Why some patients worry about GLP-1 drugs and thyroid cancer

GLP-1 receptor agonists (such as semaglutide, liraglutide, etc.) are now widely used for managing type 2 diabetes, and also for weight loss in some settings. But their rise in use has reawakened a long-standing question: Could they increase a person’s risk of thyroid cancer?

That worry originates from older animal studies (in rodents) showing that GLP-1 drugs triggered a form of thyroid cancer called medullary thyroid carcinoma (MTC). Because of those findings, the FDA requires warnings that these drugs not be used in people with a personal or family history of MTC or MEN2 (multiple endocrine neoplasia type 2).

However, translating findings from animals to humans is tricky. Over the years, human studies have been inconclusive, and doctors and patients have been left without definitive guidance.

What the new study found: detection bias, not causation

A new, well-designed study is helping to clarify things. It was published recently in JAMA Otolaryngology – Head & Neck Surgery, and was summarized by Mayo Clinic.

Here are the key takeaways in everyday terms:

1. No clear increase in thyroid cancer risk overall

• The researchers compared a large group (351,000+ adults with type 2 diabetes) who started GLP-1RAs to those who began other diabetes medications (SGLT-2 inhibitors, DPP-4 inhibitors, sulfonylureas).
• Over the full follow-up period, there was not a statistically significant increase in the risk of being diagnosed with thyroid cancer among the GLP-1RA users. The hazard ratio was 1.24 (95% confidence interval 0.88–1.76), which crosses the “no increased risk” line.
• In absolute numbers, thyroid cancer was still very rare: it occurred in about 0.17% of the ~41,100 people who started a GLP-1RA.
  So, in plain language: starting a GLP-1RA does not appear to substantially raise the long-term risk of developing thyroid cancer.

2. A “spike” in first-year diagnoses

—but there was a twist. When the data were broken down by time:

• In the first 12 months after starting a GLP-1RA, there was a statistically higher rate of new thyroid cancer diagnoses (hazard ratio ~1.85). Mayo Clinic
• But in the 1–2 year period, and beyond 2 years, that elevated risk vanished (i.e. no significant association). Mayo Clinic
• That pattern — a front‐loaded “bump” early on — doesn’t match how cancer biology typically works (it usually takes years of changes). So the authors concluded this was likely not because GLP-1RAs caused new cancers. Mayo Clinic

3. Detection (surveillance) bias is the likely explanation

The authors make a persuasive case that detection bias is at play:

• People starting GLP-1RAs may receive more attention, more frequent follow-ups, and more imaging (ultrasounds), especially in light of existing concerns. The FDA warning likely makes both patients and doctors more vigilant.
• Indeed, the study found that at 6 and 12 months, GLP-1RA users were more likely to have thyroid ultrasounds: ~2.1% by 12 months vs. ~1.5% in the comparator drug group.
• That “extra search” can lead to finding small, subclinical thyroid nodules or cancers that would otherwise have gone unnoticed. In other words, the drug didn’t cause the cancer — we just looked harder and found something that was already there.

Thus the authors argue that the higher first-year diagnoses are better explained by “we looked more in those patients” rather than “the drug caused new thyroid cancers.”

What this means for patients (and your thyroid)

Here’s how you can interpret these findings in the real world:

• For most people, especially those without a personal/family history of medullary thyroid cancer or MEN2, this study is reassuring. The metabolic and cardiovascular benefits of GLP-1RAs (for appropriate patients) likely outweigh the very small and likely spurious risk of thyroid cancer.
• The study does not support routinely screening everyone who begins a GLP-1RA with thyroid ultrasound or other imaging solely out of cancer concerns.
• If a thyroid nodule or abnormality is found in someone taking a GLP-1RA, it should be handled the same way as for anyone else — with standard guidelines (ultrasound, fine-needle biopsy as indicated, etc.), rather than assuming it’s drug-caused.
• If you have a known history or family risk of medullary thyroid cancer or MEN2, your doctor will likely continue to avoid GLP-1RAs (as per FDA labeling), because the animal data and stronger biological plausibility still warrant caution in that subgroup.

What remains uncertain & future directions

• Even though this is one of the strongest real-world human studies so far, it’s still observational, not a randomized trial. Some residual confounding can’t be completely ruled out.
• The study focused on diagnosed thyroid cancer, not on molecular changes or extremely early “micro-lesions” that may never become clinically relevant.
• Longer follow-up in diverse populations will help further confirm (or refine) these findings.
• Additional research may explore whether certain subgroups (older age, radiation history, thyroid disease, etc.) show different risks.

Tips for patients with thyroid concerns who are on (or considering) GLP-1 therapy

1. Talk openly with your endocrinologist or primary care physician — mention any personal/family thyroid cancer history, previous thyroid nodules, or risk factors.
2. Don’t demand or expect routine ultrasounds just because you’re starting a GLP-1 drug. If there’s a specific symptom or prior thyroid abnormality, the doctor can evaluate as usual.
3. Let standard thyroid workup guide decisions — e.g. if a nodule is found, use ultrasound and biopsy (if indicated) per standard guidelines, rather than attributing it automatically to the drug.
4. Focus on benefits vs. risks — GLP-1RAs have clear benefits for glucose control, weight, and cardiovascular outcomes in the appropriate patients. This new research helps put the thyroid concern into perspective.

Brito JP, Herrin J, Swarna KS, et al. “GLP-1RA Use and Thyroid Cancer Risk.” JAMA Otolaryngology–Head & Neck Surgery. 2025;151(3):243–252. doi:10.1001/jamaoto.2024.4852