Two published figures describe vomiting on tirzepatide. Its pooled relative risk across 23 randomized trials is 13.23 — the largest estimate in the entire gastrointestinal panel of that analysis, three times semaglutide’s.[1] Its labeled incidence at the maximum dose is 13%, against 24% on the semaglutide label.[2][3] One number says tirzepatide is the far worse choice for this symptom; the other says it is roughly half as likely to produce it. Neither is wrong, and a page that quotes only one of them is describing a drug that does not exist.
What each trial program recorded
The semaglutide 2.4 mg weight-reduction pool put vomiting at 24% of 2,116 treated adults against 6% of 1,261 on placebo. The narrative paragraph in the same section of the same label prints 25% against 6% — a one-point internal disagreement worth knowing about before treating any single published percentage as exact.[2] The tirzepatide pool recorded 8%, 11% and 13% across 5, 10 and 15 mg against 2% on placebo.[3]
The other labels fill in the class. Liraglutide 3 mg: 15.7% of 3,384 treated adults against 3.9% of 1,941 on placebo.[4] Oral semaglutide in diabetes: 6% at 7 mg and 8% at 14 mg against 3% on placebo.[5] Tirzepatide in diabetes: 5%, 5% and 9% across the three doses against 2%.[6] And in the pediatric semaglutide trial, 133 adolescents against 67 on placebo, vomiting was reported by 36% against 10% — half again the adult rate on the same drug at the same dose.[2]
Those placebo arms are the reason the ratio and the count disagree. Six percent of people taking nothing vomited over 68 weeks in the semaglutide program; 2% did in the tirzepatide program. Dividing one label by the other is not a comparison, which is why the molecule question belongs in the head-to-head write-up rather than in a subtraction between package inserts.
This is the effect that does follow the dose
Vomiting behaves the way titration advice assumes all gastrointestinal effects behave, and it is close to the only one that does. It rises monotonically across every dose-stratified table above: 8, 11, 13 on tirzepatide for weight; 5, 5, 9 on tirzepatide for diabetes; 6 then 8 on oral semaglutide. The higher-dose semaglutide trials extend the line — against 6% on placebo, vomiting was reported by 16% at 2.4 mg and 22% at 7.2 mg weekly.[2]
Nausea in the same tirzepatide table does not do this: 25%, 29%, 28%, peaking at the middle dose. Constipation actively falls. Vomiting is the symptom for which “a smaller dose means less of it” is supported by the tables rather than assumed, which makes it the one most worth raising before a scheduled escalation. The schedule those decisions attach to is in the titration article, and the nausea that usually arrives first has its own time course in the nausea write-up.
The largest pooled effect, on the weakest evidence grade
An umbrella review published in 2026 reanalyzed 60 meta-analyses covering 1,751 randomized trials and 3,580,616 participants under random effects. Vomiting came out at an odds ratio of 2.78 (95% CI, 1.91 to 4.06) — the largest of the three cardinal gastrointestinal effects, ahead of nausea at 2.47 and diarrhea at 1.94. It is also the only one of the three graded moderate rather than high quality of evidence.[7] The biggest number in the set rests on the least certain footing, and the reviewers note that heterogeneity and the prediction intervals leave real residual uncertainty on top of that.
Agent by agent, the network meta-analysis of 33,354 adults without diabetes ranks vomiting risk as tirzepatide 13.23 (95% CI, 4.85 to 36.09), semaglutide 4.21 (3.58 to 4.95), liraglutide 3.87 (3.15 to 4.76) and orforglipron 3.87 (1.45 to 13.56). Cagrilintide at 1.37 (0.81 to 2.32) and exenatide at 4.52 (0.25 to 82.77) did not reach significance.[1] Tirzepatide’s interval runs from 4.85 to 36.09, which is a width that should temper the headline: what the analysis establishes is that the increase is real, not that it is thirteen-fold.
The dehydration pathway, and the evidence that complicates it
Both labels carry a warning for acute kidney injury due to volume depletion and state that the majority of reported events occurred in patients experiencing gastrointestinal reactions leading to dehydration. In the pooled tirzepatide weight trials, acute kidney injury was reported in 0.5% of treated patients against 0.2% on placebo.[3] That is the mechanism worth taking seriously: repeated vomiting removes fluid and electrolytes from someone whose intake has already fallen, and the people with least reserve are those covered in the kidney article.
Two datasets complicate the population-level version of that story. A global retrospective cohort drew 3,729,925 adults with obesity from the TriNetX network and propensity-matched 12,123 new GLP-1 users against 12,123 untreated controls, all without type 2 diabetes. Over up to five years, treatment was associated with a lower risk of acute kidney injury, alongside lower all-cause mortality (hazard ratio 0.23; 95% CI, 0.15 to 0.34) — a figure large enough to signal the residual confounding an observational design cannot remove.[8]
A disproportionality analysis of 133,872 adverse event reports adds the molecule split, and it runs opposite to the vomiting ranking. Acute kidney injury appeared in 0.47% of 92,807 tirzepatide reports and 1.07% of 41,065 semaglutide reports, a reporting odds ratio of 0.44 (95% CI, 0.38 to 0.50) for tirzepatide against semaglutide.[9] The molecule with the largest vomiting ratio carries the smaller renal signal.
None of that lowers the individual risk. A cohort average describes people who mostly tolerated the drug; the warning describes what happens to someone in the middle of three days of vomiting. Those are different questions, and only the second one is actionable at 2 a.m. What to do during an acute illness on top of it is set out in the sick-day article.
When it stops being a side effect
Discontinuation is the measurable version of that threshold. In the semaglutide weight-reduction trials, 6.8% of treated patients and 3.2% on placebo permanently stopped for an adverse reaction, with vomiting specifically accounting for 1.2% against 0%.[2] In the tirzepatide trials, discontinuation for gastrointestinal reactions ran 1.9%, 3.3% and 4.3% across the three doses against 0.5% on placebo, and the label states that the majority of nausea, vomiting and diarrhea occurred during dose escalation and decreased over time.[3] The tirzepatide diabetes program reported the same gradient at 3.0%, 5.4% and 6.6%.[6]
So roughly one person in a hundred stops semaglutide because of vomiting, and between two and seven in a hundred stop this class for a gastrointestinal reason overall. That is the empirical answer to how often this becomes untenable, and it is low — which also means that persistent vomiting is not the expected course and does not need to be endured to prove commitment. A dose held or reduced is the intervention the tables support.
What compounded product changes about this
Every percentage above was produced under randomization, with a known molecule at a known dose. Most sellers covered here dispense compounded semaglutide or tirzepatide, which is not FDA-approved and is not reviewed by the FDA for safety, efficacy or quality before it is dispensed, and no randomized trial has measured a vomiting rate in one. A pharmacovigilance analysis of 81,078 GLP-1 reports, 707 of them compounded, found higher reporting odds of nausea at 1.27 (95% CI, 1.05 to 1.52) and of hospitalization at 2.35 (1.94 to 2.83); vomiting is not among the reactions it reports as elevated.[10] The same analysis found reporting odds of 48.92 for preparation errors and 8.51 for compounding or manufacturing issues, which is the relevant risk when the question is whether a vial contains the dose it claims.
Where the line sits
Vomiting on this class is dose-driven, concentrated in the escalation weeks, reported by 13% to 24% of adults at maintenance doses depending on the molecule, and by 6% of people taking placebo. It leads about one patient in a hundred to stop. The broader tolerability picture is in what the trials recorded.
It stops being a side effect and becomes a reason to stop when fluids cannot be kept down for a day, when it produces lightheadedness on standing or a stretch without urinating, when blood appears, or when severe abdominal pain radiating to the back accompanies it — that last combination is the pancreatitis presentation both labels describe, and it is a same-day call rather than a wait-and-see.