The GLP-1 Boxed Warning: What It Says, What It Rests On, and What It Does Not Claim

Answer first: the warning is uniform, the evidence behind it is not

All five of the approved labels examined here carry a boxed warning titled Risk of Thyroid C-Cell Tumors. All five reach the same contraindication in the same words: the drug is contraindicated in patients with a personal or family history of medullary thyroid carcinoma, abbreviated MTC on the labels, or in patients with Multiple Endocrine Neoplasia syndrome type 2, abbreviated MEN 2.

That much is a genuine class rule, and it is worth stating plainly because so little else about these labels is uniform.

What differs is the evidence each label cites underneath the warning. Two of them describe tumors in rats. Two describe tumors in mice and rats. The fifth describes a molecule that is not pharmacologically active in either species and did not produce tumors in rodents at all — and still carries the identical boxed warning and the identical contraindication.

Every one of the five then says a version of the same careful sentence: the human relevance of the rodent finding has not been determined. That sentence is the honest center of this warning, and it is the part most often dropped when the warning is summarized secondhand.

What the boxed warning actually states

The tirzepatide labels are the most direct. In rats, the label says, tirzepatide causes thyroid C-cell tumors. It is unknown whether the drug causes thyroid C-cell tumors, including medullary thyroid carcinoma, in humans, as the human relevance of tirzepatide-induced rodent thyroid C-cell tumors has not been determined.

The semaglutide labels say the same thing about a slightly wider set of animals. In rodents, semaglutide causes thyroid C-cell tumors at clinically relevant exposures, and it is unknown whether the drug causes them in humans, as the human relevance of the rodent finding has not been determined.

Both then give the contraindication and a counseling instruction: counsel patients regarding the potential risk of MTC and the symptoms of thyroid tumors.

Notice the grammatical shape of all of this. The animal finding is stated as fact. The human question is stated as unknown. The contraindication is stated as a rule. Those are three different kinds of claim in one short block, and the label keeps them separate on purpose.

The orforglipron label is the reason this is worth reading closely

The Foundayo label carries the same boxed-warning title and the same contraindication, and then says something the other four do not.

Its warning explains that in products with GLP-1 receptor agonist activity that are pharmacologically active in rats and mice, rodent thyroid C-cell tumors have been observed at clinically relevant exposures, and that these are considered GLP-1 receptor-dependent effects in rodents. Then comes the sentence that makes this label distinctive: orforglipron is not pharmacologically active in rats or mice and did not produce tumors in rodents.

It still carries the boxed warning. The stated reason is that while orforglipron is pharmacologically active at the human GLP-1 receptor, the human relevance of GLP-1 receptor-dependent thyroid C-cell tumors observed in rodents has not been determined.

This is a precautionary warning attached to a receptor rather than to a result. The label is not claiming this molecule caused tumors in animals. It is saying the animal test that produced the class finding could not be run meaningfully on this molecule, because the molecule does not engage the rodent receptor, and the class question therefore stays open.

That distinction is invisible in any summary that says only that GLP-1 drugs carry a thyroid cancer warning. It is also the single clearest example on these labels of a warning that reflects an unanswered question rather than an observed harm.

The human evidence the labels do cite, and how they characterize it

Three of the five labels add a sentence about actual human reports, and it is worth quoting for its caution. Cases of MTC in patients treated with liraglutide, another GLP-1 receptor agonist, have been reported in the postmarketing period, and the data in these reports are insufficient to establish or exclude a causal relationship between MTC and GLP-1 receptor agonist use in humans.

Insufficient to establish or exclude. That phrasing does two jobs at once, and both matter. It refuses to claim the drugs cause MTC in people, and it equally refuses to claim they do not.

That sentence appears on the orforglipron label and on both semaglutide labels. It does not appear on either tirzepatide label — the word liraglutide appears nowhere on those two documents at all, while MTC appears seventeen times on each and calcitonin six times, so the absence is a real editorial difference rather than a failure to look.

What no label claims, anywhere, is a demonstrated human cancer risk. The strongest human statement in this set is an explicit refusal to draw a conclusion in either direction.

Why the labels discourage routine screening

This is the part of section 5.1 most likely to surprise someone who has just read the boxed warning, and all five labels state it in nearly identical language.

Routine monitoring of serum calcitonin, or the use of thyroid ultrasound, is of uncertain value for early detection of MTC in patients treated with these drugs. Such monitoring may increase the risk of unnecessary procedures, due to the low specificity of the calcitonin test and a high background incidence of thyroid disease.

That is a warning about the test, not about the drug. A blood test with low specificity applied to a population where thyroid nodules are common produces a large number of results that lead somewhere without leading anywhere useful.

The labels do describe what an elevated result would mean if one were obtained. Significantly elevated serum calcitonin values may indicate MTC, and patients with MTC usually have calcitonin values above 50 ng/L. If serum calcitonin is measured and found to be elevated, the labels say the patient should be further evaluated, as should patients with thyroid nodules noted on physical examination or neck imaging.

So the position is consistent rather than contradictory. Do not screen everyone as a matter of routine; do investigate a result or a finding that already exists. Whether any of that applies to you is a question for a clinician who knows your history, not something to settle from a label.

The contraindication is about family history, not about symptoms

The contraindication these labels state is unusual among drug warnings in that it turns almost entirely on family history. A personal or family history of medullary thyroid carcinoma, or a diagnosis of Multiple Endocrine Neoplasia syndrome type 2, is what triggers it.

MEN 2 is an inherited condition, which is why the labels reach for family history rather than for a current test result. Someone can carry the relevant risk without having had a single symptom or an abnormal scan.

That is why intake questionnaires ask about relatives rather than only about you. It is not a formality and it is not a liability exercise; it is the one screening question the boxed warning actually specifies.

It is also a question a person can get wrong in good faith. Thyroid cancer is not one disease, and the medullary type is a small minority of thyroid cancers. A relative treated for thyroid cancer years ago may or may not fall inside what this contraindication describes, and that is worth resolving with the family and the prescriber rather than guessing in either direction.

One warning these labels used to carry, and no longer do

Two of these documents contain a line in their Recent Major Changes list recording that a Warnings and Precautions section on suicidal behavior and ideation was removed in February 2026.

That is worth knowing for two reasons. The first is simply that the warning is not on the current labels, and articles written before that change describe a document that no longer exists in that form.

The second is a lesson about reading labels at all. A word search for suicidality on these files returns a hit on two of the five — and on both, the hit is the record of a removal, not a warning. A count that nobody reads would have reported the exact opposite of what the label says.

This is the general hazard with drug labels. They are revised documents with a change log attached, and the change log mentions sections that are gone. What is currently in force is what appears under the numbered headings, not what appears in the list of what changed.

How to read this warning without over- or under-reading it

Take the contraindication literally and the cancer inference cautiously. The contraindication is a firm rule about family history that a prescriber needs an accurate answer to. The tumor finding is an animal result whose human relevance every one of these labels says has not been determined.

Ask which molecule you are being prescribed, by its generic name. The boxed warning is worded the same way across all five, but the evidence cited underneath it differs enough that the generic name is the only way to know what you are actually reading about.

If you have thyroid disease of any kind in your family, say so before you are asked twice, and be specific about which kind if anyone knows. Medullary is the word that matters here.

Then read the leaflet in your own carton. It is the label for the exact product you were dispensed, it is revised more often than any article, and its Recent Major Changes list will tell you what moved since the last version.