The boxed warning is about rats, and the human studies disagree with each other

What this comes down to

• The boxed warning states that it is unknown whether these drugs cause thyroid C-cell tumors, including medullary thyroid carcinoma, in humans, because the human relevance of the rodent finding has not been determined.
• The same warning makes a personal or family history of medullary thyroid carcinoma, or Multiple Endocrine Neoplasia syndrome type 2, a contraindication. That is a hard stop, not a caution.
• A French nested case-control study of 2,562 thyroid cancer cases reported an adjusted hazard ratio of 1.58 for all thyroid cancer after one to three years of use. A Scandinavian cohort of 145,410 users reported 0.93, with an upper confidence limit its authors read as no more than a 31% increase.
• The label also says routine monitoring of serum calcitonin or thyroid ultrasound is of uncertain value for early detection in patients treated with these drugs.

What the box actually says

The wording is close to identical across the class, which is itself informative — it is a class warning rather than a finding about one product^[1].

“In rodents, semaglutide causes dose-dependent and treatment-duration-dependent thyroid C-cell tumors at clinically relevant exposures. It is unknown whether WEGOVY 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.”

^[1]

The tirzepatide label says the same thing with rats in place of rodents and its own brand name in place of the other^[2]. Both then do something stronger than warn: they contraindicate the drug in patients with a personal or family history of medullary thyroid carcinoma or with Multiple Endocrine Neoplasia syndrome type 2. A contraindication is not a risk to weigh with a prescriber. It is a line the label says not to cross, and an intake form that never asks about your family's thyroid history has not collected the information that line depends on. How a telehealth prescription actually happens is where that question should appear.

Two large studies, two answers

The French study was a nested case-control analysis in a national diabetes cohort. It included 2,562 thyroid cancer cases matched to 45,184 control subjects, and reported that use of a GLP-1 receptor agonist for one to three years was associated with an adjusted hazard ratio of 1.58 for all thyroid cancer and 1.78 for medullary thyroid cancer^[3].

The Scandinavian study was an active-comparator new-user cohort across Denmark, Norway and Sweden from 2007 to 2021, comparing people starting a GLP-1 receptor agonist against people starting a DPP-4 inhibitor. Seventy-six thyroid cancers occurred among 145,410 GLP-1 users and 184 among 291,667 DPP-4 users. The hazard ratio was 0.93, and the authors state that the upper limit of the confidence interval was consistent with no more than a 31% increase in relative risk^[4].

Those two results are not reconcilable by picking the one you prefer. They differ in design in a way that points at the most likely explanation: people who start one of these drugs get looked at more, and thyroid cancer is the tumor most sensitive to being looked for. The Scandinavian study chose an active comparator — another diabetes drug — precisely to hold that kind of surveillance roughly constant between the groups. What would settle it is a randomized trial with this cancer as an endpoint, and no such trial is indexed^[5].

(semaglutide[tiab] OR tirzepatide[tiab]) AND "medullary thyroid"[tiab] AND "randomized controlled trial"[pt] AND "humans"[mh]

How to read it as a buyer

• The contraindication is the actionable part. If medullary thyroid carcinoma or MEN 2 is in your family, the label already answers the question, and a prescriber who never asked has skipped a step.
• The boxed warning is not evidence of human harm and it is also not decoration. It is the FDA recording an animal finding whose human relevance is undetermined, which is a different statement from either of the ones marketing makes about it.
• The observational disagreement is the honest state of the evidence. Anybody quoting one of those two studies without the other is selling you a number, not a finding.
• The same labels carry other warnings that are far better quantified, including the gastrointestinal ones and the hair loss rates, with their sex split. Side effects by how often the trials reported them is where the rest of those numbers live.

Questions people actually ask