GLP-1 and Abdominal Pain: Why the Label Says 20% and the Trial Registry Says 9.7%

Ask "does Ozempic cause stomach pain?" and the answer that comes back is usually a number from the prescribing information: 20% of adults on Wegovy 2.4 mg reported abdominal pain in the weight-reduction trials, against 10% on placebo. Both halves of that sentence are true. What almost never comes with it is a third figure from the same trial programme: STEP 1, the largest of the three trials the label pools, posts abdominal pain on ClinicalTrials.gov at 9.7% on semaglutide against 5.5% on placebo. Same drug, same dose, the same 68 weeks, in a trial that is part of the label's own pool — and half the percentage. The gap between 20% and 9.7% is not a disagreement between two sources. It is the difference between a row that groups seven related terms and a row that counts one, and it is the first thing to understand before quoting either.

This is the tenth piece in our evidence series, after which effects have enough evidence to calibrate a model, the corpus-derived pooled rates, what patients themselves report, gallstones end to end, what is inside a personalised risk number, hair loss end to end, why trial vomiting rates run from 6.5% to 58%, constipation read against the placebo arm and headache read against the placebo arm. Like those, every figure below is quoted from a primary source with the arm and the number of participants it belongs to, and every one of them was read on 16 September 2026 — from the current US prescribing information, from each trial's own results table on ClinicalTrials.gov, or from our own live production API.

What the label says, and what its footnote says

The current US prescribing information for Wegovy (semaglutide), revised June 2026, lists abdominal pain among the *"most common adverse reactions (incidence ≥5%) in adults or pediatric patients aged 12 years and older"*. Its three adverse-reaction tables each give a drug column and a placebo column:

Table 3 is the source of the 20%. Its footnote is the part that matters: the row *"includes abdominal pain, abdominal pain upper, abdominal pain lower, gastrointestinal pain, abdominal tenderness, abdominal discomfort and epigastric discomfort"* — seven separate adverse-event terms, counted as one. A person who reported any of them appears once in the 20%. Table 4 carries the identical footnote. Further down the same table, *abdominal distension* — bloating — is listed separately, at 7% against 5%, and is not part of the abdominal-pain group. (Quoted from the prescribing information via DailyMed, set ID ee06186f, version 19, effective 18 June 2026.)

For comparison, the same Table 3 puts nausea at 44% against 16% on placebo, vomiting at 24% against 6%, constipation at 24% against 11% and headache at 14% against 10%. Of the five effects this series has now read against the placebo column, abdominal pain is the one where the placebo figure is exactly half the drug figure — closer than nausea, vomiting or constipation, further apart than headache.

What the trial registries say, arm by arm

Trial registries do not use the label's grouping. ClinicalTrials.gov posts each MedDRA preferred term as its own row, so "abdominal pain", "abdominal pain upper" and "abdominal pain lower" are three rows, and a participant who reported two of them is counted in two. Our collector tracks the single term "abdominal pain". That is why STEP 1 reads 9.7% in the registry and sits inside a 20% in the label: the registry row is one of the seven terms the label adds together. In the same STEP 1 record, "abdominal pain upper" is posted beside it at 125 of 1,306 (9.6%) against 35 of 655 (5.3%) — nearly the same size — and since one person can be in both rows, the two single terms do not simply add to the grouped figure; each is a floor on it.

Our database cited 16 distinct studies for a stated abdominal pain rate when this article was written at 21:50 BST on 16 September 2026. Eight of those 16 post a placebo arm of at least 20 people; every row below was read from that trial's own ClinicalTrials.gov adverse-events module on 16 September 2026, from its non-serious events table, for the single term "abdominal pain". The counts are the registry's; the percentages and the differences are our division and subtraction.

The other eight sources in our abdominal pain base are absent from this table for two different reasons. Three are placebo-controlled but split their placebo across parts of 2 to 15 people, which is too few to read a difference from: two single- and multiple-ascending-dose studies of orforglipron, in healthy volunteers (NCT03929744, placebo parts of 8, 15 and 2; 21 drug arms from 0.0% to 16.7%) and in Japanese participants with type 2 diabetes (NCT05086445, placebo parts of 4 and 11; one drug arm at 1 of 14, the other four at 0); and a 24-person dose-titration study in Chinese participants (NCT06023095, placebo arm of 4), which reports 3 of 10 in each of its two orforglipron cohorts against 0 of 4 — the highest percentage in the base after the appetite study's 31.9%, resting on six people. Five post no placebo arm at all: a tablet-versus-capsule bioequivalence study of orforglipron in 533 adults with obesity or overweight (NCT06440980, 38 dose-level arms from 0.0% to 6.5%), a 52-person multiple-dose formulation study (NCT05841238, 18 dosing periods from 0.0% to 7.3%), the long-term safety study ACHIEVE-J (NCT06010004, 0.8%, 1.5% and 2.2% across its three doses), ACHIEVE-3 (NCT06045221), where orforglipron 12 mg and 36 mg read 21/424 = 5.0% and 23/423 = 5.4% against oral semaglutide 7 mg and 14 mg at 13/426 = 3.1% and 14/425 = 3.3%, an active comparator rather than a placebo, and a drug-interaction study of orforglipron with carbamazepine (NCT06370728, 0/30 on orforglipron alone and 1/28 on the combination; its carbamazepine-only period was withheld from our base earlier today, as the headache article describes).

Three of the largest trials of the class are not in our abdominal pain base at all, and the reason is the same one the headache article found. ClinicalTrials.gov's non-serious adverse-event table lists only terms that reached the trial's reporting threshold, 5% for all three. STEP 2 (NCT03552757, 1,210 people, semaglutide in weight management with type 2 diabetes) posts no non-serious abdominal pain row, so the single term did not reach 5% in any of its arms — while sitting inside the label's pooled 20%; its serious-events table lists 3 of 402 on 1.0 mg, 0 of 403 on 2.4 mg and 0 of 402 on placebo. SELECT (NCT03574597, 17,604 people, semaglutide 2.4 mg against placebo in a cardiovascular-outcomes population) posts abdominal pain only as a serious event, 9 of 8,803 on semaglutide against 18 of 8,801 on placebo. SURPASS-CVOT (NCT04255433, tirzepatide against dulaglutide in 13,299 people with type 2 diabetes) likewise, 8 of 6,647 against 10 of 6,647. Under our eligibility rules a serious-events row is admitted, as a floor on incidence, only for an effect the registry posts in no other table anywhere in the corpus, which abdominal pain is not, so none of the three contributes a rate.

The difference, not the headline

In the adult weight-management trials of injectable semaglutide 2.4 mg, the grouped term adds ten points and the single term adds four. Table 3 (20 − 10 = 10) and STEP 1 (9.7 − 5.5 = 4.2, with "abdominal pain upper" beside it at 9.6% against 5.3%, a 4.2-point gap) are not two independent readings — STEP 1 is one of the three trials Table 3 pools — so the point is that the size of the difference depends on how many terms you count, not on which trial you read. Quote 20% against 10% and you are describing any pain or discomfort anywhere in the abdomen; quote 9.7% against 5.5% and you are describing the one term coded as "abdominal pain". Neither is wrong. Citing one as if it were the other is.

In the tirzepatide and orforglipron Phase 3 trials the single-term gap is one to four points. SURMOUNT-1's three tirzepatide doses sit 1.5 to 2.3 points above their own placebo arm, and the 15 mg dose reports slightly *less* abdominal pain than the 5 mg or 10 mg dose. Across the nine orforglipron Phase 3 arms in three trials the difference runs from 0.7 points below placebo (ACHIEVE-1's 3 mg arm) to 3.5 above (ATTAIN-2's 12 mg and ACHIEVE-1's 36 mg), with the Japanese trial's three doses within 1.6 points of placebo. Four of the eight placebo-controlled sources in the table report at least one drug arm at or below its own placebo: ACHIEVE-1, ATTAIN-J, the retatrutide study and the adolescent oral-semaglutide trial, where the two arms are identical at 6 of 66.

One row sits well above the rest, and we are not going to smooth it into the picture. The 120-person appetite-and-eating-behaviour study reports 31.9% against 12.5% — a 19-point gap, in a trial whose own title says it was studying appetite and eating behaviour, with 72 people on the drug and 48 on placebo. Its second phase, weeks 60 to 72, reports 0 of 10 on continued semaglutide and 0 of 20 on continuous placebo. It is in the table because it is in our base, and a table that left it out would be quieter than the evidence.

So the honest one-line answer to "how much abdominal pain does the drug itself add?" is: in every placebo-controlled trial arm of more than 300 people that posts an abdominal pain rate, the drug arm sits between 1.5 and 4.2 points above its own placebo arm — seven arms across STEP 1, SURMOUNT-1 and ATTAIN-2, and unlike headache, none of them sits below. That is a real, consistent, modest excess for the single term. The label's ten-point excess for the grouped term is also real. Which one you should quote depends on whether the question is "pain coded as abdominal pain" or "any abdominal pain or discomfort", and the honest citation names the term.

Our own pooled estimate, and why it is not the Wegovy number

Our abdominal pain endpoint published a pooled clinical estimate of 6.5% (95% CI 1–42%), from 115 eligible stated rates across 16 distinct studies, source diversity "high", when this article was written at 21:50 BST on 16 September 2026. It is a third of the label's 20% for two reasons that stack: the pool is the whole class, not one drug at one dose in one population, and every registry row in it is the single term, not the label's group of seven.

Our pooled figure weights each source by its own stated sample size (a source stating none counts as 1), multiplied by a discount for how confident our extraction was in reading the figure, after collapsing each study to one entry — the unweighted mean of its posted arms, carried at the size of its largest arm. So the entries with the most weight are STEP 1 (9.7%, n=1,306), SURMOUNT-1 (5.7%, n=636), ACHIEVE-3 (4.2%, n=426), ATTAIN-2 (5.7%, n=331) and the 533-person bioequivalence study (1.7%, n=215), and beside them sit six Phase 1 pharmacology studies of orforglipron — dose-escalation, formulation and drug-interaction studies dosed for days rather than months, in healthy volunteers or small cohorts — which are six of the 16 studies, contribute 86 of the 115 stated rates, and span 0% to 30% across arms of 6 to 215 people. The two highest entries in the pool are also two of its smallest — the appetite study (16.0%, the mean of its 31.9% and 0% phases, n=72) and the 24-person Chinese study (30.0%, n=10) — and the interval's width is the between-study spread those two entries create: the other fourteen entries run from 1.3% to 9.7%. Whether Phase 1 arms of that kind belong in a pooled incidence estimate at all, or should be kept and labelled, is a methodology question we have open at the time of writing; for now they are in, and the live endpoint names every one of them so you can take them out yourself.

If you want the Wegovy 2.4 mg number, quote the label: 20% against 10% on placebo, N=2,116 and N=1,261, Table 3, seven terms grouped. If you want the single registry term for one trial, quote STEP 1: 9.7% against 5.5%. If you want the whole class with its heterogeneity visible, quote ours, with its interval and its source count. None of the three is the other two.

What the FDA's adverse-event reports show

A different kind of number sits outside the incidence pool on purpose. FDA FAERS publishes the share of *spontaneous adverse-event reports* for each drug that mention a term; the term our collector files under abdominal pain is "abdominal pain upper", the most-reported of the group: 2,951 of 99,460 semaglutide reports (3.0%), 1,425 of 51,432 liraglutide reports (2.8%), 2,428 of 103,767 dulaglutide reports (2.3%), 3,709 of 182,230 tirzepatide reports (2.0%) and 36 of 1,154 orforglipron reports (3.1%), as read from our own endpoint on 16 September 2026. A share of complaints is not a share of patients — it answers "of the people who filed a report about this drug, how many mentioned upper abdominal pain," not "how many people on the drug got it" — so our eligibility rules report these separately rather than average them in. Our endpoint also lists orforglipron's "abdominal distension" and "abdominal discomfort" shares under the same effect; each row names its own term, and we quote only the pain term here.

What patients themselves report

Of the 26 distinct community reports in our corpus, 4 mention abdominal pain — a 15.4% reporting frequency. That is not an incidence rate and we never treat it as one: it is the share of people who chose to post about their GLP-1 experience who mentioned abdominal pain at all, and at four reports it is a count, not a rate. See that article for why the pool is small and cannot currently grow.

Correction, 16 September 2026 — one row withheld before this article was published

Reading every source of the 17 our base cited this evening turned up one whose rate is not an abdominal pain rate. The only PubMed abstract in the base, PMID 37062422 — a placebo-controlled single-ascending-dose Phase 1 study of TG103, an investigational GLP-1/Fc fusion protein, in 32 healthy Chinese subjects, 24 of them dosed — lists among treatment-related adverse events decreased appetite (41.7%), nausea (20.8%) and *abdominal distension* (12.5%). It states no abdominal pain figure. Our extractor had stored the 12.5% as an abdominal pain rate on 16 August, and it had sat in the base for a month. Abdominal distension is bloating, a separate adverse-event term — the label above lists it as its own row, outside the abdominal-pain group — so the row was withheld from both our stores at 21:49 BST today, before this article was published. The abstract's nausea and decreased-appetite figures are genuinely stated and stay. The effect on published figures: abdominal pain 116 → 115 stated rates, 17 → 16 distinct studies, 6.6% → 6.5%, interval 1–42% unchanged; site-wide 1,125 → 1,124 eligible rates from the same 30 studies. No guard caught it, because nothing checks that the sentence a rate was read from names the effect it is filed under; that check is the next thing to build, and the correction is described here rather than quietly applied because a number filed under the wrong term is exactly what a hostile reader would ask about first.

If you are on a GLP-1 now

This is educational content, not medical advice. The label lists abdominal pain among the most common adverse reactions, at rates above placebo in every table. Two other places it appears in the prescribing information are worth knowing rather than paraphrasing, because they describe a different kind of pain from the common one. Under acute pancreatitis, the label instructs clinicians to *"observe patients carefully for signs and symptoms of acute pancreatitis, which may include persistent or severe abdominal pain (sometimes radiating to the back), and which may or may not be accompanied by nausea or vomiting"*, and to discontinue if pancreatitis is suspected. Under gallbladder problems, the patient information tells you to call your healthcare provider for *"pain in your upper stomach (abdomen)"* together with *"yellowing of skin or eyes (jaundice)"*, *"fever"* or *"clay-colored stools"*. Abdominal pain that is persistent or severe, that radiates to the back, or that arrives with fever or jaundice is a reason to contact a clinician rather than to wait for it to settle. Nothing here should be used to start, stop or change a medication or its dose without your prescriber.

Every figure in this article was read on 16 September 2026 from the cited primary source or from our live production API, and is stated as of that date. Query everything behind it — all 15 effects, both tracks, every source with its URL — at magistra.health/en/data-api, with the full eligibility methodology at magistra.health/en/methodology.