GLP-1 Bowel Obstruction and Ileus: Label Warning and Evidence Review

Bowel obstruction and ileus are listed on the Ozempic, Wegovy, Mounjaro and Zepbound labels, but only as reports received after the drugs reached the market, which give no rate and do not prove the drug caused them.^[1][3] Whether these drugs raise the risk is not settled, and the weight of the evidence has moved against a large effect. The 2023 study behind most headlines found no obstructions at all among people on semaglutide; its signal came from liraglutide.^[5] Larger studies since, in Scandinavia, the US and Japan, found no increase against another diabetes drug class,^[7][8][9] and in the biggest semaglutide trial serious obstruction was reported at almost the same frequency as on placebo.^[13] One UK cohort did find more.^[6] Either way, obstruction is uncommon. If it happens, it is an emergency: belly pain with vomiting, a swollen abdomen and no gas or stool means going to an emergency department.

The evidence at a glance

What the labels say, and when they started saying it

Every current label lists the same bowel terms in section 6.2, the list of adverse reactions reported after approval:^[1][2][3][4]

Gastrointestinal: acute pancreatitis and necrotizing pancreatitis, sometimes resulting in death; ileus, intestinal obstruction, severe constipation including fecal impaction

— Ozempic (semaglutide) prescribing information, section 6.2, June 2026

The same section explains why no number comes with it. The reports are voluntary and the population they come from has no known size, so the label cannot always estimate how often the event happens or establish that the drug caused it.^[1] The labels do have a warning in section 5 that “use has been associated with gastrointestinal adverse reactions, sometimes severe,” and patients are told to contact their prescriber about stomach problems that are severe or will not go away.^[1][4] That warning does not name ileus or obstruction.

The widely repeated story that FDA “added ileus in September 2023” is true for Ozempic and incomplete for the others. Archived label versions on DailyMed show different dates for each drug:

So Wegovy, the weight-loss version of semaglutide, listed ileus about nine months before Ozempic did.^[2][1] For Ozempic, intestinal obstruction itself joined the list only in late 2025.^[1] None of the four has moved ileus or obstruction into the warnings section by name.

The 2023 JAMA study: where the headline came from

In October 2023 a research letter in JAMA used US insurance claims to compare people without diabetes who had an obesity diagnosis and took semaglutide (613 people) or liraglutide (4,144) with people taking the weight-loss pill bupropion-naltrexone (654).^[5] Against bupropion-naltrexone, the two GLP‑1 drugs together carried an adjusted hazard ratio of 4.22 for bowel obstruction, with a 95% confidence interval from 1.02 to 17.40.^[5]

The case counts behind that number tell a different story. Obstruction ran at 8.1 per 1,000 person-years on liraglutide (73 cases), 1.7 on bupropion-naltrexone (2 cases) and zero on semaglutide.^[5] Semaglutide users were followed for a median of only 0.6 years. When the authors dropped high cholesterol from the adjustment, the hazard ratio became 3.63 with an interval from 0.87 to 15.10, which no longer excludes no effect.^[5] An interval that wide comes from having only two events in the comparison group. The study raised a real question. It did not show that Ozempic or Wegovy causes obstruction, and it could not, because no one on semaglutide had one.

What larger studies have found since

Most later studies compared GLP‑1 users with people starting an SGLT2 inhibitor, such as Jardiance or Farxiga, a common second-line diabetes drug with no known effect on the gut. That design compares people at a similar stage of diabetes rather than treated against untreated people.

The UK study is the one large cohort that found more obstruction: 1.9 against 1.1 hospitalizations per 1,000 person-years, with the gap widest after about a year and a half of use. Its authors calculated that one extra obstruction would occur for every 1,223 people treated for a year.^[6] The Scandinavian registries, with almost five times as many GLP‑1 users, found a slightly lower rate on GLP‑1 drugs and concluded that earlier safety signals were not confirmed.^[7] A US study of over a million people with diabetes found no difference against any other class except insulin, where the GLP‑1 rate was lower, and the result held at one, three and five years and in people with and without obesity.^[8] The Japanese study counted only obstructions that needed a tube to decompress the bowel or surgery, and found 0.14% against 0.17% over a year.^[9]

These are people with diabetes, most taking older drugs at diabetes doses. No large cohort has yet reported obstruction rates for people taking Wegovy or Zepbound at full weight-loss doses.

What the randomized trials show

Trials randomize, so they avoid the comparison problems of cohorts, but obstruction is rare enough that most trials see only a handful. Pooled, they have not found an increase. A 2025 meta-analysis of 55 placebo-controlled trials with 106,395 people looked at intestinal obstruction and paralytic ileus among other events and concluded the drugs probably have little or no effect on them.^[11] A 2026 network meta-analysis of 50 trials and 192,359 people found no significant increase for any GLP‑1 drug, and a lower risk on liraglutide, the drug that drove the 2023 claims signal.^[10] A meta-analysis that mixed trials with observational studies found no overall increase but huge disagreement between studies, and a higher risk for liraglutide in a drug-by-drug subgroup.^[12]

The single largest placebo-controlled data set is SELECT, which randomized 17,604 adults with heart disease and overweight to semaglutide 2.4 mg, the Wegovy dose, or placebo for up to 240 weeks. Its posted serious adverse events break obstruction into several terms:^[13]

Added together, the obstruction and ileus terms come to 27 on semaglutide and 28 on placebo, about 3 reports per 1,000 people in each arm over up to four and a half years. Obstruction was not a planned outcome, and the trial enrolled people with heart disease, not the younger population many weight-loss prescriptions go to. It is still the closest thing to a direct answer for Wegovy, and it does not show an excess.

Why the signal keeps appearing in report databases

Analyses of adverse-event reporting databases do keep finding obstruction reported more often with these drugs than with other diabetes medicines. In the World Health Organization’s database, intestinal obstruction was reported about three times as often with GLP‑1 drugs as with other diabetes drugs; for DPP‑4 inhibitors the figure was more than eight times.^[14] Reports reflect publicity, label changes and who takes a drug, and they carry no denominator. A genetic study that used natural variation in the GLP‑1 receptor to mimic lifelong drug exposure found a lower risk of paralytic ileus, not a higher one, though the confirming analysis in UK Biobank did not reach significance.^[15]

Ileus, obstruction, constipation and gastroparesis are different things

The terms get used interchangeably online, but they name different problems, and the difference matters for what to do.

MedlinePlus lists the causes of ileus as including abdominal surgery, infection, low potassium and “certain medicines, especially narcotics,” and the causes of mechanical obstruction as including adhesions from earlier surgery, hernias and impacted stool.^[23] GLP‑1 drugs slow the gut, which is why a drug-induced ileus is biologically plausible, but no study has shown that this is how obstruction happens on them. Constipation is covered in GLP-1 constipation: what works and slow stomach emptying in Ozempic and gastroparesis.

The published case reports show the range. A 65-year-old woman on semaglutide for weight loss needed part of her small bowel removed because a ball of undigested plant fiber, a bezoar, had blocked it; her surgeons suggested, without being able to prove, that slowed stomach and bowel transit allowed it to form.^[20] Two reports describe functional small-bowel obstruction, with no physical blockage, in women who had never had abdominal surgery, both after a tirzepatide dose increase.^[21][22] One of those patients was managed without surgery, and the authors noted that this kind of obstruction can look mechanical on a scan.^[22] Case reports show what can happen, not how often.

Who is likely at higher risk

No study has built a risk score for obstruction on GLP‑1 drugs. What the evidence does point to:

• People starting an opioid. Among 411,188 adults with diabetes on a GLP‑1 drug who started an opioid, the 30-day risk of severe constipation, bowel obstruction or gastroparesis was 0.51% with oxycodone, against 0.35% with hydrocodone and 0.33% with tramadol.^[16] The study had no group without an opioid, so it shows oxycodone carrying more risk than the alternatives, not the size of the combined risk.
• People with earlier abdominal surgery or hernias. Adhesions and hernias are the classic causes of mechanical obstruction regardless of medication.^[23]
• Severe constipation. Impacted stool is itself a cause of obstruction, and fecal impaction is now on every label’s postmarketing list.^[23][1]
• Dose increases. Two of the case reports describe obstruction soon after a step up in tirzepatide dose.^[21][22] That is a pattern in anecdotes, not a measured risk.

Two groups people often worry about look reassuring so far. In a Danish nationwide study of people with inflammatory bowel disease, GLP‑1 use was not linked to more ileus or obstruction.^[17] And among 13,361 adults with diabetes having surgery under general anesthesia, ileus within a week after the operation was no more common in GLP‑1 users.^[18] How to handle a dose before a procedure is set out in GLP-1 drugs before endoscopy and surgery.

Warning signs and what to do

MedlinePlus lists the symptoms of intestinal obstruction as abdominal swelling, fullness and gas, pain and cramping, inability to pass gas, constipation or sometimes diarrhea, and nausea and vomiting. It advises contacting a clinician for any of these:^[23]

• You cannot pass stool or gas.
• Your abdomen is swollen and stays that way.
• You keep vomiting.
• You have abdominal pain that does not go away and has no explanation.

A partial small-bowel obstruction without complications can often be treated without surgery: nothing by mouth, fluids through a vein, and a tube through the nose to drain the stomach. Surgery is needed if that does not relieve it or if the bowel tissue is at risk of dying.^[23] Repeated vomiting also drains fluid, and every GLP‑1 label warns of kidney injury from dehydration; see GLP-1 drugs and your kidneys.

After an obstruction: restarting the drug

No guideline says whether to restart a GLP‑1 after an obstruction, and no study has measured how often obstruction recurs on rechallenge. In a Japanese study of 196 people with diabetes hospitalized for obstruction or ileus while on a GLP‑1, 43.9% did not resume it within six months. Stopping was more likely after surgery or a long stay. The authors noted that how much the patient stood to gain from the drug seemed to play little part in the decision.^[19] That decision belongs with the prescriber and the surgical team, and it should weigh what caused the blockage: a hernia or adhesion points away from the drug more than an ileus with no mechanical cause.

The lawsuits: what the court record shows

Federal claims over stomach and bowel injuries from GLP‑1 drugs are combined in In re: Glucagon-Like Peptide-1 Receptor Agonists (GLP-1 RAs) Products Liability Litigation, MDL No. 3094, before Judge Karen S. Marston in the Eastern District of Pennsylvania. When the Judicial Panel on Multidistrict Litigation created it in February 2024, the injuries named were “gastroparesis, ileus, intestinal obstruction or pseudo-obstruction, or other gastrointestinal injury.”^[24] On October 1, 2026, 4,025 cases were pending.^[26] Novo Nordisk and Eli Lilly deny the allegations.^[27]

Most cases center on gastroparesis: in 2024 plaintiffs’ lawyers told the court they expected over 95% to allege it.^[25] The label placement is a central dispute. In setting the schedule, the court summarized the plaintiffs’ position that a passing mention of ileus in the postmarketing section, separate from the warnings, does not help prescribers identify or reduce the risk.^[25] The court split the case into three cross-cutting questions: how gastroparesis must be diagnosed, whether the warnings were adequate or claims are preempted by federal law, and general causation.^[25] An expert-evidence hearing on the second and third was set for September 14 to 18, 2026, with closing arguments on October 13, 2026.^[27][28] No ruling on them had been posted as of mid-October 2026. The gastroparesis testing ruling and the earlier milestones are covered in Ozempic and gastroparesis.

What to ask a prescriber

1. I have had abdominal surgery, a hernia or a previous blockage. Does that change the plan? These are the main causes of mechanical obstruction regardless of drug.
2. What should I do if constipation lasts more than a few days? Agree on a plan before it happens, since impacted stool is itself a cause of obstruction.
3. I am about to be prescribed an opioid. Does the choice matter? In people on a GLP‑1, oxycodone carried more short-term bowel risk than hydrocodone or tramadol.
4. Should the next dose increase wait? Ask whether to hold at the current dose if bowel symptoms are not settled.
5. Which symptoms mean emergency care rather than a call? Write down the answer.