A cancer history changes how a boxed warning reads. Someone treated for breast cancer eight years ago, or colon cancer at 45, opens a label, finds the word carcinoma inside a black box, and reasonably concludes the drug is not for them. On the evidence as it stands, that conclusion is usually wrong for the wrong reason — and the reassurance that replaces it is also weaker than it looks.
What the contraindication actually names
Read on DailyMed on September 15, 2026, the contraindications section of the Wegovy label has exactly two entries: “a personal or family history of MTC or in patients with MEN 2,” and a prior serious hypersensitivity reaction to semaglutide or an excipient. The Zepbound label has the same two. MTC is medullary thyroid carcinoma, a tumor of the calcitonin-producing C-cells; MEN 2 is the inherited syndrome that causes most of the familial cases. The reasoning behind the warning — rodent C-cell tumors and a receptor that primates barely express on those cells — is set out in the thyroid cancer article and is not re-argued here.
What matters for a survivor is what the labels do not say. Across all four manufacturer documents — Wegovy, Ozempic, Zepbound and Mounjaro — the word “malignancy” does not appear at all, and every occurrence of the word “cancer” refers to medullary thyroid carcinoma. There is no sentence about breast cancer, no sentence about a solid tumor in remission, no sentence about active chemotherapy, and no sentence excluding anyone on the basis of a cancer history of any other kind. The absence is not permission. It is silence.
The population the contraindication removes is small, and it is not the population most people picture. Medullary thyroid carcinoma accounted for 3,833 diagnoses and 536 deaths across the SEER 17 and 12 registries between 2000 and 2020, with incidence rising at an average annual percentage change of 1.64 and mortality at 3.46.[1] Most people excluded by that clause have never had cancer themselves — the clause covers family history too, which is why the question belongs on every intake form rather than only in front of survivors.
Both labels then decline to recommend screening. Routine monitoring of serum calcitonin or routine thyroid ultrasound is described as of uncertain value for early detection, and as something that may increase the risk of unnecessary procedures given the low specificity of the test and the high background rate of thyroid disease. The labels add that patients with MTC usually have calcitonin values above 50 ng/L.
The reassuring cohorts excluded the reader they are quoted at
Three large observational studies now report lower cancer incidence among people taking these drugs, and they are repeated constantly. All three removed anyone with a cancer history as a condition of entry.
The largest drew on electronic records covering 113 million US patients and analyzed 1,651,452 adults with type 2 diabetes who had no prior diagnosis of any of 13 obesity-associated cancers.[2] A target-trial emulation in a Florida research network matched 86,632 adults without prior cancer history, 43,317 on these drugs and 43,315 not, and reported 14 cancers at 13.6 against 16.4 per 1,000 person-years, hazard ratio 0.83 (95% CI 0.76 to 0.91).[3] A third emulation restricted to adults with obesity and without diabetes matched 161,798 patients without a prior obesity-associated cancer and, over a median two years, found any such cancer at hazard ratio 0.59 (95% CI 0.53 to 0.67).[4]
Those are large, consistent and genuinely interesting numbers about cancer incidence in people who have never had cancer. They carry no information about recurrence, about a second primary tumor, or about anyone in survivorship care, because the study designs took those people out first.
The direction depends on which drug the comparison uses
The 13-cancer analysis is also where this literature stops being tidy. Against insulin, GLP-1 receptor agonists were associated with lower risk in 10 of the 13: gallbladder cancer at hazard ratio 0.35 (95% CI 0.15 to 0.83), meningioma 0.37, pancreatic cancer 0.41 (0.33 to 0.50), hepatocellular carcinoma 0.47, ovarian cancer 0.52, colorectal cancer 0.54 (0.46 to 0.64), multiple myeloma 0.59, esophageal cancer 0.60, endometrial cancer 0.74 and kidney cancer 0.76. Postmenopausal breast cancer and thyroid cancer did not move.[2]
Against metformin, in the same patients, not one of those cancers showed a significant reduction — and kidney cancer went the other way, at hazard ratio 1.54 (95% CI 1.27 to 1.87).[2] The Florida emulation found the same direction on the same organ against untreated controls: kidney cancer at 1.38 (95% CI 0.99 to 1.93), an interval that touches one.[3]
That is the shape worth carrying away. Insulin is a comparator with its own associations, and a drug that looks protective against it may be doing nothing more than not being insulin. The paper carries a published correction, and its authors describe the findings as preliminary evidence supporting further study rather than as a prevention claim. A page that quotes the insulin column and stops has reported real figures and drawn a false picture.
The one cohort built from survivors
A single-institution retrospective study identified 1,022 patients with nonmetastatic breast cancer — ductal carcinoma in situ or stage 1 to 3 — who received a GLP-1 receptor agonist between 2005 and 2024. Seventy-nine percent had type 2 diabetes. Median weight at initiation was 86.8 kg and median body-mass index 33.5.[5]
Among the 442 on semaglutide or tirzepatide, median weight change was −1.9% at three months, −3.1% at six and −2.6% at twelve, with individual results spanning −27.8% to +11.5%.[5] Set that against the roughly 15% reported in the registration trials, summarized in the semaglutide weight-loss article. Endocrine therapy, which most hormone-receptor-positive survivors take for five to ten years, was associated with weight gain in the same analysis, and the twelve-month figure was smaller than the six-month one.
Then the survival result, which has to be read carefully in both directions. Overall survival differed sharply between 810 users and 1,620 matched non-users: hazard ratio 0.37 (95% CI 0.27 to 0.53, P < 0.0001). Disease-free survival was not associated with treatment at all.[5]
A drug acting on breast cancer would be expected to move the cancer endpoint. This one moved all-cause death by nearly two-thirds and left disease recurrence untouched, in a retrospective cohort at one cancer center where the people prescribed a weight-loss drug were by definition well enough to be prescribed one. That pattern is what confounding by indication looks like, and the authors call for clinical trials rather than claiming an effect. It is not evidence of harm either.
What nobody has studied
A narrative review from the National Cancer Institute reaches the same place from the other side: the literature on these drugs in cancer survivors is limited and largely retrospective, and there is a notable absence of well-designed randomized trials measuring weight, clinical endpoints or cancer outcomes in this population.[6] It names the interactions nobody has characterized — between these drugs and anticancer therapy, and their joint effect on treatment efficacy and nutritional status — and flags two groups for caution: people with cancer cachexia, and people with sarcopenic obesity, where the muscle that comes off matters more than the fat. The trial figures for lean-mass loss are in the muscle article.
Several specific questions have no data at all. Whether these drugs change the pharmacokinetics of an oral targeted therapy by slowing gastric emptying. Whether delayed emptying interacts with chemotherapy-induced nausea. Whether weight loss during endocrine therapy changes recurrence risk. Whether anything about the class affects surveillance imaging. Each of those is a real clinical question, and for each the honest answer is that no study has reported it.
What this means in practice
The narrow, defensible reading is this. A history of most cancers is not a labeled contraindication, and no label sentence excludes it. A history of medullary thyroid carcinoma, or a family history of it, or multiple endocrine neoplasia type 2, absolutely is, and that rule is treated as absolute regardless of everything else. The incidence findings in cancer-free populations do not transfer to survivorship, and the one survivor cohort in existence reports a fifth of the weight loss the trials did, alongside a survival figure that its own disease-free analysis argues against reading as a cancer effect.
One market fact belongs here too. Most cash-pay sellers dispense compounded semaglutide or tirzepatide, which is not FDA-approved and is not reviewed by the FDA for safety, efficacy or quality before it is dispensed — the distinction set out in the compounding article and what “not FDA-approved” means. A compounded vial does not arrive with the labeling quoted throughout this page, so the document a prescriber would use to check a cancer history is not in the box, and the sellers listed on the price boards are generally not the people managing survivorship care.
The decision belongs to an oncologist who knows the tumor type, the current therapy and the nutritional picture. What this page can establish is the size of the gap that decision is being made across.