GLP-1 Medications and Ileus: The FDA's Postmarketing Label Warning, the JAMA Cohort, the Scandinavian Counter-Finding, and the ER Symptoms You Cannot Ignore

In late 2023, the FDA added “ileus” and “intestinal obstruction” to the postmarketing experience section of every GLP-1 receptor agonist label — Ozempic [8], Wegovy [7], Zepbound [9], Mounjaro [10], Saxenda [11], and later Foundayo [11]. The published evidence on the magnitude of the signal is genuinely mixed: Sodhi et al JAMA 2023 (PMID 37796527) reported a hazard ratio of 4.22 (95% CI 1.02-17.40) for bowel obstruction in a US weight-loss cohort [1], while a 121,254-patient Scandinavian cohort (PMID 37716613) [2] and a 2025 Gastroenterology meta-analysis of 55 RCTs (PMID 40499738) [3] did not replicate the elevated signal. The absolute risk is small either way — but the consequences of a missed ileus include emergency surgery, NG-tube decompression, and rare death. This article walks through the verbatim FDA label language, all three published datasets, the enteric-nervous-system mechanism, and the symptom profile that distinguishes ileus from ordinary GLP-1 nausea so the ER-level red flags are unmistakable.

The verbatim FDA label language

The labels are nearly identical across the GLP-1 class. All six FDA-approved drugs list ileus and intestinal obstruction in Section 6.2 (Postmarketing Experience), pulled 2026-05-07 directly from the DailyMed-hosted prescribing information for each:

The Postmarketing Experience section — distinct from the Adverse Reactions section that summarizes clinical-trial rates — is FDA shorthand for “reported voluntarily from a population of uncertain size, so we cannot calculate a frequency.” That is exactly why the labels do not provide a percentage and exactly why the JAMA cohort study below matters.

Sodhi et al, JAMA 2023 — the strongest signal

The Sodhi et al letter in JAMA 2023 (PMID 37796527) is the single most-cited pharmacoepidemiology paper on GLP-1 and gastrointestinal adverse events in the weight-loss population [1]. The study used the PharMetrics Plus commercial-claims database (2006-2020) and compared semaglutide and liraglutide users to bupropion-naltrexone users — all three indicated for weight loss. n=5,411 total. Hazard ratios for prespecified outcomes:

• Bowel obstruction: HR 4.22 (95% CI 1.02-17.40) — statistically significant.
• Pancreatitis: HR 9.09 (95% CI 1.25-66.00) — statistically significant, very wide CI reflecting small event count.
• Gastroparesis: HR 3.67 (95% CI 1.15-11.90) — statistically significant.
• Biliary disease: HR 1.50 (95% CI 0.89-2.53) — not statistically significant in this dataset (the broader gallbladder evidence including bariatric-surgery rapid-loss data is in our companion GLP-1 + gallbladder article).

The wide confidence intervals — particularly for pancreatitis (1.25-66.00) — are the result of small absolute event numbers in a cohort of 5,411. The point estimate matters; the precision is genuinely limited. Sodhi et al stated this directly in the letter conclusion: “Use of GLP-1 agonists for weight loss compared with use of bupropion-naltrexone was associated with increased risk of pancreatitis, gastroparesis, and bowel obstruction but not biliary disease.” That sentence, plus the coincident timing of the FDA postmarketing label addition (September 2023), is what made this letter clinically load-bearing despite its size.

The Scandinavian counter-finding

A much larger Scandinavian cohort study published the following year in Clinical Gastroenterology and Hepatology (Ueda et al, PMID 37716613) compared 121,254 GLP-1 receptor agonist users to 185,027 SGLT2 inhibitor users across Sweden, Denmark, and Norway from 2013 to 2021 [2]. The finding for intestinal obstruction:

• GLP-1 RA vs SGLT2i: HR 0.83 (95% CI 0.69-1.01) — not statistically elevated.
• Incidence rate: 1.3 per 1,000 patient-years (GLP-1 RA) vs 1.6 per 1,000 patient-years (SGLT2i).

Three differences plausibly explain the gap from Sodhi's finding: the Scandinavian cohort was mostly type 2 diabetes patients (not weight-loss patients with higher BMI and faster loss); the comparator was a different diabetes drug class (SGLT2i, not bupropion-naltrexone); and the Scandinavian sample was 22-fold larger, giving tighter confidence intervals. Either way, the Scandinavian dataset does not support a clinically meaningful excess of ileus on a GLP-1 background.

The 2025 RCT meta-analysis

A 2025 Gastroenterology systematic review and meta-analysis of 55 randomized controlled trials including 106,395 participants (PMID 40499738) examined GLP-1-associated gastrointestinal adverse events comprehensively [3]. For intestinal obstruction and paralytic ileus, the conclusion was that GLP-1 RAs “probably have little or no effect” on the risk versus placebo. The same meta-analysis confirmed elevated risks for cholelithiasis (RR 1.46, 95% CI 1.09-1.97) and for gastroesophageal reflux disease (RR 2.19, 95% CI 1.48-3.25) — so it is not the case that the meta-analysis simply failed to find any signal.

RCTs systematically under-detect rare events because they enroll healthier subjects, exclude patients with prior abdominal surgery or other ileus risk factors, and are shorter in duration than real-world cohorts. That makes the RCT meta-analysis well-powered for common GI symptoms but underpowered for ileus specifically.

Why the studies disagree

The three datasets are not as contradictory as they look on first read. Sodhi captured a US weight-loss-only commercial population with relatively rapid loss and a comparator (bupropion-naltrexone) that is less GI-active than the comparators in the other studies. The Scandinavian study captured a mostly-T2D population with slower titration, less weight loss, and an SGLT2i comparator that itself has no ileus signal. The RCT meta-analysis pooled trials with protocol-driven titration and pre-screened patients. The most coherent reading of all three together: a small absolute excess risk that is most visible in the rapid-weight-loss US weight-management context and washes out in the broader T2D and trial populations.

The mechanism — GLP-1 receptors in enteric neurons

Two mechanism papers are load-bearing. Schirra et al (Neurogastroenterol Motil 2008, PMID 18298441) [4] documented that physiologic GLP-1 doses suppress the migrating motor complex (MMC) in the human jejunum, reducing the median fasting MMC count and motility index. Amato et al (Neurogastroenterol Motil 2010, PMID 20158614) [5] characterized the peripheral motor action of GLP-1 through enteric neuronal receptors and showed that GLP-1 inhibits small-intestine motility via presynaptic GLP-1 receptors on enteric neurons, modulating nitric-oxide release and decreasing cholinergic neurotransmission.

The clinical consequence: GLP-1 receptor agonists slow the whole gut, not just gastric emptying. In a patient with good baseline motility, that produces ordinary GLP-1 symptoms (early satiety, mild constipation, occasional nausea). In a patient with predisposing factors — adhesions from prior abdominal surgery, opioid co-administration, narrowed bowel from prior radiation, baseline severe constipation — the same motility suppression can tip the bowel into an actual ileus. None of these risk modifiers has a published adjusted hazard ratio in the GLP-1 literature yet, so any patient with these conditions should discuss them with the prescribing clinician.

How to tell ileus from ordinary GLP-1 GI symptoms

The single most useful clinical skill for a GLP-1 patient is the ability to distinguish typical post-injection nausea from the signs of actual ileus. The contrast is sharp once you know what to look for:

Any one of the right-column items is an ER trigger. Two or more, especially distension plus absent flatus, is an emergency-room presentation. Do not call your prescriber first — go directly. Acute small-bowel obstruction progressing to strangulation or perforation can become irreversible within hours.

What to do if you suspect ileus

1. Stop the GLP-1. Do not take your next weekly injection if you have ileus symptoms. The drug is worsening the picture.
2. Stop oral intake of food and fluids. Do not try to push solids or liquids through. Continued intake into an obstructed bowel is dangerous.
3. Go to the emergency department. Call 911 if pain is severe or you are vomiting blood or coffee-ground material.
4. Tell the ED team you are on a GLP-1 receptor agonist. Name the specific drug. They will need this for both the differential diagnosis and any imaging or surgical planning. The American Society of Anesthesiologists has formal guidance on perioperative GLP-1 management — see our GLP-1 + surgery and anesthesia article.
5. Bring your medication list. Opioid analgesics worsen ileus; the ED needs to know what else you are on.

Risk factors and the not-yet-quantified questions

The published pharmacoepidemiology has not yet quantified adjusted hazard ratios for the most clinically obvious ileus risk modifiers in a GLP-1 context. Specifically: prior abdominal surgery with adhesions, opioid co-administration, prior radiation enteritis, severe baseline constipation, and aggressive titration speed are all mechanistically plausible risk amplifiers, and societies including the American Society of Anesthesiologists and the American Society for Parenteral and Enteral Nutrition have published consensus guidance reflecting that plausibility, but no peer-reviewed cohort study has published adjusted HRs by risk modifier. If you have any of these conditions and you are starting a GLP-1, raise them with your prescriber explicitly — the clinical-judgment adjustment is real even if the formal adjusted estimate is not yet on PubMed.

The bottom line

• Every FDA-approved GLP-1 label lists ileus and intestinal obstruction in Section 6.2 (Postmarketing Experience), added in 2023.
• Sodhi et al JAMA 2023 (PMID 37796527) reported HR 4.22 (95% CI 1.02-17.40) for bowel obstruction in a US weight-loss cohort — the strongest single signal.
• Ueda et al Scandinavian cohort 2024 (PMID 37716613, n=121,254) found HR 0.83 — not significant.
• 2025 Gastroenterology meta-analysis of 55 RCTs (PMID 40499738) found “little or no effect” on intestinal obstruction.
• Best reading: small absolute excess risk most visible in the rapid-weight-loss US weight-management context; washes out in broader populations.
• The clinically actionable skill is recognizing the symptom profile — distension, no stool/gas for >24h, severe constant pain, persistent vomiting — and going to the ER, not the prescriber, when those appear.

Related research

• GLP-1 side effects: what the trials actually showed — the broader STEP/SURMOUNT adverse-event picture this article zooms in on.
• Practical GLP-1 nausea management — the ordinary side-effect baseline that ileus diverges from.
• GLP-1 + gallbladder/biliary disease — companion piece on the other major GI label warning.
• GLP-1 + surgery and anesthesia (ASA guidance) — directly relevant if a bowel obstruction requires surgical intervention.
• How to taper off a GLP-1 safely — relevant if a clinician recommends stopping after an ileus episode.