Inflammatory bowel disease is two diseases with one name. Crohn’s disease can involve any part of the gut and can narrow it; ulcerative colitis is confined to the colon. Obesity is now common in both, and the people who have both are being prescribed the same weight-loss drugs as everyone else. What almost nobody quoting this arrangement mentions is how little of it has been tested.
A PubMed search restricted to the randomized-controlled-trial publication type, crossing the five main incretin molecules with Crohn’s disease, ulcerative colitis and inflammatory bowel disease, returns zero records. The same query without the type restriction returns 97, so the zero is an absence of randomized evidence rather than a failed search. Everything below comes from matched electronic-record cohorts, and the gap between those and a trial is the subject of this page. The broader tolerability picture the cohorts sit inside is in what the trials recorded on side effects.
The labels say nothing, and the trials never reported on it
Four labels were pulled from DailyMed as structured product labeling and stripped to plain text. Across Wegovy, Ozempic, Zepbound and Mounjaro — 212,786, 137,730, 154,687 and 142,530 characters of extracted label respectively — the strings inflammatory bowel, Crohn, ulcerative colitis and colitis appear a combined zero times. In the same four extractions, diarrhea appears 8 to 11 times, pancreatitis 16 to 19, and gastroparesis 2 to 3. The zeros are real rather than an artifact of a broken fetch.
That is silence, not permission and not prohibition. The obesity and cardiovascular outcome programs never published a result in this population either, which leaves the entire question to registries and claims databases. Five interventional studies are registered. Two are phase 3b, pairing an interleukin-23 antibody with tirzepatide in 350 people with ulcerative colitis and 290 with Crohn’s disease, with primary completion dates in April and May 2028. The other three enroll 60, 24 and 20 participants. None has reported.
The retrospective record is enormous, and it is all retrospective
A 2025 systematic review pooled 11 studies and 16,242 patients with IBD treated with this class. Weight fell by 9.6 kg on semaglutide (95% CI −12.0 to −7.2), 9.4 kg on liraglutide and 11.8 kg on tirzepatide at three months, and IBD-related surgery was less frequent, at an odds ratio of 0.46 (95% CI 0.32 to 0.67, I² = 42%).[1] A 2026 single-arm meta-analysis of 10 studies and 7,831 patients found total weight loss of 7.55% and fecal calprotectin falling from 289.46 to 210.59 µg/g, with 11% requiring intravenous steroids and 6% requiring advanced therapy.[2]
The individual cohorts are larger still and their effects are larger again. A propensity-matched analysis of 11,016 patients with IBD — 5,508 exposed, 5,508 not — reported five-year mortality of 8.05% against 10.03%, a hazard ratio of 0.65 (95% CI 0.54 to 0.78), with IBD-related surgery at HR 0.53 and complications at HR 0.81.[3] A Crohn’s-specific cohort matched 9,766 pairs and found one-year mortality of 0.7% against 4.2%, a relative risk of 0.17 (95% CI 0.13 to 0.22).[4]
A relative risk of 0.17 for death is not a plausible effect of a weight-loss drug over twelve months, and the authors say so themselves: the mortality difference is “unlikely to reflect a true causal effect” and should be treated as hypothesis-generating. Corticosteroid use in that same cohort did not move at all, at a relative risk of 1.00 (95% CI 0.96 to 1.03).[4] What the number is measuring is which patients get offered an elective weight-loss prescription, and which do not.
The one study built to behave like a trial found nothing
A 2026 target trial emulation set out to remove exactly that bias. It drew on a claims database, required patients to be stable — no steroids, no IBD-related hospitalization or surgery, and an unchanged IBD therapy for more than six months — and matched initiators of semaglutide or tirzepatide one-to-one against non-initiators. In the larger cohort of 2,028 matched pairs, one-year relapse occurred in 7.9% against 7.9%, a relative risk of 1.01 (95% CI 0.82 to 1.24), with safety outcomes at 7.0% against 7.9%. In the 346 pairs on advanced therapies or immunomodulators, relapse was 12.4% against 15.6%, relative risk 0.80 (95% CI 0.55 to 1.15). The stated conclusion is that adjunctive treatment does not improve outcomes in stable IBD.[5]
The same study recorded something a buyer should weigh separately: 36% and 33% of initiators stopped within one year.[5]A single-center series of 272 patients with IBD found adverse events in 40%, 93% of them gastrointestinal, and discontinuation in 24%, with nearly half of those stops attributed to tolerability.[6] A drug that a third of people put down inside a year is a different purchase from the one the annual pricing pages assume, which is what the twelve-month cost article works through.
A conclusion of “safe and effective” over an interval of 0.00 to 313.88
A second 2025 meta-analysis pooled 10 observational studies and 10,362 patients, 3,479 of them exposed. Pooled incidences among the exposed were 11% for hospitalization, 11% for corticosteroid use, 8% for treatment escalation, 14% for flare and 3% for surgery. Against non-users, the odds ratios were 0.93 for corticosteroid use (95% CI 0.17 to 5.16), 0.70 for treatment escalation (95% CI 0.06 to 7.62) and 0.32 for IBD-related surgery — on a 95% CI of 0.00 to 313.88, with I² of 92.5%. The paper’s conclusion is that these drugs “appear to be safe and effective.”[7]
An interval running from zero to three hundred is not a measurement of anything. It is a statement that the underlying studies disagree so completely that pooling them produced no information. Two other cohorts pointed the same non-direction from the opposite end: a 271-patient before-and-after analysis found no significant change in gastrointestinal adverse events, bowel surgery or IBD-related hospitalization across the year on either side of starting, at p ≥ 0.10 for every comparison,[8] and the 272-patient series found flares in 17% of the year before and 13% of the year after (P = 0.40).[6] Nothing got worse. Nothing was shown to get better either.
One head-to-head does separate the molecules. A propensity-matched analysis of 3,042 pairs found tirzepatide associated with less intravenous steroid use than the GLP-1 agonists, at an adjusted hazard ratio of 0.81 (95% CI 0.68 to 0.94), and a lower composite of steroids and surgery at 0.86 (95% CI 0.73 to 0.97), with hospitalization, emergency visits and intestinal surgery unchanged.[9] How the two molecules differ generally is in the tirzepatide and semaglutide comparison.
Why the biggest positive result is also the hardest to read
The largest reported benefit comes from a single tertiary center: 150 adults with ulcerative colitis starting semaglutide or liraglutide, matched to 150 who did not, with symptomatic remission at 12 weeks of 66.7% against 25.3% and an adjusted odds ratio of 5.90 (95% CI 3.52 to 9.86). Weight loss was not associated with remission.[10]
The endpoint was a partial Mayo score of 2 or less with a rectal bleeding subscore of zero — a symptom index, built largely from stool frequency and bleeding. Symptoms are a weak proxy for inflammation in this disease even before a drug is added: a meta-analysis of 27 studies and 3,169 patients with IBD in remission found IBS-type symptoms in 32.5% (95% CI 27.4 to 37.9), and still in 23.5% of those in endoscopic remission, rising to 36.6% in Crohn’s disease against 28.7% in ulcerative colitis.[11]
A drug whose defining effect is on gut motility and appetite moves the inputs to a stool-frequency score directly. The counterweight is inside the same paper: in the subset that underwent endoscopy, remission was 58% against 38% and endoscopic improvement 69% against 47%.[10] That is a real objective signal in a small subgroup of a single-center retrospective study, and it is the reason this result cannot simply be dismissed — only held open until 2028.
Telling a side effect from a flare
Nausea, abdominal pain, diarrhea, appetite loss and weight loss are on every one of these labels and are also how a flare announces itself. The overlap is the practical problem, and no study has measured how often it is resolved correctly.
What separates them in practice is objective testing rather than symptom description. Fecal calprotectin is the usual instrument, and its measured performance is good: pooled across 17 studies and 1,956 patients, sensitivity of 85.8% (95% CI 78.3 to 91) and specificity of 91.7% (95% CI 84.5 to 95.7).[12] Those figures come with a condition attached. They were measured on a different question — distinguishing IBD from irritable bowel syndrome at first presentation, against colonoscopy or radiology — and all included studies were rated at high or unclear risk of bias. Its accuracy for separating a drug adverse event from a flare in someone with an established diagnosis has not been established.
Rectal bleeding, fever, symptoms that wake someone at night, and a rising calprotectin or C-reactive protein point toward inflammation rather than tolerability, while symptoms that arrive within days of a dose increase and settle before the next one point the other way. That is clinical reasoning, not a trial result, and it is stated here as reasoning. The one thing that does not tolerate a wait-and-see approach is obstruction: a published case describes a 39-year-old man developing small-bowel pseudo-obstruction with ileitis five weeks after starting semaglutide, after four weekly 0.25 mg doses and one 0.5 mg dose, with no mechanical cause and resolution after the drug was stopped.[13] In stricturing Crohn’s disease the distinction matters more, and the motility evidence behind it sits in the gastroparesis article; the ordinary version of the same symptom is in the article on diarrhea.
The intestinal peptide is the other one
The mechanistic enthusiasm here partly borrows from a neighbor. Glucagon-like peptide-2, not peptide-1, is the gut-trophic hormone, and the GLP-2 analog teduglutide holds an approval for short bowel syndrome in patients dependent on parenteral support — read from its label rather than inferred. A 2026 review of both peptides as intestinal reparative therapies makes the necessary distinction explicitly: the human data on GLP-1 in IBD are observational, and separating a metabolic benefit from an intrinsic disease-modifying effect is the open question.[14] No GLP-1 receptor agonist has an approval in inflammatory bowel disease anywhere.
What this leaves at a checkout
A cash-pay service prescribes for weight, on a weight indication, with an intake built around eligibility and dosing rather than around a chronic inflammatory disease. Most of these sellers dispense compounded semaglutide or tirzepatide, and a compounded preparation does not hold FDA approval: the agency does not review it for safety, for efficacy or for quality before it reaches a patient. The distinction is worked through in what not FDA-approved actually means, and the sellers themselves are collected on the compounded semaglutide board.
The evidence supports a narrow set of statements. Weight comes off in people who have IBD at roughly the rate it comes off in people who do not. Nothing in the observational record shows the class making IBD worse, across more than 16,000 patients. Nothing in the one analysis designed to imitate a randomized trial shows it making IBD better. A third of people stop within a year, mostly because of gastrointestinal symptoms that in this population are harder to interpret than in any other. Whether a gastroenterologist should be told before a vial is ordered is not a close question, and the first randomized answer to everything else is three years away.