Ozempic and Thyroid Cancer: What Is the Real Risk?

Ozempic carries a black box warning — the FDA's most serious warning level — for thyroid C-cell tumours. This warning has caused significant concern for many patients and potential patients. Understanding what it is actually based on, who it applies to, and how to contextualise the risk requires looking at the evidence carefully.

The short summary: the warning exists because of animal data, not confirmed human cases. The risk appears to apply primarily to people with a specific genetic predisposition. For most patients, the thyroid concern does not change the risk-benefit calculation. But it is a real concern for a specific, identifiable group.

What the Black Box Warning Says

The Ozempic (and Wegovy) prescribing information states: "Ozempic causes dose-dependent and duration-dependent thyroid C-cell tumours at clinically relevant exposures in both genders of rats and mice. It is unknown whether Ozempic causes thyroid C-cell tumours, including medullary thyroid carcinoma (MTC), in humans, as the human relevance of semaglutide-induced rodent thyroid C-cell tumours has not been determined."

Critically: "Ozempic is contraindicated in patients with a personal or family history of MTC or in patients with Multiple Endocrine Neoplasia syndrome type 2 (MEN 2)."

The Animal Data

In rodent studies, GLP-1 receptor agonists — including semaglutide — produced thyroid C-cell hyperplasia and medullary thyroid tumours at high doses. Rodent thyroid C-cells have a much higher density of GLP-1 receptors than human thyroid C-cells. Whether the same stimulation of C-cell proliferation that occurs in rodents occurs in humans at clinical doses is the central unanswered question.

Human thyroid C-cells do express GLP-1 receptors, but at far lower levels than in rodents. This is why the warning says "human relevance has not been determined" — the mechanism exists in theory but the magnitude of the effect in humans at approved doses is not confirmed.

What the Human Data Shows

Several large observational studies have attempted to assess whether GLP-1 agonist use is associated with increased thyroid cancer incidence in humans:

A 2023 study in JAMA Network Open examining over 1.6 million GLP-1 agonist users found a statistically significant association between GLP-1 receptor agonist use and all thyroid cancers, but the absolute risk increase was very small (1 to 2 excess cases per 10,000 person-years of use).

The same study found the association was stronger for the non-medullary thyroid cancers (papillary, follicular) than for medullary, which is the type suggested by the animal data. This complicates interpretation.

Longer-term data is still accumulating. The cardiovascular outcome trial for semaglutide (SUSTAIN-6) and the SELECT trial did not identify a significant thyroid cancer signal over five-year follow-up periods.

Who This Warning Genuinely Applies To

The contraindication is clear and should be followed for:

• People with a personal history of medullary thyroid carcinoma (MTC)
• People with a family history of MTC in a first-degree relative
• People with MEN2 syndrome (Multiple Endocrine Neoplasia type 2) — a genetic syndrome that causes MTC, parathyroid tumours, and pheochromocytoma

For these individuals, the theoretical GLP-1-driven C-cell stimulation combined with existing genetic predisposition to C-cell tumours represents an unacceptable risk.

For the general population without these risk factors, the thyroid concern is one to monitor but does not constitute a contraindication.

Monitoring During Treatment

For patients without the contraindications above, routine thyroid monitoring is not formally required but is reasonable practice. Patients should report any new neck lump, difficulty swallowing, hoarseness, or persistent throat discomfort to their physician immediately — these can be symptoms of thyroid masses.

Calcitonin levels (a marker for medullary thyroid C-cell hyperactivity) can be monitored, though this is not standard practice and has its own interpretive limitations.