Cancer questions and the GLP-1 labels

What the labels contain, and what they do not

None of the US labels we reviewed for semaglutide, liraglutide, tirzepatide or orforglipron says the medicine causes cancer in people, and none claims it protects against cancer. What they contain is narrower: the thyroid C-cell boxed warning, some case counts from the liraglutide trials, and a section on animal studies ^{[1][2][3][4][5]}. The thyroid warning, the MTC exclusion and the calcitonin debate are covered in the thyroid article; everything below is in addition to that.

Reading the liraglutide counts

Breast cancer in the Saxenda® trials included invasive cancers (13 women on liraglutide, 2 on placebo) and ductal carcinoma in situ, an early form confined to the milk ducts (4 and 1). Most were positive for estrogen and progesterone receptors. The label says the cases were too few to determine whether they were related to the medicine, and that the data are not enough to know whether it affects a breast tumor that is already present ^[1].

Papillary thyroid cancer is a different type from MTC. In the Saxenda® trials, four of the eight tumors were smaller than 1 cm and four were found in tissue from thyroid operations prompted by findings that predated treatment ^[1]. In the Victoza® trials, most were under 1 cm and were found after surgery prompted by calcitonin or ultrasound screening that the study protocol required ^[2]. Looking closely at the thyroid tends to uncover small cancers, which makes the medicine’s own contribution hard to separate; the authors of a 2025 meta-analysis raised a similar point about colorectal cancer ^[8].

The colorectal counts come with no conclusion from the label, and the single placebo case was a different tumor type, a neuroendocrine tumor of the rectum ^[1]. These were weight-management and diabetes trials, not studies built to measure cancer.

Animal studies beyond the thyroid

• Semaglutide did not damage genes or chromosomes in standard laboratory tests ^[3]. The EU product information adds that no treatment-related tumors other than the thyroid C-cell findings were seen ^[6].
• Tirzepatide was not genotoxic in the test reported and caused no tumors in a six-month study in genetically modified (rasH2) mice ^[4].
• Liraglutide caused fibrosarcomas, tumors of connective tissue, in the skin at the injection site of male mice at the highest dose; the US label attributes them to the high local drug concentration ^[1]. The EU document states that no other treatment-related tumors were found and that the thyroid tumors did not appear in monkeys treated for 20 months ^[7].
• Orforglipron (Foundayo™) was not carcinogenic in rats or transgenic mice and did not damage genes in testing, but it is inactive at the rodent GLP-1 receptor, so these studies cannot reveal a receptor-driven effect ^[5].

What larger studies report

Trials submitted for approval are too small and too short to settle cancer questions on their own, so researchers pool many trials or study health records. Each design has limits that matter here.

• Meta-analysis of randomized trials, 2025: 50 trials of at least 52 weeks comparing a GLP-1 medicine with placebo or another treatment. Overall cancer did not differ (odds ratio 1.05). Thyroid cancer was more frequent (1.55), more clearly in longer trials. Colorectal cancer was also more frequent (1.27), but only in shorter trials, which the authors thought might reflect extra diagnostic tests prompted by stomach side effects. Uterine cancer was less frequent in the obesity trials ^[8].
• US health-record cohort, 2024: among 1.65 million people with type 2 diabetes, those prescribed a GLP-1 medicine had lower rates of 10 of 13 obesity-associated cancers than those prescribed insulin. Against metformin, no cancer was less common, and kidney cancer was more common (hazard ratio 1.54). Postmenopausal breast cancer and thyroid cancer were not lower in either comparison ^[9].
• Israeli population cohort, 2024: in 543,595 adults with diabetes, GLP-1 use compared with basal insulin gave no support for a rise in pancreatic cancer over seven years; the authors call for monitoring beyond that period ^[10].
• National Cancer Institute summary: it notes that observational findings on weight loss and cancer need caution, and that a meta-analysis of randomized trials with about three years of follow-up found no association between GLP-1 medicines and gastrointestinal cancers ^[11].

Why these results are hard to read

• Small numbers: a handful of cases in either group moves a percentage a lot, which is why the labels speak of too few cases to determine a relationship ^[1].
• Detection: people on these medicines may have more tests, either built into a trial or prompted by side effects, and more testing finds more early cancers ^[2][8].
• Comparison group: the same health-record data looked protective against insulin and mostly neutral, with more kidney cancer, against metformin, so the comparator shapes the answer ^[9].
• Time: cancers develop over years, and the pancreatic study asks for follow-up beyond seven years ^[10].

Obesity, diabetes and the background risk

According to the National Cancer Institute, overweight and obesity are linked to at least 13 types of cancer, with the largest increases for endometrial and esophageal cancer. Among the explanations it gives are the extra estrogen made by fat tissue and the high insulin levels that often come with obesity ^[11].

Many people prescribed GLP-1 medicines therefore start with a higher background risk. Most evidence that losing weight lowers cancer risk comes from observational studies, which NCI says should be read with caution, and weight loss after bariatric surgery has also been associated with lower cancer risk ^[11]. Randomized trials of GLP-1 medicines have not shown an overall change in cancer in either direction ^[8].

Symptoms that overlap with side effects, and what to tell your prescriber

Some warning signs of colorectal cancer overlap with common GLP-1 side effects: diarrhea, constipation, a change in bowel habits and lower abdominal pain. MedlinePlus advises contacting a provider about black or tarry stools, blood during a bowel movement, a change in bowel habits or weight loss with no known reason ^[12]. On a medicine meant to cause weight loss the last sign is harder to judge, so discuss any change that surprises you, and do not assume that blood or black stools are a side effect.

Before starting, tell your prescriber about any personal or family history of cancer. Keep routine screening appointments as your clinician advises; MedlinePlus notes that colon cancer screening is recommended for adults from age 45 ^[12]. A new breast lump, nipple discharge or dimpled breast skin is a reason to contact a provider whatever medicine you use ^[13].

Saxenda®, Victoza® and Wegovy® are registered trademarks of Novo Nordisk A/S. Zepbound® is a registered trademark and Foundayo™ is a trademark of Eli Lilly and Company. PepLabo is not affiliated with or endorsed by Novo Nordisk or Eli Lilly and Company.