The thyroid C-cell tumor warning, explained

Where the warning appears, and how documents differ

In the United States the warning sits in a box at the top of the prescribing information, above every other section. The same box, titled risk of thyroid C-cell tumors, appears on the labels for semaglutide (Ozempic®, Rybelsus®, Wegovy®), liraglutide (Victoza®, Saxenda®), dulaglutide (Trulicity®), tirzepatide (Mounjaro®, Zepbound®) and orforglipron (Foundayo™). Each label backs it with a contraindication in section 4, an explanation in section 5.1 and the animal data in section 13.1 ^{[1][2][3][4][5]}.

The European product information handles the same findings differently. For Wegovy®, hypersensitivity is the only contraindication, and the thyroid tumors are described under preclinical safety data rather than in a warning box ^[6]. The brand pages on this site, such as the Ozempic® and Zepbound® side-effect guides, cover the rest of each label.

What the animal studies found

• Semaglutide: in two-year studies, mice and rats developed more thyroid C-cell adenomas, and carcinomas rose as well, at exposures comparable to those seen in patients ^[1]. The EU document adds that no other treatment-related tumors appeared ^[6].
• Liraglutide: benign C-cell adenomas increased with dose in mice and rats, and malignant C-cell carcinomas were also seen. Mouse experiments showed the cell growth depended on the GLP-1 receptor ^[2].
• Dulaglutide: in a two-year rat study, C-cell tumors became more frequent with higher doses and longer treatment; a six-month study in transgenic mice found no C-cell growth or tumors ^[3].
• Tirzepatide: adenomas rose in rats, and adenomas and carcinomas counted together rose at every dose tested; a six-month transgenic mouse study found no tumors ^[4].
• Orforglipron: no tumors in rats or transgenic mice, but the drug does not act on the rodent receptor, so the label expects it would share the class effect if it did ^[5].

What is and is not known in people

The central statement in each US label is that it is unknown whether the medicine causes C-cell tumors, including MTC, in humans, because the human relevance of the rodent results has not been determined ^[1][4]. The labels neither confirm nor rule out a risk in people.

A few human data points appear. After approval, MTC has been reported in people using liraglutide, and the labels say those reports are not enough to establish or exclude a cause ^[1][5]. The Trulicity® label describes one MTC case in a trial participant whose calcitonin was already about eight times the upper limit of normal before treatment, and one case of C-cell overgrowth with raised calcitonin in its cardiovascular trial ^[3].

European assessors describe the semaglutide rodent tumors as the result of a non-genotoxic, receptor-driven process to which rodents are especially sensitive, and consider the relevance to humans low without being able to exclude it completely ^[6].

What studies in people have reported

Observational studies compare people who used a medicine with similar people who did not. Differences in who is prescribed what, and in how closely each group is examined, can shape their results, so they show associations rather than cause and effect. Randomized trials avoid that problem but rarely run long enough, or include enough people, to count a rare cancer reliably.

• France, nested case-control study: among adults with type 2 diabetes in the national health insurance database, 2,562 people with thyroid cancer were compared with 45,184 matched controls. One to three years of GLP-1 use was associated with more thyroid cancer overall (hazard ratio 1.58) and more MTC (1.78) ^[7].
• Denmark, Norway and Sweden, cohort study: 145,410 people starting a GLP-1 medicine were compared with 291,667 starting a DPP-4 inhibitor, another diabetes medicine, over a mean of 3.9 years. Thyroid cancer rates were similar (hazard ratio 0.93), and the MTC estimate was too imprecise to read either way ^[8].
• Meta-analysis of 64 randomized trials lasting at least a year: GLP-1 treatment was linked to more thyroid cancer overall (odds ratio 1.52), but a single case could have changed the result, and the authors estimated that about 1,349 people would need five years of treatment for one extra case. Papillary cancer and MTC on their own showed no significant difference ^[9].
• European Medicines Agency safety committee, 2023: after reviewing the French study, other publications and company data for exenatide, liraglutide, dulaglutide, semaglutide and lixisenatide, it found that the evidence did not support a causal association with thyroid cancer and asked companies to keep monitoring ^[10]. Tirzepatide and orforglipron were not part of that review.

Why the rodent result may differ in people, and why caution remains

• Supported by the labels: in rodents the C-cell growth depends on the GLP-1 receptor, and the Foundayo™ label treats it as an expected effect of the whole class ^[2][5].
• Proposed by regulators but not proven in people: rodents appear particularly sensitive to this mechanism, which is why the EU document rates the human relevance as low ^[6].
• General medical background: MTC is very rare. Most cases are sporadic, but hereditary MTC runs in families, and a family history of MTC or of multiple endocrine neoplasia raises the risk ^[11].
• Why the exclusion stays: MTC arises from the same C cells as the rodent tumors, and the studies so far cannot exclude a small risk, so people already at high risk are kept off these medicines ^[1][8][9].

Who is excluded, and what else to mention

The US exclusion is the same on every label: MTC in yourself or a close relative, or MEN 2 ^[1][4]. Other thyroid conditions, such as an underactive thyroid, nodules or other types of thyroid cancer, are not contraindications. Your prescriber will still want to hear about them, and about any relative with a thyroid cancer whose type you are not sure of.

If you take levothyroxine, mention that too. In an interaction study, semaglutide tablets raised levothyroxine exposure by about a third, and both the US and EU documents ask clinicians to monitor thyroid treatment in that situation ^[1][6]. This is about absorption of a tablet, not about tumors. Other reasons a GLP-1 may not suit someone are collected in the article on who should not take one.

Calcitonin tests and neck ultrasound

Calcitonin is a hormone made by thyroid C cells, and a blood test for it is used when MTC is suspected or being followed ^[11][14]. You might expect regular testing during treatment. The US labels instead describe routine calcitonin or thyroid ultrasound monitoring as of uncertain value for finding MTC early, with a risk of unnecessary procedures, because the test is not very specific and thyroid disease is common ^[1][5]. They note that people with MTC usually have values above 50 ng/L, and that a raised result, or a nodule found on examination or imaging, calls for further evaluation ^[1].

Calcitonin can shift slightly on treatment. In the Saxenda® trials, 1.2% of people on liraglutide and 0.6% on placebo ended with values at least twice the upper limit of normal, 0.5% and 0.2% ended above 20 ng/L, and nobody who started below 20 ng/L rose above 50 ng/L ^[2].

Other things raise calcitonin as well: MedlinePlus lists kidney disease, smoking, higher body weight and some medicines that reduce stomach acid ^[14]. A single raised value therefore needs a clinician to interpret it in context.

Neck symptoms: what to report and what else they can mean

The Medication Guides ask you to tell your healthcare provider about swelling or a lump in your neck, a hoarse voice, difficulty swallowing or breathlessness ^[1]. MedlinePlus lists similar symptoms for MTC, along with cough and diarrhea caused by high calcitonin ^[11].

Each of these has far more common causes. Most neck lumps in adults are not cancer; enlarged lymph nodes from an infection are the usual reason ^[12]. Hoarseness most often follows a cold or throat infection and settles within about two weeks, and acid reflux is another cause ^[13], which matters because reflux is also a GLP-1 side effect with its own article. Even so, MedlinePlus advises having any neck lump checked promptly, and hoarseness that lasts more than two to three weeks in an adult ^[12][13].

Ozempic®, Rybelsus®, Wegovy®, Victoza® and Saxenda® are registered trademarks of Novo Nordisk A/S. Mounjaro®, Zepbound® and Trulicity® are registered trademarks and Foundayo™ is a trademark of Eli Lilly and Company. PepLabo is not affiliated with or endorsed by Novo Nordisk or Eli Lilly and Company.

When to get help