News

Is Semaglutide Safe? What the Evidence Shows

Creator: Nadia Okonjo, Health Writer
Published: 2026-08-25

Semaglutide has years of trial data and a defined risk profile, but safe is not universal. Eligibility depends on personal history, other medicines, monitoring and exactly what product is dispensed.

Nadia Okonjo, Health Writer August 25, 2026 6 min read

Two women in full-length casual clothes sit in armchairs for a calm conversation at home

The short version

In STEP 1, gastrointestinal events led 4.5% of participants assigned semaglutide to stop treatment versus 0.8% assigned placebo; common does not always mean tolerable.
STEP 5 followed treatment for 104 weeks: gastrointestinal events occurred in 82.2% with semaglutide and 53.9% with placebo, and most were mild to moderate.
SELECT followed 17,604 adults for a mean 39.8 months: major cardiovascular events were 6.5% versus 8.0%, while discontinuation for adverse events was 16.6% versus 8.2%.
The thyroid boxed warning comes from rodent C-cell tumors; the relevance to humans is unknown, and personal or family MTC or MEN 2 history is a contraindication.
NAION remains rare and the observational evidence conflicts: one 2025 cohort estimated 1.41 additional cases per 10,000 person-years, while another large cohort found no association.
Compounded-vial safety depends on accountable prescribing, pharmacy labeling and concentration-specific measurement; poison-center reports include 10-fold dosing errors.

Is semaglutide safe? For many eligible adults, randomized trials support a known and monitorable safety profile, but the honest answer is conditional. Most adverse effects are gastrointestinal and often settle; some people stop treatment because of them, and rare but serious risks change who should not receive the medicine. Safety also depends on the product. Branded pens and properly prescribed compounded vials are not interchangeable, and an anonymous vial removes the clinical and pharmacy safeguards that make dosing traceable. A clinician weighs thyroid history, gallbladder and pancreatic history, eye health, diabetes medicines, pregnancy plans and current symptoms before prescribing.

Is semaglutide safe long term? What STEP and SELECT show

The evidence is larger and longer than a social-media anecdote, but it is not lifetime evidence. In STEP 1, 1,961 adults with overweight or obesity and without diabetes were followed for 68 weeks. Gastrointestinal events were usually transient and mild to moderate, yet 4.5% of participants assigned semaglutide stopped because of those events, compared with 0.8% assigned placebo (Wilding et al., New England Journal of Medicine, 2021). That difference matters: a common effect can still make treatment unacceptable for an individual.

STEP 5 extended randomized observation to 104 weeks. Gastrointestinal events occurred in 82.2% of the semaglutide group and 53.9% of the placebo group; investigators described most as mild to moderate, and 92.8% of all randomized participants attended the end-of-trial safety visit (Garvey et al., Nature Medicine, 2022).

SELECT then followed 17,604 adults with established cardiovascular disease and overweight or obesity, but without diabetes, for a mean 39.8 months. Major cardiovascular events occurred in 6.5% with semaglutide and 8.0% with placebo, a 20% relative reduction as the trial reported it. Adverse events led to discontinuation in 16.6% versus 8.2% (Lincoff et al., New England Journal of Medicine, 2023); serious adverse events overall were 33.4% versus 36.4% (Kushner et al., Obesity, 2025). Cardiovascular benefit in that specific high-risk population does not cancel tolerability problems, and roughly 40 months is not the same as decades.

The risks that change a prescribing decision

The boxed warning is based on thyroid C-cell tumors in rodents at clinically relevant exposures. It is unknown whether that finding applies to humans. Semaglutide is contraindicated for someone with a personal or family history of medullary thyroid carcinoma, or with multiple endocrine neoplasia syndrome type 2. That history is a screening question, not a risk that can be managed by simply watching for symptoms later.

Pancreatitis remains a labeled warning even though randomized data have not shown an excess in every trial. In SELECT, confirmed acute pancreatitis occurred in 0.2% with semaglutide and 0.3% with placebo. Gallbladder-related disorders were slightly more frequent, 2.8% versus 2.3%. A history of pancreatitis, gallstones, severe gastroparesis, diabetic retinopathy or medicines that can lower glucose changes the clinical conversation; none of those facts can be reduced to a universal yes or no.

The everyday symptom-by-symptom discussion belongs in the separate guide to semaglutide side effects. The safety question here is whether a person's history, current medicines and ability to tolerate treatment make the overall tradeoff reasonable. At Promise, a licensed provider reviews every request and not everyone qualifies.

If semaglutide is not a good fit, a prescriber may compare another incretin medicine such as tirzepatide. That is a fresh eligibility decision, not an assumption that two related medicines carry identical risks.

What the newer eye and mental-health reviews mean

Nonarteritic anterior ischemic optic neuropathy, or NAION, is a sudden loss of blood flow to the optic nerve. A 2025 Danish-Norwegian cohort found 32 events among semaglutide users and estimated an adjusted hazard ratio of 2.81 versus SGLT-2 inhibitors, but the absolute difference was 1.41 additional cases per 10,000 person-years (Simonsen et al., Diabetes, Obesity and Metabolism, 2025). A separate multinational cohort of nearly 300,000 people with diabetes, obesity or both did not find a statistically significant association (Chou et al., Ophthalmology, 2025). Both were observational, so neither proves or disproves causation.

Regulators weighed that signal differently. In June 2025, the European Medicines Agency's safety committee classified NAION as a very rare adverse effect, potentially affecting up to 1 in 10,000 people taking semaglutide. The U.S. prescribing information available on August 25, 2026 did not list NAION. A sudden, painless loss of vision or rapidly worsening eyesight needs urgent same-day assessment regardless of which regulator's label applies.

The mental-health review has moved in the other direction. The FDA's January 2024 preliminary evaluation and the EMA committee's April 2024 review found no evidence of a causal association with suicidal thoughts. In January 2026, the FDA said its completed review found no increased risk and requested removal of suicidal-behavior wording from affected weight-management labels. That is consistent with a large electronic-health-record study that found no greater risk with semaglutide (Wang et al., Nature Medicine, 2024) and with SELECT, where serious suicide or self-injury events were 0.11% in both groups. New or worsening mood symptoms still deserve prompt clinical attention; a population-level finding never makes an individual's symptoms unimportant.

Is compounded semaglutide safe in the same way?

Product source adds a separate layer of risk. Compounded semaglutide is different from the FDA-approved products Ozempic and Wegovy and is not FDA-approved. A compounded preparation does not undergo the same premarket review for safety, effectiveness and quality as a branded product. It can still be lawfully prepared by a licensed U.S. compounding pharmacy for an individual patient on a prescription. Regulatory status is one input, not a marketing gate: a licensed provider may still prescribe a compounded formulation when clinically appropriate, and that decision belongs to the patient and prescriber.

The FDA's July 2024 alert focused on measurement errors with multi-dose compounded vials, especially confusion among milligrams, milliliters and syringe units. A poison-center case series documented two 10-fold errors and symptoms lasting for days (Lambson et al., Journal of the American Pharmacists Association, 2023). The compounded semaglutide explainer covers the regulatory and pharmacy distinction; the GLP-1 units-to-mg guide explains why numbers from one concentration cannot be copied to another. The prescribed amount and measurement belong on the pharmacy label and in the clinician's instructions.

That is the practical legitimacy test: a named prescriber reviewed the history, a licensed pharmacy dispensed a labeled product, and there is a clinical contact for questions. A grey-market vial sold without a prescription offers none of that traceability. Clinical oversight does not erase semaglutide's risks, but it makes the product, measurement and response to problems accountable.

The most useful answer is therefore conditional: semaglutide has a substantial safety record for eligible, monitored patients, clear reasons to exclude some people and unresolved questions that deserve honest follow-up. The right comparison is not risk versus no risk. It is a documented medical decision with monitoring versus a product and dose nobody accountable has verified.

Common questions

Is semaglutide safe for long-term use?

For many eligible adults, the available evidence supports a defined, monitorable safety profile over several years. STEP 5 followed participants for two years and SELECT for a mean of about 40 months. Those studies do not answer what happens after decades of exposure, so long-term treatment still requires periodic review of benefit, tolerability, medical history and other medicines.

Does semaglutide cause thyroid cancer?

Human evidence has not established that semaglutide causes thyroid cancer. The boxed warning exists because thyroid C-cell tumors occurred in rodents, and it remains unknown whether that finding applies to humans. Semaglutide is contraindicated for people with a personal or family history of medullary thyroid carcinoma or multiple endocrine neoplasia syndrome type 2.

Can semaglutide cause pancreatitis or gallbladder problems?

Pancreatitis is a labeled warning, while gallbladder events occurred slightly more often with semaglutide in SELECT. Confirmed acute pancreatitis was 0.2% with semaglutide and 0.3% with placebo; gallbladder-related disorders were 2.8% and 2.3%, respectively. A prescriber considers prior pancreatitis, gallstones and current symptoms rather than treating those trial averages as an individual forecast.

Is compounded semaglutide as safe as a branded pen?

It does not have the same product-level evidence or premarket review as a branded pen. Safety depends partly on whether a licensed pharmacy prepared and clearly labeled the medication for an individual prescription. Multi-dose vials also introduce measurement risk, so the prescribed amount and syringe measurement need to come from that product's pharmacy label rather than a generic conversion found elsewhere.

Can semaglutide affect eyesight?

It can raise more than one eye-safety question, and they are distinct. Rapid glucose improvement can worsen diabetic retinopathy in some people with diabetes, while observational studies of the rare optic-nerve condition NAION have reached conflicting results. EMA classified NAION as very rare in 2025. Sudden painless vision loss or rapidly worsening eyesight needs urgent assessment.

Does semaglutide increase suicidal thoughts?

Current large-scale reviews have not found an increased risk. The FDA's preliminary 2024 evaluation and EMA's 2024 review found no causal association; after a fuller review, the FDA requested removal of suicidal-behavior wording from affected weight-management labels in January 2026. Any new or worsening mood symptoms still warrant prompt clinical attention, regardless of the population-level evidence.