.[17][2]
• In the U.S., most people keep taking it. The 2024 guidance from the ASA and four other societies says to continue unless your risk is higher. Higher risk means early in treatment or while the dose is going up, a high dose, a weekly drug, stomach symptoms, or another condition that slows the stomach.[9][8]
• Other countries differ. European anesthesiologists say to stop weekly shots at least a week before, and two weeks if the drug is for weight loss. Australia, New Zealand and one U.S. society instead require a 24-hour clear-liquid diet.[10]
• The new study is about benefit, not stomach risk. It’s a records study, so it can’t show the drug caused the small differences, and it didn’t report aspiration.[1][2]
• The evidence is thin. One randomized trial of 60 people has results. It looked at sedated endoscopy, not surgery under general anesthesia.[11][20]

What the new 2-million-patient study found

The researchers at UTHealth Houston used Epic Cosmos, a pooled health-record database from more than 1,700 hospitals. They took U.S. adults who had elective surgery lasting over an hour, with at least a day in hospital, from 2016 to 2024. Of 2,028,601 patients, 145,610 were on a GLP-1 drug. Users were matched with similar nonusers.[1] After matching, major cardiovascular events happened in 1.0% of users and 1.1% of nonusers. Those events were heart attack, heart injury, stroke or death from any cause. Infections, meaning wound, urinary tract, pneumonia or sepsis, happened in 4.9% and 5.3%.[2] That works out to about one fewer heart event and four fewer infections per 1,000 operations.[3]

Three limits matter for a reader. First, it is observational. People prescribed and staying on a GLP-1 may have better access to care or other differences that matching can’t remove, and the authors say it “does not prove cause and effect.”[1] Second, Newsweek reports that it “did not establish whether or when patients stopped taking GLP-1 medication before surgery,” and that it didn’t report aspiration.[2] So it can’t tell you whether to take this week’s shot. Third, it has been presented at a meeting but not yet published as a peer-reviewed paper.[3] Newsweek’s outside expert called the results “reassuring and hypothesis-generating.” She is Dr. Jessica Duncan, chief medical officer of Ivím Health, a telehealth company that sells compounded tirzepatide and semaglutide and is listed as a partner of the peptide-medicine group lobbying the White House this week. Her advice was still careful: tell your team which drug, the dose and when you last took it, and don’t start a GLP-1 just to prevent complications.[2][19]

Other records studies point the same way, with the same caveat

An analysis by Mass General Brigham anesthesiologists, published in Anesthesiology on September 25 compared surgical patients with type 2 diabetes on a GLP-1 against those on other diabetes drugs. Against metformin, 0.98% vs 2.20% died within 14 days. It found “no increased aspiration risk compared with any other drug class.”[16] A drop of more than half in deaths within two weeks is hard to credit to a weekly shot alone. It is the kind of result that warns how much these comparisons can reflect who gets prescribed what.[16]

Two insurance-claims studies looked straight at lung complications. In 23,679 people with diabetes having same-day emergency surgery, who had no time to hold the drug, respiratory complications were 3.5% with a GLP-1 vs 4.0% on other diabetes drugs (adjusted odds ratio 1.03).[14] In 366,476 people having common operations, aspiration pneumonia was 0.7% vs 0.6% (adjusted odds ratio 0.78, not significant). Acute respiratory failure was more common in GLP-1 users before adjustment, at 13.5% vs 11.1%.[15] Both studies predate wide tirzepatide use, and in the larger one the drugs were mostly semaglutide and dulaglutide.[15]

The stomach problem is real, even if aspiration is rare

The Zepbound label says tirzepatide delays stomach emptying, with the delay “largest after the first dose,” fading over time.[17] The Wegovy label added a warning in November 2024, and the current Zepbound label carries the same one. It cites “rare postmarketing reports” of aspiration under anesthesia in people who still had food in the stomach after fasting as told. FDA adds that the data are “insufficient” to say whether changing fasting or pausing the drug lowers the risk.[18][17]

When anesthesiologists look with ultrasound, they often find food. In a Belgian study, 44 semaglutide users were matched with 44 similar patients for age, BMI and diabetes. After fasting, 49% of users had a full stomach, against 18% of the comparison group, and 42% vs 7% had solid food. The authors say the higher risk held “even after withholding one administration of semaglutide.”[12] In 316 GLP-1 users at three Mayo Clinic hospitals, 35.8% had a lot left in the stomach. Among weekly injectors, those patients had held their shot for a median of 6 days, against 8 days for those with emptier stomachs. Holding 7.5 days or less and fasting from solids for 21.3 hours or less were the cutoffs linked to more food.[13]

The one randomized trial

OCULUS, at two Cleveland Clinic hospitals, randomized people on a stable dose for at least a month. Before a sedated upper endoscopy, they either skipped one dose or kept taking the drug. For weekly drugs, skipping meant 7 to 13 days without it. No one had general anesthesia.[11] Food that blocked or cut short the exam occurred in 25.0% of those who continued, against 3.1% of those who held a dose. The gap crossed the trial’s preset stopping line at 60 of 120 planned patients.[11] Among people who ate normally the day before, it was 46.7% vs 5.0%. Among the 25 who had clear liquids the day before for a colonoscopy, there were no cases in either group.[11]

Two details cut against parts of current guidance. First, none of the patients with a full stomach reported nausea, heartburn or other stomach symptoms that day. The authors conclude that “symptom-based strategies may be insufficient.”[11] Screening for symptoms is one of the main ways U.S. guidance sorts patients into higher risk.[9] Second, the trial was small, and more continuers were on tirzepatide (54% vs 25%). There were no aspirations, unplanned breathing tubes or hospital stays, but the authors say the trial was not powered to show that, and warn against reading that as proof of safety.[11]

What the guidelines say, country by country

U.S. advice reversed in 16 months. In June 2023 the ASA said to hold weekly shots for a week before any procedure.[7] In October 2024 it joined four other societies in saying most patients can continue.[8] That guidance calls itself “guidance, and not an evidence-based guideline.” It says the right hold time is “unknown,” and it adds that it may need to change for dual and triple agonists.[9] Other bodies landed in different places:[10]

All seven recommend a bedside stomach ultrasound, the Canadian one only when there are stomach symptoms on the day.[10]

What we counted on ClinicalTrials.gov

We pulled 2,174 studies that list a GLP-1 drug or match GLP-1 plus anesthesia, sedation, endoscopy, aspiration or post-surgery terms. Then we read the 239 keyword hits. 40 are about these drugs around a procedure. Only 1, OCULUS, has posted results.[20]

• 7 randomized tests of what to do beforehand. Three compare holding with continuing, all before endoscopy or colonoscopy. Two test a liquid diet or longer fasting, and two test a drug that speeds the stomach. None compares holding with continuing before surgery under general anesthesia, the setting the labels warn about.[20][17]
• 3 drugmaker studies. Novo Nordisk and Lilly are each running ultrasound studies of fasting length and pausing their own drugs. Lilly’s dulaglutide study finished in August, and none has posted results.[20]
• 3 randomized trials asking the new study’s question. They test whether giving a GLP-1 prevents heart injury, delirium or wound problems after surgery. None has results, and the earliest is due in March 2027.[20]
• 27 observational stomach-content studies, none with results posted on the registry. Some, like the Belgian study, are published as papers.[20][12]
• Retatrutide: none of the 33 registered retatrutide studies is about procedures.[20]

Every record, the 27 observational studies and the ones we excluded, with reasons, are in our source file.[20]

Strengths and limits

Strengths. The records studies are very large. They agree with each other: no clear rise in aspiration pneumonia in routine care, and slightly fewer heart and infection complications.[1][14][15][16] OCULUS was randomized, and the endoscopists and anesthesiologists didn’t know who had held a dose. The stomach was seen directly rather than estimated.[11]

Limits. Records studies can’t see who skipped a dose or what anyone ate. They also count only aspiration that got coded, and they can’t separate the drug from the kind of patient who takes it.[2][16] The ultrasound studies measure food in the stomach, a stand-in for risk, not aspiration itself.[12][13] OCULUS was 60 people, stopped early, and used sedation for endoscopy, not general anesthesia for surgery.[11] The new study isn’t peer reviewed yet.[3]

Bottom line

Should you stop Ozempic, Wegovy, Zepbound or Mounjaro before surgery? The honest answer is that it depends on your situation and your team, and the evidence behind any fixed rule is thin. U.S. guidance says most people can keep taking these drugs. Holding a dose or a 24-hour liquid diet is suggested if you’re early in treatment, on a rising or high dose, or having stomach symptoms.[9][8] The one randomized trial suggests holding one dose, or clear liquids the day before, cuts the chance of food in the stomach a lot. It also found symptoms don’t reliably warn you.[11] The new study suggests GLP-1 users don’t do worse after surgery and may do slightly better. It can’t show the drug caused that, and it didn’t look at whether a dose was held.[1][2]

The question every anesthesia team will ask is when you took your last dose, and how much. A dated log answers that in seconds. Jabbit keeps one privately on your iPhone.

Frequently asked questions

Should I stop Ozempic, Wegovy, Zepbound or Mounjaro before surgery?

Ask your surgical and anesthesia team, and tell them early. Since late 2024, U.S. guidance from the American Society of Anesthesiologists and four other societies says most people can keep taking these drugs before elective surgery. It suggests holding them, or a 24-hour liquid diet, for people at higher risk of food staying in the stomach: early in treatment or while the dose is going up, on higher doses, with nausea, vomiting or other stomach symptoms, or with other conditions that slow the stomach. Some countries’ guidelines say to stop weekly shots for one to two weeks. Don’t start, stop or skip doses on your own.

How long before surgery do you stop a GLP-1 if your team says to?

U.S. guidance says that if a hold is chosen, skip the weekly shot in the week before the procedure, or a daily drug on the day of the procedure, while admitting “the duration to hold therapy is unknown.” The European anesthesia society says at least one week for weekly shots and two weeks when the drug is taken for obesity. In one 316-patient ultrasound study, food was more often left in the stomach when the shot had been held 7.5 days or less.

Why do anesthesiologists worry about GLP-1 drugs?

These drugs slow how fast the stomach empties. Under general anesthesia or deep sedation, food left in the stomach can come back up and be breathed into the lungs (pulmonary aspiration). The Wegovy and Zepbound labels now warn about this, citing rare reports in people who still had food in the stomach despite fasting as instructed. In stomach ultrasound studies, roughly a third to a half of users still had food in the stomach after fasting. In one matched study the figure was 18% for similar people not on the drug.

Does the new study mean GLP-1 drugs make surgery safer?

Not yet. The study of about 2 million surgical records presented October 18, 2026 found GLP-1 users had 11% lower odds of heart complications and 8% lower odds of infections. In absolute terms that is about 1.0% vs 1.1% for heart events and 4.9% vs 5.3% for infections. It is observational, so people on GLP-1s may differ in ways the matching missed. It did not establish whether people held their dose, and it did not report aspiration. Three randomized trials are testing whether a GLP-1 prevents a complication after surgery, and none has results.

Does a clear liquid diet the day before help?

It looks promising. In the OCULUS randomized trial, none of the 25 patients who had clear liquids the day before (for a colonoscopy) had a problem from food left in the stomach, whether or not they kept taking their drug. That was a subgroup that wasn’t randomized by diet. Two randomized trials testing a 24-hour liquid diet or longer fasting are recruiting, and Australia, New Zealand and one U.S. society already require it.

What about compounded semaglutide or tirzepatide, or retatrutide bought as a peptide?

Compounded versions use the same active ingredients, so the same stomach-emptying effect is expected, but no study we found looked at them separately. Retatrutide, which is not approved, has no registered study on procedures at all, and the U.S. guidance says it may need to change for dual and triple agonists. Tell your team exactly what you inject, the dose and the date of your last shot, even if it came from a compounding pharmacy or a peptide seller.

This page analyzes a conference presentation, published studies, guidance documents, drug labels, press reports and trial-registry records. It is not medical advice and does not recommend starting, stopping or changing any medicine before a procedure; follow your surgical and anesthesia team. The registry review is Jabbit’s own; its inputs are cited above.