Question 1

Is GLP-3 safe to take?

Accepted Answer

That question cannot be answered for a name, because side effects are counted in people who received a particular molecule, and no molecule among these sources is called GLP-3. It can be answered for what the name usually stands in for, which the the GLP-3 overview page traces to retatrutide. A 2026 phase 3 report describes retatrutide as a triple hormone receptor agonist under clinical development for type 2 diabetes, obesity, and related complications. Under clinical development means it is still being tested and has not been approved for anything. A 2023 commentary on the early trial work put the open question plainly: the safety of retatrutide needs to be determined in larger and longer trials. So the answer comes in two parts. For retatrutide given inside a trial, a documented record exists, and the sections below set out what it holds. For a vial sold as GLP-3, no record exists at all, because no one can say which molecule it holds, and that is a different kind of risk rather than a smaller version of the same one.

Question 2

What side effects come with the triple agonist the name usually means?

Accepted Answer

The 2026 phase 3 trial randomly assigned 537 adults with type 2 diabetes to retatrutide at 4, 9 or 12 mg, or to placebo, as a once-weekly injection for 40 weeks. Its authors summed up the harms in one phrase, as an adverse event profile consistent with molecules with GLP-1 agonist activity. The 2023 phase 2 trial, with 281 participants, landed in the same place, reporting a safety profile consistent with GLP-1 receptor agonists and GIP and GLP-1 receptor agonists. Consistent with the class is a comparison rather than a count. It says the third receptor did not throw up a new kind of problem in those trials, and it says nothing about how often the familiar problems occurred, because neither abstract prints the rates. The counted rates for retatrutide itself are set out on the Retatrutide safety page.

Question 3

What side effects did the two-receptor molecules show?

Accepted Answer

The dual GLP-1 and glucagon agonists are the other molecules a GLP-3 label might stand for, and their safety data come from earlier stages still. SAR425899, an early molecule of this type, was tested in two placebo-controlled trials in healthy volunteers and in patients with type 2 diabetes, and the most frequently reported adverse events were gastrointestinal. One detail runs against expectation: gastrointestinal side effects were less pronounced in patients with type 2 diabetes compared with healthy volunteers. That matters for anyone reading a trial rate as a forecast. A rate measured in patients with diabetes may understate what a person without diabetes feels, at least for that molecule, and those trials were small enough that the point is a signal rather than a rule. The authors also left the larger question open, writing that whether dual agonism is superior to pure GLP-1R agonists for obesity and diabetes remains to be confirmed.

Question 4

When do the stomach side effects start, and do they pass?

Accepted Answer

A 2022 review of high-dose GLP-1 medicines and the newer co-agonists puts the gut side effects at 30% to 70% of patients, mostly arising within the first 2 weeks of the first dose. The same review describes them as being mild or moderate in severity, and transient, which means most of them passed rather than lasting. A 2025 review of the development pipeline adds that the nausea, vomiting and bowel problems often can be mitigated by slow up titration, meaning a slow climb to the target amount. Put together, the pattern is a front-loaded problem that the trials managed by pacing the dose. That is a description of what happened under a protocol, with material of known identity and strength, and it does not tell you how the first weeks go when the amount in each injection is itself uncertain. How fast the trials climbed, and what they climbed to, is laid out on the GLP-3 dosage page.

Question 5

How many people stopped the trials early?

Accepted Answer

In the retatrutide phase 3 trial, 490 of the 537 participants, or 91%, stayed on study drug to the end, and 94% completed the study. The phase 2 trial lost more people along the way, since 237 (84%) participants completed the study and 222 (79%) completed study treatment. Neither abstract says how many of those who stopped did so because of side effects, so the drop-outs cannot be split by reason here. Completion is still worth knowing. It is the share of people in whom every reported rate was measured to the end, and roughly one participant in five in the phase 2 trial did not finish on the injection. A side-effect rate counted among the people who stayed describes a group that had already tolerated the first weeks, which is weaker than it looks as a guide to anyone starting out.

Question 6

Could these molecules push blood sugar too low?

Accepted Answer

The glucagon receptor is the unusual part. Glucagon is the hormone that raises blood sugar, so a molecule that pulls on it is working partly against what a diabetes medicine usually does. A 2025 review sets out the usual GLP-1 side of that balance, writing that these agonists increase insulin release but suppress glucagon release. Why add glucagon at all is, in the words of a 2023 commentary on retatrutide, poorly defined and still needing to be clarified. How far sugar can fall on the newer molecules shows up in a 2022 review, which reports up to 62% of people with type 2 diabetes attaining normoglycaemia, a normal blood sugar, with 15-mg tirzepatide. The abstracts behind this page do not report how often blood sugar fell too low on retatrutide or the dual agonists. That is a gap in what these summaries print rather than evidence that it never happens, and it matters most for anyone already taking another medicine that lowers blood sugar.

Question 7

Do the dual and triple agonists affect the kidneys?

Accepted Answer

One study among these sources looked at kidneys directly. A 2025 analysis followed 247 participants with chronic kidney disease and type 2 diabetes given cotadutide, a dual GLP-1 and glucagon agonist, or semaglutide, or placebo. Its model predicted that protein leaking into the urine, measured as UACR, would fall by 45·6% against placebo after 26 weeks at the highest cotadutide amount in those participants. Against that sits a 2025 review that lists renal function disorders among the more severe side effects of the GLP-1 class, beside pancreatitis and allergic reactions. A model prediction is a forecast fitted to trial data rather than a measured outcome, and protein in the urine is a marker rather than kidney failure itself. The two readings do not cancel. One is a hoped-for benefit in diseased kidneys, the other a rare harm named across a whole class, and neither was measured in anyone holding an unlabelled vial.

Question 8

Why is the safety record for these molecules so thin?

Accepted Answer

Because they are new, and the slow questions about harm take years and thousands of people to answer. The longest trial among these sources is still running. The survodutide obesity trial in Japan is a 76-week, randomised, double-blind study, and so far it has reported its design and the baseline characteristics of the people enrolled. Its authors frame what it will add as Japan-specific efficacy, safety and tolerability data for survodutide, which is a promise about future results rather than a result. For the heart, a 2026 framework paper places retatrutide as an emerging triple-agonist candidate awaiting cardiovascular outcome data. Thin is not the same as worrying. It means the rare and the late effects have not had the time or the numbers to show up yet, in either direction, and a short clean record is a different thing from a long one.

Question 9

Whose bodies does the safety record describe?

Accepted Answer

Every rate above was measured in a particular group, and the groups were narrower than a reader might assume. The phase 2 retatrutide trial was run at centres in the USA, and 84% of its participants were White, with a mean age of 56. The phase 3 trial widened the map to 48 sites in the USA, Mexico, and India. A separate trial in Japan was set up to produce Japan-specific results, which makes the same point from the other side: a rate from one population is not automatically a rate for another. Young people are a near blank. A study of bone metabolism in 12 to 21 year olds on GLP-1 therapy is registered and still recruiting. None of this makes the molecules riskier for anyone outside those groups. It means the published figures were measured on somebody else, which is a different statement from a finding of harm.

Question 10

Do they interact with alcohol or other medicines?

Accepted Answer

Alcohol has the most direct data, and all of it comes from rats. A 2026 study trained male and female rats to recognise the feel of alcohol, and found that acute administration of semaglutide, tirzepatide, and retatrutide each attenuated alcohol discrimination in those animals. In plain terms, the alcohol registered less strongly in rats given any of the three, which is a finding about how alcohol feels to an animal rather than about harm. A 2025 review makes the wider point for the class, noting that because these agonists act on reward signalling, they could modify the effect of cocaine, alcohol and nicotine. Among medicines, metformin is the one the trials were built around, since the phase 2 retatrutide trial let people stay on a stable dose of metformin (≥1000 mg once daily). Everything else is untested among these sources rather than cleared, and a pairing no trial has examined is not a pairing with a known result.

Question 11

What is the safety record of the GLP-2 medicine?

Accepted Answer

Some readers meet GLP-3 next to GLP-2 and assume the two are neighbours. GLP-2 is a separate gut hormone, and its approved copy, teduglutide, works on the intestine rather than on appetite. Its safety record among these sources is a single case report, of a patient with parenteral nutrition-dependent short bowel syndrome and cardiovascular disease with 4-year exposure to teduglutide, who died of heart disease. Parenteral nutrition means feeding through a vein. One case does not show that the medicine caused the death, and the report sets out a history rather than a cause. Its intended use is narrow, and a 2024 gut-medicine guideline describes glucagon-like peptide-2 agonists as able to help people with short bowel syndrome move off long-term vein feeding. None of that is about weight, and a GLP-2 record says nothing about anything sold under a GLP-3 label.

Question 12

What does an unidentified vial add to the risk?

Accepted Answer

More than it seems, because the molecules this label could cover are given in amounts that differ by a factor of a hundred or more. Cotadutide was tested at 100, 300 or 600 μg, which are micrograms, thousandths of a milligram. Retatrutide was tested at 4 mg, 9 mg, or 12 mg, which are whole milligrams. A vial that holds one molecule while the person using it believes it holds another is therefore not a small error. It is the difference between a studied amount and one that no trial among these sources gave, rather than a rounding problem. Manufacture adds a second layer. Even for teduglutide, a licensed peptide, a chemistry paper describes its synthesis as made challenging by extensive aspartimide formation, a by-product of how such chains are built. A vial outside any licensed system has no batch test anyone can look up, which is why identity, rather than the side-effect list, is the first safety question for anything sold as GLP-3.

GLP-3 side effects and safety data