Are GLP-1 medications safe?

The short answer

For the FDA-approved medicines, taken by the people they were approved for, the safety profile is well characterised and mostly reassuring. That is a narrower sentence than "GLP-1s are safe," and the narrowness is the point.

Three things are true at once, and most articles only tell you one of them.

The first is that these drugs have a longer track record than the news cycle suggests. Exenatide, the first GLP-1 receptor agonist, was approved in 2005. Two decades of diabetes prescribing sits underneath the obesity indications that arrived in 2021.

The second is that the serious risks are not zero and not vague. There is a boxed warning. There is a rare eye condition that European regulators added to the label in 2025. There is a gastrointestinal signal that one large claims analysis put at a nine-fold increase in pancreatitis. Every one of those has a shape you can learn.

The third is the one that gets the least coverage and causes the most measurable harm: a large fraction of GLP-1 injections in the United States over the past two years did not come from an approved pen. They came from a compounded vial and a syringe, and the FDA has spent that entire period publishing alerts about people overdosing themselves by a factor of five to twenty because milligrams, millilitres and syringe units are three different things.

What the biggest safety trial actually measured

SELECT was designed to answer whether semaglutide prevents cardiovascular events, and it also functions as the largest long-term safety dataset for the 2.4 mg dose. It enrolled 17,604 adults with established cardiovascular disease and a BMI of 27 or above, none of whom had diabetes, and followed them for a mean of 39.8 months.

Read that hazard ratio carefully, because it is routinely mangled. A 20% relative reduction here means 1.5 fewer events per 100 people over roughly three and a half years. That is a genuinely good result for a preventive medicine. It is not "20% less likely to have a heart attack this year."

The safety relevance is that a trial this size, running this long, is where uncommon harms surface. Nothing in SELECT stopped the drug from being recommended. That is meaningful evidence, and it is evidence that did not exist in 2021.

Common side effects, with the real numbers

The common side effects are gastrointestinal, they cluster during dose escalation, and they cause far fewer people to quit than the raw percentages suggest. These figures come from the FDA-approved Wegovy label, covering 2,116 adults treated for up to 68 weeks.

The gap between "44% felt nauseated" and "1.8% quit over it" is the most useful number on this page. It tells you that for most people the nausea is real, unpleasant and survivable, and that it usually settles.

Two practical patterns show up repeatedly and are worth knowing before you start. Symptoms concentrate in the first 48 to 72 hours after an injection, and they spike again after each dose increase rather than accumulating steadily. If your worst days are predictable, you can plan around them. Our side effect timeline guide breaks down which symptoms show up in which week, and which ones are not normal at all.

Serious risks: proven, probable, unresolved, retired

Most safety coverage flattens every concern into "may cause," which is useless for deciding anything. The table below sorts the named risks by how strong the evidence actually is, which is the distinction that changes what you do.

The pattern worth noticing: the risks with the loudest media coverage (suicide, pancreatic cancer) are the ones that regulators investigated and largely closed. The risks that actually changed labels are quieter and more specific.

Who should not take a GLP-1

The contraindications are short, specific, and non-negotiable. They are the part of the safety conversation with the least ambiguity, and the part a telehealth intake form is most likely to skim.

• Personal or family history of medullary thyroid carcinoma, or MEN 2. This is a hard contraindication on the label, not a caution.
• Pregnancy, or planning pregnancy. Semaglutide should be stopped well before conception; discuss timing with your clinician, because the washout is measured in weeks, not days.
• Prior pancreatitis. Not an absolute bar, but it changes the risk calculation and needs an explicit conversation.
• Known hypersensitivity to the drug or its excipients.
• Type 1 diabetes without specialist supervision, where these are used off-label and require close monitoring.

Beyond the contraindications sit the interactions that matter more than people expect: insulin and sulfonylureas, where the combination can produce hypoglycaemia and doses usually need adjusting; and anything with a narrow therapeutic window taken orally, where slower gastric emptying changes absorption timing.

If you want to check the standard criteria before an appointment, our eligibility checker walks through the BMI and comorbidity thresholds used for the approved indications.

The muscle loss question

Losing lean mass alongside fat is a feature of large, rapid weight loss in general, not a unique toxicity of GLP-1 medication. The STEP 1 body composition substudy used DEXA scanning and found total fat mass fell 19.3% while the proportion of lean mass to total body mass rose — body composition improved overall, even though lean tissue was part of what was lost.

That nuance keeps getting lost in both directions. "You lose 40% muscle" and "muscle loss is a myth" are both wrong. What is true: a meaningful share of the scale drop is lean tissue, the ratio is broadly what you would expect from any comparable weight loss, and it is the most modifiable risk on this entire page.

Two levers, both boring, both effective:

• Protein. Harder than it sounds when appetite has gone quiet, which is exactly why it slips. Our protein target calculator gives you a daily gram number based on weight and activity.
• Resistance training. Two sessions a week is the threshold where the evidence gets consistent. This is not about becoming an athlete; it is about giving the body a reason to keep the tissue.

Appetite suppression makes protein the first thing to fall and the last thing anyone notices falling. It is worth tracking deliberately rather than estimating.

The part of this market that is genuinely dangerous

If you are trying to work out where GLP-1 risk is concentrated right now, it is not in the pen. It is in the vial. This is the section that most safety articles skip, and it is the one with the clearest documented harm.

During the 2022–2024 shortages, US law allowed compounding pharmacies to produce copies of semaglutide and tirzepatide. Volume exploded. So did the failure modes, and the FDA has been publishing them ever since.

Dosing errors, at scale

In July 2024 the FDA alerted providers and patients that people were injecting five to twenty times the intended dose of compounded semaglutide. The cause was not exotic. Patients were being handed a multi-dose vial and a syringe and asked to convert between milligrams, millilitres and insulin syringe units — three units of measurement that look interchangeable and are not. By early 2025 the agency had logged more than 455 adverse event reports tied to compounded semaglutide.

Salt forms that are not the drug

The FDA has received reports of compounders using semaglutide sodium and semaglutide acetate. These are different chemical entities from the semaglutide base in the approved products. They have not been shown to be safe or effective, and there is no reason to assume the dose equivalence carries over.

"Research peptides" are not a legal grey area, they are unregulated

Vials sold as research chemicals, not for human consumption, carry no requirement for identity, purity, sterility or potency testing. Nobody is checking what is in them. This is the category where contamination and mislabelled concentration are not hypothetical.

Where the rules landed

On 30 April 2026 the FDA proposed permanently excluding semaglutide, tirzepatide and liraglutide from the 503B Bulks List — the list of ingredients outsourcing facilities are permitted to compound from — and has sent warning letters to companies making false or misleading claims about compounded GLP-1 products.

None of this means every compounded vial is dangerous or that everyone using one is at risk. It means the safety question splits in two. "Is semaglutide safe" and "is the specific product in my fridge what the label says it is" are different questions, and only the first one has a trial behind it. We cover the sourcing question in more depth in are GLP-1s a scam? and in our guide to peptides and unsupervised use.

Stopping is a safety event too

Discontinuation has its own risk profile, and it is the one people plan for least. In the STEP 1 trial extension, participants who came off semaglutide after 68 weeks regained roughly two-thirds of the weight they had lost over the following year, and the cardiometabolic improvements moved back toward baseline alongside it.

For someone using a GLP-1 for weight management, that is a disappointment. For someone using one for type 2 diabetes or cardiovascular risk reduction, it is a clinical event: blood glucose rises again, and the protective effect measured in SELECT is not something you keep after you stop.

The practical failure is stopping abruptly and without a plan — usually because of cost, supply or side effects rather than a clinical decision. If any of those are pushing you toward stopping, that is a conversation to have before the last dose, not after. Our Ozempic transition guide and the weight regain field guide cover what to track through that window.

What to ask your prescriber

The single highest-value thing you can do for your own safety is turn "is this safe" into five specific questions. Bring these:

1. Given my history — thyroid, pancreas, gallbladder, eyes, pregnancy plans — which of the named risks apply to me specifically?
2. What symptom would make you want to hear from me the same day, rather than at my next appointment?
3. Which of my other medications need dose adjustment or timing changes because gastric emptying is slower?
4. What is the plan if I need to stop — for cost, supply, side effects, or pregnancy? What does the off-ramp look like?
5. If I am using a compounded product: what concentration is it, what is the exact volume I should draw, and who do I call if the next vial arrives at a different strength?

Question five is the one people skip and the one the FDA alerts are about.

Common questions

Are GLP-1 medications safe long term?

The longest high-quality safety data on the obesity dose comes from SELECT, which followed 17,604 adults for a mean of 39.8 months without a safety finding that changed the recommendation. GLP-1 receptor agonists as a class have been prescribed for type 2 diabetes since 2005. "Long term" beyond a decade at the 2.4 mg dose is genuinely not yet measured, and anyone telling you otherwise is guessing.

Do GLP-1 drugs cause thyroid cancer?

No human causal link has been established. The boxed warning exists because semaglutide caused dose-dependent thyroid C-cell tumours in mice and rats at clinically relevant exposures, and the label states plainly that human relevance is unknown. The medication is contraindicated if you or a close relative has had medullary thyroid carcinoma or MEN 2.

Is compounded semaglutide as safe as Ozempic or Wegovy?

It is not the same safety question. Approved products have a known identity, purity and concentration; compounded products vary by pharmacy and by batch, and the FDA has documented dosing errors of five to twenty times the intended dose along with the use of unapproved salt forms. If you use a compounded product, verify the concentration on every individual vial before drawing a dose.

What are the warning signs I should not ignore?

Severe persistent abdominal pain radiating to the back, vomiting you cannot stop for more than 24 hours, sudden vision loss or rapidly worsening eyesight, signs of an allergic reaction, and symptoms of low blood sugar if you also take insulin or a sulfonylurea. Each of these is a call today, not at your next appointment.

Do I have to tell my surgeon or anaesthetist I take a GLP-1?

Yes, always, including for dental sedation and endoscopy. Delayed gastric emptying means your stomach may not be empty even after a standard fast, which is an aspiration risk under sedation. The October 2024 multi-society guidance is covered in full in GLP-1 before surgery.

Does the weight come back if I stop?

For most people, a large share of it does. In the STEP 1 trial extension, participants regained about two-thirds of their lost weight in the year after stopping, and cardiometabolic markers drifted back toward baseline. This is why discontinuation is worth planning with a clinician rather than improvising.

Keep reading