Gallstones are the side effect with the readiest explanation attached. Lose weight quickly, the story goes, and bile chemistry shifts; the gallbladder is collateral damage from the diet, not from the drug. It is a tidy account, it appears on most pages that cover this, and the three FDA labels for weight-management GLP-1s do not support it.
Every figure below was read on 2026-09-11 from labels published through openFDA: Zepbound effective 2026-08-28, Wegovy effective 2024-04-23, and Saxenda effective 2026-02-25. All three carry acute gallbladder disease as a warning. What they do not carry is the same number.
The three labels, side by side
| Product | Molecule | Cholelithiasis, drug vs placebo | Cholecystitis, drug vs placebo |
|---|---|---|---|
| Saxenda | liraglutide 3.0 mg daily | 2.2% vs 0.8% | 0.8% vs 0.4% |
| Wegovy | semaglutide 2.4 mg weekly | 1.6% vs 0.7% | 0.6% vs 0.2% |
| Zepbound | tirzepatide, weekly | 1.1% vs 1.0% | 0.7% vs 0.2% |
Read the cholelithiasis column against what each product does to body weight and the ordering is upside down.
Saxenda produces the smallest weight reduction of the three by a wide margin. In the one trial that put it head to head against semaglutide, liraglutide 3.0 mg produced a mean change of −6.4% at week 68 against −16.4% on semaglutide 2.4 mg — figures we set out in full in the semaglutide and liraglutide comparison. Saxenda is also the product with the largest gallstone excess: nearly three times the placebo rate, with the label adding that most of the patients who had these events required surgery to remove the gallbladder.
Zepbound produces the largest weight reduction of the three. Its cholelithiasis figure is 1.1% against 1% on placebo. That is not a small excess; it is not an excess.
If rapid weight loss were driving gallstone formation, the column would run the other way.
The manufacturers reach opposite conclusions, in writing
The labels do not merely report different numbers. They report different causal readings, and each one is stated in the manufacturer's own words.
Novo Nordisk, in the Wegovy label and again in the Saxenda label, writes that substantial or rapid weight loss can increase the risk of cholelithiasis — and then immediately rules it out as the whole explanation: the incidence of acute gallbladder disease was greater in treated patients than in placebo-treated patients even after accounting for the degree of weight loss. That sentence appears in both documents.
Eli Lilly, in the Zepbound label, writes the opposite about its own product: acute gallbladder events were associated with weight reduction.
It would be easy to read this as two companies spinning the same finding in different directions. The numbers say otherwise. Each statement fits the data in the document that carries it. Novo Nordisk is describing products with a real excess over placebo that survives adjustment. Lilly is describing a product whose treated and placebo arms reported cholelithiasis at 1.1% and 1%, where there is nothing left for the drug to explain. Both readings are defensible on their own evidence, which is precisely why the disagreement is informative rather than promotional.
What that leaves unexplained
The honest position is that the pattern across three labels is inconsistent with the simple weight-loss story, and that nothing in these documents replaces it with a better one.
Two features complicate any neat account. The first is cholecystitis — inflammation rather than stones — where the three products converge rather than diverge. Every label reports roughly three times the placebo rate: 0.8% against 0.4%, 0.6% against 0.2%, 0.7% against 0.2%. Whatever produces stones behaves differently across these drugs; whatever produces inflammation appears to behave similarly. Zepbound is the clearest case: essentially no excess in stones, a three-and-a-half-fold excess in inflammation, in the same trial pool.
The second is age. In the single Wegovy trial in patients aged 12 and older, cholelithiasis reached 3.8% against 0% on placebo — more than double the adult figure in the same label. Novo Nordisk states the conclusion directly rather than leaving it to the reader: the incidence was higher in treated pediatric patients than in treated adults. It rests on one trial, and a placebo arm that reported none at all makes the ratio unstable, but it is the label's own conclusion and not an inference drawn here.
Why the ordering may not be a paradox at all
There is a mechanism that would fit, and it is worth stating as a hypothesis rather than a finding, because none of these three documents tests it.
Gallbladder emptying is regulated in part by the same gut hormone signalling these drugs act on, and the three products do not act on it identically. Tirzepatide is a dual agonist, working at both the GIP and the GLP-1 receptor; liraglutide and semaglutide act at GLP-1 alone. Liraglutide is dosed daily and semaglutide weekly, which produces very different concentration profiles across a week. Any of those differences could plausibly matter more to the gallbladder than the number on the scale.
What can be said from the labels is narrower and firmer: the excess does not track weight loss, two manufacturers say so explicitly for their own products, and the drug with the biggest effect on weight has the smallest effect on stones. The mechanism behind that is not in these documents, and this page will not invent one.
The limitation that applies to all of it
Every figure here is an adverse event reported during a clinical trial, not a systematic imaging study. Nobody scanned these participants at fixed intervals looking for stones. Asymptomatic gallstones are common in the general population and would not appear in any of these counts, which means all six numbers are floors rather than incidences, and the gap between the treated and placebo arms is the only part designed to be compared.
That caveat cuts the same way in every row of the table, which is what makes the comparison across rows worth making — and what stops any single number in it from being a risk estimate for an individual person.
For the broader picture of what these trials counted, our side-effects hub sets out the gastrointestinal figures across the same labels, and the muscle loss page covers the outcome these documents measure least well of all.