GLP-1 agonists are the most widely prescribed weight-loss drugs in history, and the safety database is now substantial. By the start of 2026, semaglutide alone had accumulated over 30 million patient-years of exposure across diabetes and obesity indications. Tirzepatide is approaching 5 million. The acute tolerability profile (nausea, diarrhoea, constipation) is well-defined. The mid-range safety questions (pancreatitis, gallbladder, thyroid) have been actively monitored for over a decade. The longest-running outcomes trial, SELECT (Lincoff et al., NEJM, 2023, PMID: 37952131), provides cardiovascular follow-up of 17,604 patients over a median 39.8 months.
This article walks through the safety profile of the class as a whole rather than re-doing per-drug side effect tables, which are covered in Semaglutide vs Tirzepatide. The triple-agonist-specific dysesthesia signal seen with retatrutide is covered separately in Triple vs Dual Agonism, because that signal does not apply to the dual or single agonists that make up the rest of the class.
For background on the receptor biology that produces both the efficacy and the side-effect spectrum, see How GLP-1 Agonists Work.
What “safety” means for the GLP-1 class
GLP-1 agonists are not narrow-target drugs. The receptor is expressed in the pancreas, gut, brain, cardiovascular tissue, and several other locations. A single molecule activating a single receptor type produces effects in five organ systems, because that receptor lives in five organ systems. This is the strength and the weakness of the class. The breadth of effects produces remarkable efficacy. It also produces a spectrum of side effects broader than what most drug classes show.
The safety profile breaks into four tiers, each with a different evidence base and a different clinical priority:
- Tier 1: Acute GI tolerabilityNausea, vomiting, diarrhoea, constipation. Common, mostly mild to moderate, dose-dependent, and resolve with tolerance. Drive most of the 6 to 8% discontinuation rate seen across the class. Managed with slow titration and timing strategies. Well understood.
- Tier 2: Class-wide adverse eventsPancreatitis, gallbladder disease, gastroparesis, retinopathy progression in pre-existing diabetic eye disease, acute kidney injury from volume depletion. Rare but serious. Carry boxed or label warnings. Clear contraindications and monitoring guidance exist.
- Tier 3: Reviewed and not substantiatedThyroid C-cell carcinoma (rodent signal, not seen in humans) and suicidal ideation (FDA review concluded January 2024 found no causal link). Both retain label language out of caution despite negative findings. Clinically actionable as contraindications rather than active monitoring targets.
- Tier 4: Long-term unknownsEffects beyond five years of continuous use are still being characterised. Open questions include muscle and bone changes over decades, rare events that emerge only with millions of patient-years, and whether the metabolic adaptations to chronic GLP-1 receptor activation produce any signals that have not yet been detected.
The rest of this article walks through each tier in turn. The one signal not covered here is dysesthesia (abnormal skin sensation), reported in 20.9% of retatrutide 12mg participants in TRIUMPH-4 versus 0.7% on placebo. That finding is specific to triple-receptor agonism and is covered in detail in the Triple vs Dual Agonism explainer.
The acute GI profile and how it resolves
Gastrointestinal side effects are intrinsic to the mechanism. Slowed gastric emptying is a primary effect that drives part of the appetite suppression, which is why nausea is the most common side effect across every compound in the class. Direct GLP-1 receptor activity in the area postrema (the brain region responsible for vomiting) also contributes.
SURMOUNT-5 (Aronne et al., NEJM, 2025) reported the head-to-head GI rates for the two leading drugs. Nausea occurred in 50.0% of semaglutide participants and 44.0% of tirzepatide participants over 72 weeks. Diarrhoea ran 20.4% versus 24.7%. Vomiting was 21.1% versus 20.7%. Constipation was 17.6% versus 21.6%. Discontinuation due to adverse events landed at 8.0% on semaglutide and 6.1% on tirzepatide. The tirzepatide GI profile is slightly more favourable than the semaglutide profile despite higher efficacy, which is one of the more interesting findings of the head-to-head trial.
Symptoms peak in the first one to two weeks after each dose increase, then decline as tolerance develops. Slow titration protocols exist specifically to soften this peak. Wegovy titrates over 16 weeks (5 dose steps to reach 2.4mg). Zepbound titrates over 20 weeks (6 dose steps to reach 15mg). Skipping titration steps causes preventable nausea and is the single most common cause of avoidable discontinuation.
Serious adverse events the class is monitored for
Five class-wide serious events drive the boxed warnings, contraindications, and monitoring guidance attached to every GLP-1 agonist label. Absolute rates are low. The clinical priority is recognising risk factors that exclude a patient from the class rather than monitoring active drug users for emerging events.
The thyroid C-cell concern sits in a separate category because it is a label-level contraindication based on rodent data rather than human evidence. Rodent studies showed dose-dependent C-cell tumours with liraglutide and other GLP-1 agonists. Whether the signal applies to humans is unclear, because human C-cells express far fewer GLP-1 receptors than rodent C-cells. No causal signal has emerged in human trials or post-marketing data through 2026. The FDA has nevertheless contraindicated the entire class in anyone with a personal or family history of medullary thyroid carcinoma or Multiple Endocrine Neoplasia type 2, out of caution.
The signals the FDA has reviewed
Two safety signals have been formally reviewed by regulators and not substantiated. Both retain label language out of caution. Understanding what was reviewed and what was concluded matters because both signals have circulated widely in the public conversation and are often cited as established risks.
Suicidal ideation and self-harm
The signal was raised in mid-2023 when the European Medicines Agency announced a review of post-marketing reports. The FDA followed with its own review of clinical trial data, FAERS adverse event reports, and pharmacovigilance signals. In January 2024 the FDA concluded that the available evidence did not support a causal association between GLP-1 agonists and suicidal thoughts or self-harm. The EMA reached a similar conclusion in April 2024. Both agencies noted that patients with a history of mental health concerns should be monitored and should discuss the risk-benefit picture with their prescriber, as is standard for most chronic medications.
Cardiovascular safety
The cardiovascular question is settled in the opposite direction: the data show benefit rather than harm. SELECT (Lincoff et al., NEJM, 2023, PMID: 37952131, NCT03574597) followed 17,604 patients with established cardiovascular disease on semaglutide 2.4mg over a median 39.8 months. The trial showed a 20% reduction in major adverse cardiovascular events versus placebo, driven by reductions in cardiovascular death, non-fatal myocardial infarction, and non-fatal stroke. SELECT is the largest cardiovascular outcomes trial ever completed for a weight-loss drug and is the basis for semaglutide's specific cardiovascular protection labelling. Tirzepatide's outcomes trial (SURPASS-CVOT) is expected to read out in 2026 or 2027. Until it does, semaglutide is the only GLP-1 weight-loss drug with published cardiovascular outcomes data.
Pregnancy and fertility
Animal reproductive toxicity studies show developmental toxicity at clinical exposure levels for every approved GLP-1 agonist. Human pregnancy data are limited because trials exclude pregnant women and accidental exposures are under-reported. The FDA recommendation is to discontinue any GLP-1 agonist at least two months before planned conception, accounting for the long half-lives. Anyone who becomes pregnant while on a GLP-1 should stop the drug and contact their prescriber. The class is also not recommended during breastfeeding. A separate fertility question has emerged with reports of unintended pregnancies among women on GLP-1s who had previously been infertile, possibly related to weight loss restoring ovulation. This is a counselling point rather than a safety concern, and contraception guidance should reflect it.
Who should not take a GLP-1 agonist
The contraindications are stable across the class and have changed little since liraglutide's 2010 approval. They divide into absolute exclusions and relative cautions.
| Category | Condition | Reason |
|---|---|---|
| Absolute | Personal or family history of medullary thyroid carcinoma | Rodent C-cell tumour signal, label contraindication |
| Absolute | Multiple Endocrine Neoplasia type 2 (MEN 2) | Same C-cell concern, genetic predisposition |
| Absolute | Personal history of pancreatitis | Boxed warning, recurrence risk |
| Absolute | Pregnancy or planned conception within 2 months | Animal teratogenicity, long drug half-life |
| Relative | Severe gastroparesis or gastric outlet obstruction | Mechanism worsens motility |
| Relative | Active gallbladder disease | Rapid weight loss can precipitate events |
| Relative | Proliferative diabetic retinopathy | Rapid glucose normalisation can worsen course |
| Relative | Severe renal or hepatic impairment | Limited safety data, monitor closely |
Pre-operative guidance has been added more recently. The American Society of Anesthesiologists issued recommendations in 2023 to hold daily GLP-1 agonists for one day and weekly ones for one week before elective procedures requiring anaesthesia, to reduce the risk of aspiration from delayed gastric emptying. Some centres have moved to two-week holds for the longer-acting drugs. The guidance is evolving as real-world aspiration data accumulate.
The long-term picture beyond five years is the area where the evidence base is least mature. Continuous receptor activation over decades has not been studied. Open questions include effects on muscle mass and bone density (rapid weight loss produces both), whether any rare events become apparent only with millions of patient-years of exposure, and whether the metabolic adaptations seen in the first year of treatment continue to evolve. Real-world pharmacovigilance through May 2026 has not surfaced any new long-term signals beyond those already characterised. For the foundational mechanism biology that produces both the efficacy and the safety spectrum, see How GLP-1 Agonists Work.
Frequently Asked Questions
What are the most common side effects of GLP-1 agonists?
Gastrointestinal symptoms dominate. Nausea is the most common, reported by roughly 40 to 50% of participants in trials of semaglutide and tirzepatide. Diarrhoea, vomiting, and constipation each occur in 15 to 25% of users. Most symptoms are mild to moderate, peak in the first week or two after each dose increase, and decline as tolerance develops. Discontinuation due to adverse events runs around 6 to 8% across the class.
Do GLP-1 agonists cause pancreatitis?
The signal exists but the absolute risk is low. Early post-marketing reports raised concern, but the SELECT trial in 17,604 patients on semaglutide over a median 39.8 months showed no significant increase versus placebo. Current guidance is to stop the drug if pancreatitis is suspected and to avoid GLP-1 agonists in anyone with a personal history of acute or chronic pancreatitis. The label warning persists across the class.
What is the thyroid C-cell concern?
Rodent studies showed dose-dependent thyroid C-cell tumours with liraglutide and other GLP-1 agonists. Whether this translates to humans is unclear because human C-cells express far fewer GLP-1 receptors than rodent C-cells. No causal signal has emerged in human trials or post-marketing surveillance. The FDA has nevertheless contraindicated the entire class in anyone with a personal or family history of medullary thyroid carcinoma or Multiple Endocrine Neoplasia type 2.
Do GLP-1 agonists cause gastroparesis?
Severe gastroparesis case reports exist but are rare. The class slows gastric emptying as part of its mechanism, which is why nausea is common. In a small minority of users this slowing becomes severe enough to cause persistent vomiting, abdominal pain, and the inability to eat. The condition usually resolves after stopping the drug. Anyone with pre-existing gastroparesis should avoid the class. Pre-operative guidance now includes holding GLP-1 agonists for one to several days before anaesthesia to reduce aspiration risk.
Did the FDA find that GLP-1 agonists cause suicidal thoughts?
No. In January 2024 the FDA concluded its review of post-marketing reports of suicidal ideation and self-harm with GLP-1 agonists. The review found no causal link. The signal had been raised based on European Medicines Agency reports in 2023, and the FDA's analysis of clinical trial data and FAERS reports did not support a causal association. Patients with a history of mental health concerns should still discuss the risk-benefit picture with their prescriber.
Are GLP-1 agonists safe during pregnancy?
No. Animal studies show developmental toxicity at clinical exposure levels, and human pregnancy data are insufficient. The FDA recommends discontinuing GLP-1 agonists at least two months before conception, given the long half-lives. Anyone who becomes pregnant on a GLP-1 should stop the drug and contact their prescriber. The drugs are also not recommended during breastfeeding.
What does long-term safety look like beyond five years?
The data are still maturing. Semaglutide has the longest cumulative exposure (Wegovy approved June 2021, Ozempic in diabetes since 2017), with SELECT providing 39.8 months of cardiovascular outcomes data. No new long-term safety signals have emerged from real-world pharmacovigilance through 2026. Open questions include effects on muscle and bone over decades of continuous use, and whether any rare events become apparent only with millions of patient-years of exposure.
Who should not take a GLP-1 agonist?
Anyone with a personal or family history of medullary thyroid carcinoma or MEN 2, anyone with a history of pancreatitis, anyone with severe gastroparesis, and anyone who is pregnant or planning pregnancy within two months. People with severe gastrointestinal disease, gallbladder disease, or diabetic retinopathy should discuss risks with their prescriber. The drugs are not approved for paediatric use under age 12 (semaglutide) or 18 (other GLP-1s for weight loss).
This article is for informational and educational purposes only and does not constitute medical advice. GLP-1 agonists are prescription medications with a range of indications, contraindications, and side effects. The safety information presented here reflects regulatory guidance and published clinical trial data as of May 2026 and may be superseded by future evidence. Consult a licensed healthcare provider before starting, stopping, or changing any medication.
Last updated: May 2026.