Semaglutide caused thyroid C-cell tumors in mice and rats, and tirzepatide caused them in rats, at exposures the labels call clinically relevant. The Ozempic prescribing information and the Mounjaro prescribing information both state that it is unknown whether the drugs cause those tumors, including medullary thyroid carcinoma, in humans, because the human relevance of the rodent finding has not been determined. Both labels contraindicate the drug in a personal or family history of medullary thyroid carcinoma, or in Multiple Endocrine Neoplasia syndrome type 2.
The question is asked about peptides as a class. Semaglutide and tirzepatide carry this warning, as do liraglutide and dulaglutide, the other GLP-1 receptor agonists in the catalog with an approved label of this kind. Most other peptides in the catalog do not.
It is unknown whether these drugs cause thyroid C-cell tumors, including medullary thyroid carcinoma, in humans.
A boxed warning puts the rodent finding above the indications
On these labels the warning is a block at the top of the prescribing information, under the heading WARNING, before the uses of the drug are listed. The two boxes are nearly the same text.
The Ozempic box says that in rodents, semaglutide causes dose-dependent and treatment-duration-dependent thyroid C-cell tumors at clinically relevant exposures. It says it is unknown whether Ozempic causes thyroid C-cell tumors, including medullary thyroid carcinoma (MTC), in humans, "as human relevance of semaglutide-induced rodent thyroid C-cell tumors has not been determined." It contraindicates Ozempic in patients with a personal or family history of MTC or with Multiple Endocrine Neoplasia syndrome type 2 (MEN 2). It tells the prescriber to counsel patients about that potential risk and about the symptoms it lists: a mass in the neck, dysphagia, dyspnea, persistent hoarseness.
The Mounjaro box makes the same three moves and names the species. "In both male and female rats, tirzepatide causes dose-dependent and treatment-duration-dependent thyroid C-cell tumors at clinically relevant exposures." The human clause matches, and so does the contraindication: a personal or family history of MTC, or MEN 2, the inherited syndrome named beside it. The human clause still says relevance has not been determined.
Medullary thyroid carcinoma is a tumor of the thyroid's C cells, the cells that make calcitonin and the cells that formed the tumors in the animal studies. The box uses "including medullary thyroid carcinoma" as the human name for that C-cell tumor. Papillary and follicular thyroid cancers begin in a different cell. The warning is about one cancer, and about the people for whom that cancer, or MEN 2, is already in the history.
Both boxes then decline a screening program. Routine monitoring of serum calcitonin, or thyroid ultrasound, "is of uncertain value" for early detection of MTC in people treated with either drug. The Ozempic label says such monitoring may increase the risk of unnecessary procedures, because the calcitonin test is nonspecific and thyroid disease is already common. A significantly elevated calcitonin may indicate MTC, and the label says people with MTC usually have values above 50 ng/L.
The tumors appeared at exposures that overlap a human dose
"Clinically relevant" means the plasma exposure in the animals, measured as area under the curve, overlapped exposure at the maximum recommended human dose.
Under the Ozempic warning, the toxicology section describes two two-year carcinogenicity studies. In CD-1 mice, males received 0.3, 1, and 3 mg/kg/day — 2-, 11-, and 30-fold the maximum recommended human dose of 2 mg a week — and females received 0.1, 0.3, and 1 mg/kg/day, or 1-, 2-, and 7-fold. A statistically significant increase in thyroid C-cell adenomas, and a numerical increase in C-cell carcinomas, occurred in males and females at clinically relevant exposures. "Numerical" is the label's word for that carcinoma rise: recorded, and not marked statistically significant.
In Sprague Dawley rats the doses were 0.0025, 0.01, 0.025, and 0.1 mg/kg/day, mapped to an exposure below quantification and then 0.2-, 0.5-, and 3-fold the exposure at 2 mg a week. C-cell adenomas increased, statistically significantly, in males and females at all dose levels, including the dose whose exposure the label could not quantify. C-cell carcinomas increased, statistically significantly, in males at 0.01 mg/kg/day and above, the dose the label places at 0.2-fold human exposure. The warnings section calls the mouse and rat work lifetime exposure. The studies under that phrase are the two-year studies.
The toxicology section then states the limit the box is built on. Human relevance of the rat tumors "is unknown and could not be determined by clinical studies or nonclinical studies." The warnings section adds a postmarketing note and refuses to finish it. Cases of MTC have been reported in people treated with liraglutide, another GLP-1 receptor agonist, 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.
Mounjaro's rat study is the same class of result in one species. Males and females received 0.15, 0.50, and 1.5 mg/kg twice weekly for two years, equal to 0.1-, 0.4-, and 1-fold the maximum recommended human dose of 15 mg once weekly. C-cell adenomas increased, statistically significantly, in males at 0.5 mg/kg and above and in females at 0.15 mg/kg and above. Adenomas and carcinomas counted together increased, statistically significantly, in males and females at every dose. A six-month study in rasH2 transgenic mice, at 1, 3, and 10 mg/kg twice weekly, did not find tirzepatide tumorigenic. Semaglutide's warning rests on mice and rats. Tirzepatide's follows the rat study. None of the studies enrolled a person.
A mean of 3.9 years still leaves the rare cancer imprecise
One large human study shows where the gap still is. In a Scandinavian cohort published in the BMJ in 2024, Björn Pasternak and colleagues compared new users of a GLP-1 receptor agonist with new users of a dipeptidyl peptidase-4 (DPP-4) inhibitor, using nationwide registers in Denmark (2007–21), Sweden (2007–21), and Norway (2010–18). After exclusions, 145,410 people started a GLP-1 receptor agonist and 291,667 started a DPP-4 inhibitor. Mean follow-up in the GLP-1 group was 3.9 years. Seventy-six of them developed thyroid cancer, against 184 in the comparator group: 1.33 versus 1.46 cases per 10,000 person-years. The hazard ratio was 0.93, with a 95 percent confidence interval from 0.66 to 1.31.
The authors translate the upper bound as no more than a 31 percent increase in relative risk, or no more than 0.36 extra events per 10,000 person-years against a background of 1.46. They also write that a small increase cannot be excluded. Against sodium-glucose cotransporter 2 inhibitors, a second active comparator, the hazard ratio was 1.16 (0.65 to 2.05). With the first year set aside, the ratio was 0.83 (0.56 to 1.22). Restricted to that first year, it was 1.47 (0.74 to 2.93). They treat that first-year figure as more consistent with closer attention after a prescription than with a cancer that would appear on that schedule.
The cancer in the box is where the study goes thin. Medullary thyroid carcinoma carried a hazard ratio of 1.19 and a confidence interval from 0.37 to 3.86, an interval that runs from a sizable decrease to a nearly fourfold increase. Papillary thyroid cancer was the most common subtype. The paper says estimates for the other subtypes were imprecise because the events were few. A cohort can be informative about thyroid cancer as a category and still have little to say about the rare C-cell cancer the rodents developed.
Which drug, and for how long, matters just as much. Liraglutide was 57.3 percent of GLP-1 use, semaglutide 32.9 percent, dulaglutide 4.9 percent, exenatide 4.1 percent, and lixisenatide 0.9 percent. Mean follow-up after a first semaglutide prescription was 1.1 years. Tirzepatide is not in the cohort. The authors note that cancer can take longer than a mean of 3.9 years to declare itself, even though 25 percent of the GLP-1 group was followed for 6.1 years or longer and the cumulative incidence curves, as the authors describe them, showed no sign of an emerging risk increase out to ten years.
Over that follow-up, GLP-1 receptor agonist use was not associated with a substantially increased risk of thyroid cancer as a whole. The boxes rest on something else: two-year rodent experiments on C-cell tumors, which the Ozempic label also calls lifetime exposure. The cohort is a few years of human thyroid cancer, mostly papillary, with a wide interval on the medullary subtype, a short look at semaglutide, and no tirzepatide. The Ozempic and Mounjaro labels still say the human relevance has not been determined, and they still contraindicate the drugs where MTC or MEN 2 is already part of the history. Semaglutide and tirzepatide carry that warning. Most other peptides in the catalog do not — and where a peptide has no approved label, a missing box is only a missing document.