GLP-1 Cancer Fears: Pancreatic vs Thyroid (MTC) Evidence

“Does Ozempic cause cancer?” is really two completely different questions that get tangled together — and the honest answers point in opposite directions. The pancreatic-cancer fear began with a 2011 adverse-event-database hypothesis (Elashoff 2011 ^[1]), but the large cohort studies that followed did not bear it out — they found no increased pancreatic cancer, and several found lower rates (Dankner 2024 ^[3]; Wang 2025 ^[4]; Wang 2024 ^[5]). The thyroid fear is a different organ entirely: rodents given GLP-1 drugs develop thyroid C-cell tumors (Bjerre Knudsen 2010 ^[6]), which is why the FDA boxed warning exists — but the label itself says it is unknown whether this happens in humans ^[7]. These are two organs and two evidence levels. This article keeps them separate so you can see what is actually known versus unknown. For the full cancer review, see does GLP-1 cause cancer?

The honest summary

• These are TWO different fears about TWO different organs. Pancreatic cancer and thyroid (medullary) cancer have nothing to do with each other anatomically, and the evidence for each is at a completely different stage. Treating “cancer” as one question is the central confusion.
• Pancreatic cancer: the fear is largely NOT supported by current data. The original 2011 hypothesis came from a voluntary adverse-event database (Elashoff 2011^[1]); the large cohort and meta-analytic data that followed found no increased risk — and several found a lower rate (Pinto 2019^[2]; Dankner 2024^[3]; Wang 2025^[4]; Wang 2024^[5]).
• Thyroid/MTC: a real rodent finding, human relevance unknown. GLP-1 drugs cause C-cell tumors in rats and mice (Bjerre Knudsen 2010^[6]). The FDA boxed warning is precautionary — it explicitly says it is unknown whether this occurs in humans^[7].
• The MTC warning matters most if you have the specific contraindications. The label contraindicates GLP-1s only in people with a personal or family history of medullary thyroid carcinoma (MTC) or Multiple Endocrine Neoplasia syndrome type 2 (MEN 2)^[7].
• Pancreatitis is not pancreatic cancer. The labels carry a real pancreatitis precaution, but acute pancreatitis (an inflammation) is a separate event from pancreatic cancer (a malignancy) — and even pancreatitis was rare in the large trials (Steinberg 2017^[9]).

Part 1 — Pancreatic cancer: where the fear came from, and what large data actually show

The origin: a 2011 adverse-event-database hypothesis

The pancreatic-cancer worry traces back to 2011, when Elashoff and colleagues^[1] (Gastroenterology) analyzed the FDA's adverse-event reporting database (FAERS) and reported disproportionately more reports of pancreatitis and pancreatic cancer for the early GLP-1 drugs exenatide and sitagliptin. This generated real concern at the time. But a FAERS disproportionality signal is a hypothesis-generating finding, not proof: the database is voluntary, prone to reporting bias (a drug in the news gets more reports), cannot establish a denominator, and cannot prove causation. The right next step is to test the hypothesis in large, controlled cohorts — which is exactly what happened over the following decade.

What the cohort and meta-analytic data found instead

When the question was tested in large populations with proper comparison groups, the signal did not hold up:

• Pinto 2019 (Scientific Reports)^[2] — a meta-analysis with trial sequential analysis pooling randomized GLP-1 trials found no statistically significant pancreatic-cancer signal, and the sequential analysis suggested the accumulated data were already sufficient to rule out a meaningful increase.
• Dankner 2024 (JAMA Network Open)^[3] — a cohort followed for 7 years after GLP-1 initiation concluded plainly: “No support for an increased pancreatic cancer incidence over 7 years following start of GLP-1RA treatment was found.” In years 5–7 the hazard ratio was below 1.0 (0.50; 95% CI 0.15–1.71) versus basal insulin.
• Wang 2025 (Journal of the National Cancer Institute)^[4] — a target-trial-emulation study using real-world data found GLP-1 use was associated with reduced pancreatic-cancer incidence in type 2 diabetes, with hazard ratios of 0.42–0.82 across six comparison drug classes.
• Wang 2024 (JAMA Network Open)^[5] — across 13 obesity-associated cancers, GLP-1 use versus insulin was tied to a markedly lower pancreatic-cancer risk (HR 0.41; 95% CI 0.33–0.50).

Pancreatitis is not pancreatic cancer

A common source of confusion: the labels carry a real pancreatitis precaution, and people conflate that with cancer. They are different things. Acute pancreatitis is an inflammatory event; pancreatic cancer is a malignancy. And even the pancreatitis risk turned out to be small in the large trials — in the 9,340-patient LEADER trial, liraglutide raised lipase enzymes but acute pancreatitis was rare (~0.4%) and not increased over placebo (Steinberg 2017^[9]). See our deep dives on asymptomatic elevated lipase/amylase and restarting a GLP-1 after pancreatitis.

Part 2 — Thyroid / MTC: a real rodent finding, with a precautionary warning

The rodent C-cell tumor finding

The thyroid fear rests on different — and, importantly, real — evidence. Bjerre Knudsen and colleagues^[6] (Endocrinology, 2010) showed that GLP-1 receptor agonists activate GLP-1 receptors on rodent thyroid C-cells, triggering calcitonin release, C-cell hyperplasia, and ultimately C-cell adenomas and carcinomas at clinically relevant exposures in rats and mice. This finding was replicated for semaglutide and tirzepatide in pre-approval rodent toxicology, which is why those labels carry an identical boxed warning. The key scientific caveat is species difference: rodent thyroids have far more C-cells, expressing GLP-1 receptors at much higher density than human C-cells — so whether the rodent mechanism translates to humans is genuinely unknown.

The FDA boxed warning, verbatim

The Wegovy (semaglutide) label carries the FDA's most serious warning category. Reading it exactly as written is clarifying, because the wording is carefully precautionary^[7]:

“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. WEGOVY is contraindicated in patients with a personal or family history of MTC or in patients with Multiple Endocrine Neoplasia syndrome type 2 (MEN 2).”

— FDA Wegovy (semaglutide) Prescribing Information — Boxed Warning (DailyMed)

Read literally, the boxed warning is not the FDA saying “this drug causes cancer in humans.” It is the FDA saying: there is a rodent signal, the human relevance is undetermined, and as a precaution the drug is contraindicated in the small group already at elevated baseline MTC risk. The tirzepatide (Zepbound/Mounjaro) boxed warning is essentially identical and rests on the same rodent data.

What human data do — and don't — show

Human thyroid-cancer evidence is mixed but does not establish causation. The strongest human signal to date is Bezin 2023^[8] (Diabetes Care), a French national nested case-control study (2,562 thyroid-cancer cases vs 45,184 controls) reporting an adjusted hazard ratio of about 1.58 (95% CI 1.27–1.95) for all thyroid cancer and 1.78 (95% CI 1.04–3.05) for medullary thyroid cancer specifically with 1–3 years of cumulative GLP-1 use. That sounds alarming, but it carries well-recognized limitations: detection bias (GLP-1 users see clinicians and get imaged more — “more looking, more finding”), short latency relative to cancer biology, confounding by indication, and the fact that it is a single dataset that has not been replicated. The EMA's Pharmacovigilance Risk Assessment Committee and the FDA both reviewed this signal in 2023–2024 and concluded that a causal association has not been established and that the labeling did not need to change. (Our full cancer review walks through the Bezin study and the regulatory reviews in depth.)

Two fears, side by side

What this means for you

• Worried specifically about pancreatic cancer? The large controlled data are reassuring — they did not confirm the 2011 hypothesis, and several found lower rates. If you have a personal history of pancreatic cancer or pancreatitis, that is a conversation to have with your oncology/care team, but it is driven by the pancreatitis precaution, not a proven cancer link.
• Worried about thyroid cancer? The decisive question is whether you have a personal or family history of MTC or MEN 2. If yes, GLP-1s are contraindicated. If no, the residual risk — if any — is considered small enough by regulators that the warning is precautionary, not a contraindication.
• Don't conflate the two. The single biggest mistake in this topic is treating “does it cause cancer” as one question. They are different organs, different evidence, different conclusions.
• Symptoms still matter. Report a new neck mass, persistent hoarseness, trouble swallowing or breathing (thyroid), or severe persistent upper-abdominal pain with vomiting (pancreatitis) to your clinician — regardless of how reassuring the population data are.

Bottom line

The “GLP-1 cancer” fear is really two fears about two organs. The pancreatic-cancer worry began with a 2011 adverse-event-database hypothesis (Elashoff^[1]) that the large cohort and meta-analytic data that followed did not confirm — they found no increase, and several found lower rates (Pinto 2019^[2]; Dankner 2024^[3]; Wang 2025^[4]; Wang 2024^[5]). The thyroid/MTC concern is a genuine rodent finding (Bjerre Knudsen^[6]) behind a precautionary FDA boxed warning that explicitly states the human relevance is unknown^[7]; it matters mainly if you have a personal or family history of MTC or MEN 2. Bottom line: the pancreatic-cancer fear is largely not supported by current data, and the MTC warning is precautionary and chiefly relevant to a specific high-risk group.

This article is educational and is not medical advice. Every claim above is sourced to a peer-reviewed study indexed in PubMed (verified against the live PubMed database) or to the FDA-approved label (verified verbatim on DailyMed) before publication. Decisions about starting, continuing, or stopping a GLP-1 — especially with any personal or family history of cancer — should be made with your prescribing clinician.