Meta-analysis flags thyroid and colorectal cancer signals in GLP-1 trials

A meta-analysis published in the journal Diabetes, Obesity and Metabolism looked at 50 randomized controlled trials of GLP-1 receptor agonist drugs, the class that includes semaglutide (Ozempic, Wegovy, Rybelsus) and other medicines used for diabetes and obesity ^[1]. The researchers, led by Giovanni Antonio Silverii of the University of Florence, wanted to know whether these drugs raise or lower the odds of cancer compared with other treatments or placebo, using only trials that lasted at least 52 weeks ^[1].

Across all cancer types combined, the analysis found no meaningful difference between people on GLP-1 drugs and those in comparison groups, with an odds ratio of 1.05 ^[1]. That means, overall, the drugs did not appear to raise cancer risk in these trials. But when the researchers broke the results down by cancer type, two signals stood out. The odds of thyroid cancer were 55% higher in the GLP-1 group, a result that was statistically significant and appeared more pronounced in trials that ran longer ^[1]. The odds of colorectal cancer were 27% higher, but this signal showed up mainly in shorter trials, not longer ones ^[1].

The study also found a benefit for one cancer type. In trials done specifically in people with obesity, uterine cancer risk was lower in the GLP-1 group, with an odds ratio of 0.24 ^[1]. That reduction did not hold up in trials done in people being treated for diabetes rather than obesity, where the odds ratio was close to 1 ^[1]. No significant difference in risk was found for any other cancer studied ^[1].

The authors offer a possible explanation for the colorectal cancer signal that has nothing to do with the drug itself. Because GLP-1 drugs commonly cause gastrointestinal side effects, doctors may have ordered more diagnostic procedures, like colonoscopies, in patients taking these drugs, which could uncover more cancers that would have gone undetected in the comparison group during the same trial period ^[1]. That the signal appeared in shorter trials but not longer ones fits that theory, since a burst of extra screening early on would find existing but previously silent cancers rather than cause new ones to form ^[1].

The thyroid cancer finding is harder to explain away, according to the authors, who say it needs further dedicated research ^[1]. This is not an entirely new concern. Regulators including the Food and Drug Administration and the European Medicines Agency already advise caution in prescribing GLP-1 drugs to patients with a personal or family history of medullary thyroid cancer ^[1].

Why it matters for patients

This analysis pulls together randomized trial data, which is generally considered stronger evidence than observational studies because patients are randomly assigned to treatment or comparison groups. That makes the thyroid and colorectal signals worth taking seriously. But the study also complicates the picture painted by some earlier observational research, which had suggested GLP-1 drugs might lower rates of obesity-linked cancers overall ^[1]. The uterine cancer finding here supports that idea for one cancer type, but the thyroid and colorectal results point the other way.

The overall cancer odds ratio of 1.05, with a confidence interval crossing 1.0, means the trials as a group did not detect a broad increase in cancer risk tied to these medications ^[1]. Patients already know that thyroid cancer risk has been a labeled concern for this drug class. What is new is the colorectal signal, though the authors' own explanation, more scans catching more cancers, suggests this may reflect surveillance patterns rather than a direct drug effect ^[1].

What happens next

The study's authors call for further specific research into the thyroid cancer signal ^[1]. The paper does not give a timeline for such studies, and it is not yet known whether regulators will change existing prescribing guidance based on these findings.