Open the prescribing information for semaglutide and one of the first things you see is a boxed warning about thyroid tumors. Plenty of people starting GLP-1 drugs read it and ask their doctors the same question: is the semaglutide thyroid cancer risk real? A review published November 6 in Endocrine Reviews by two physicians at Cedars-Sinai Medical Center in Los Angeles makes a blunt argument. After weighing rodent, primate and human data, the authors conclude that the boxed warning is no longer necessary.
That is a scientific recommendation, not a regulatory change. The FDA warning still stands. This explainer walks through why the warning exists, what animal and human studies actually show, and why the debate is not quite over.
Where the semaglutide thyroid cancer warning came from
The story starts in rodents. In preclinical studies, rats and mice given GLP-1 receptor agonists developed dose-dependent increases in thyroid C-cell hyperplasia and tumors. C-cells are a minority population in the thyroid that make calcitonin, and they are the cells that give rise to medullary thyroid carcinoma, a rare cancer distinct from the much more common papillary type.
Those findings contributed to the FDA’s decision to require boxed warnings on products including semaglutide and contraindications for patients with a personal or family history of medullary thyroid carcinoma or the inherited tumor syndrome multiple endocrine neoplasia type 2, known as MEN 2. The labeling is not identical across every GLP-1 medication, but the C-cell tumor warning remains prominent on Ozempic and Wegovy.
That was a reasonable precaution when these medicines were being developed. The C-cell signal was consistent in rodents, medullary thyroid cancer is serious, and long-term human evidence was limited. The open question was whether a rat’s thyroid tells you much about a person’s.
Why rodents may be the wrong model
The new review, written by Elsa W. Braunstein and Glenn D. Braunstein, lays out several reasons for doubt, all drawn from differences in animal biology. Rodents have a relatively high background rate of spontaneous C-cell tumors. Their thyroids contain more C-cells, and those cells express more GLP-1 receptors. The molecular pathways involved in C-cell tumor development also differ between rodents and humans.
In other words, rodent C-cells may be primed to respond to GLP-1 stimulation in a way human C-cells are not. The review also notes that non-human primate studies and human clinical trials did not demonstrate the same C-cell stimulation seen in rodents. That absence does not prove the risk is zero, but it weakens the assumption that the animal findings translate directly to people.
A rat’s thyroid has more C-cells, more GLP-1 receptors and a different biological response. That difference matters.
The distinction is central to drug safety science. Animal studies are designed to identify hazards before medicines reach large populations. A hazard in one species can justify additional precautions even when its relevance to humans remains uncertain. But the warning’s scientific foundation should also be reassessed as human evidence accumulates.
What the human studies show
Human evidence comes mostly from observational studies, and it is not perfectly uniform. One of the strongest reassuring findings comes from a 2024 BMJ study using nationwide registers in Denmark, Norway and Sweden. Among 145,410 patients who started a GLP-1 drug, 76 developed thyroid cancer over an average 3.9 years of follow-up, an incidence of 1.33 per 10,000 person-years. Among 291,667 patients starting a DPP-4 inhibitor, the corresponding incidence was 1.46 per 10,000 person-years.
The adjusted hazard ratio was 0.93, with a 95% confidence interval from 0.66 to 1.31. That finding showed no statistically significant increase in overall thyroid cancer risk. But the upper confidence limit remained compatible with a 31% relative increase, so the study cannot exclude every possible elevation in risk.
There are also important limitations for readers specifically interested in semaglutide. Although the GLP-1 group had an average 3.9 years of follow-up, patients who started semaglutide had an average of only 1.1 years. For medullary thyroid cancer specifically, there were just four cases among GLP-1 users and eleven among DPP-4 users. The hazard ratio was 1.19, but its confidence interval stretched from 0.37 to 3.86. Rare cancers are difficult to evaluate precisely, even in very large datasets.
A 2023 Diabetes Care study from France pointed the other way. Using national health insurance data, researchers matched 2,562 people with thyroid cancer to 45,184 controls. They found that one to three years of GLP-1 drug exposure was associated with higher rates of all thyroid cancer, with an adjusted hazard ratio of 1.58, and medullary thyroid cancer, with an adjusted hazard ratio of 1.78.
The French findings deserve attention, but association is not causation. Observational research can be affected by differences in underlying health, patterns of medical surveillance and diagnostic testing. Detection bias is a plausible explanation for some apparent increases, particularly for thyroid cancers found through imaging, but it cannot simply be assumed to explain every positive result.
The question isn’t whether rats developed tumors. It’s whether GLP-1 treatment causes thyroid cancer in people.
The Cedars-Sinai review argues that the balance of randomized trial meta-analyses and large cohort studies has not established a causal connection. Positive signals also appear in pharmacovigilance databases, where clinicians and patients voluntarily report suspected adverse events. Those systems are valuable for detecting potential problems, but reporting bias and publicity can distort apparent associations.
The review also discusses U.S. thyroid cancer trends since GLP-1 drugs entered clinical use in 2005. Overall thyroid cancer incidence increased during part of that period, a pattern partly attributed to greater detection of papillary cancers, while the relative proportion classified as medullary cancer declined. Such population trends provide context but do not, by themselves, establish the safety of an individual medicine.
What remains uncertain
It is worth being clear about what a review is. It synthesizes existing evidence; it does not generate new clinical outcomes. This one has two authors and presents their interpretation of the available research. Other experts and regulators may weigh the remaining uncertainty differently.
The evidence is increasingly reassuring regarding substantial short-term increases in overall thyroid cancer risk, but it remains less definitive for very rare medullary cancers and effects that might emerge after many years of treatment. Newer regimens, including higher-dose semaglutide for obesity and emerging dual or triple agonists, also have varying amounts of long-term human safety evidence.
Regulators have already examined the issue. In October 2023, the European Medicines Agency’s safety committee concluded that the available evidence did not support a causal association between the GLP-1 receptor agonists it reviewed and thyroid cancer. The committee recommended continued safety monitoring rather than a product-information change.
The regulatory question is separate from the scientific one. Publication in a respected medical journal does not automatically change FDA labeling. The current U.S. Wegovy prescribing information still states that human relevance of the rodent tumor findings is unknown and continues to contraindicate treatment for people with a personal or family history of medullary thyroid carcinoma or MEN 2.
Patients should not treat the new review as a reason to disregard that restriction or independently change prescribed therapy. The label also notes that routine calcitonin screening or thyroid ultrasound has uncertain value for early detection in patients receiving Wegovy.
Grey Pepper has covered other safety questions where early signals and later evidence pulled in different directions, including the debate over semaglutide and a rare form of vision loss, and real-world evidence on cardiovascular outcomes with Ozempic and Mounjaro.
The bottom line
The thyroid warning on semaglutide reflects the logic of precautionary drug regulation: a strong animal signal, a serious potential outcome and uncertainty about long-term human risk. More than two decades after GLP-1 drugs first reached patients, the biology looks increasingly species-specific, the primate findings have not reproduced the rodent signal, and large human cohorts have generally offered reassurance.
But reassurance is not proof of zero risk. The French study remains a counterpoint, the Scandinavian data have important limits for medullary cancer and semaglutide-specific follow-up, and uncommon outcomes require prolonged surveillance.
That is why two researchers can reasonably argue that the boxed warning has outlived its scientific justification, while regulators may still want stronger evidence before changing it. The disagreement is not about whether the rodent tumors existed. It is about how much weight those findings should continue to carry when the question concerns human health.
Sources
- Braunstein EW, Braunstein GD. Glucagon-like Peptide-1 Receptor Agonists and Thyroid Cancer: Review of Preclinical and Human Data. Endocrine Reviews. Published online November 6, 2026.
- Pasternak B, et al. Glucagon-like peptide 1 receptor agonist use and risk of thyroid cancer: Scandinavian cohort study. BMJ. April 10, 2024.
- Bezin J, et al. GLP-1 Receptor Agonists and the Risk of Thyroid Cancer. Diabetes Care. 2023.
- U.S. National Library of Medicine, DailyMed. Wegovy (semaglutide) prescribing information. Boxed warning and contraindications.
- European Medicines Agency. PRAC meeting highlights: GLP-1 receptor agonists and thyroid cancer. October 2023.
- Society for Endocrinology, British Thyroid Association and British Association of Endocrine and Thyroid Surgeons. Joint society statement regarding GLP-1 agonists and thyroid cancer. 2026.
