Semaglutide linked to more benign growths, not cancer, in diabetes trials

Semaglutide has become one of the most widely used medicines in the world, prescribed alongside diet and exercise to improve blood sugar control in adults with type 2 diabetes. From early on, questions have circulated about whether the drug might encourage abnormal tissue growth, and that debate has never fully settled. Individual trials are rarely large enough to detect uncommon outcomes such as tumours, which is exactly the situation where pooling every available randomised trial can give a clearer answer than any single study. This analysis set out to do that pooling properly.

The researchers followed PRISMA 2020 guidance and searched Web of Science, PubMed, Embase and ScienceDirect for randomised controlled trials of semaglutide in adults with type 2 diabetes. Nineteen trials conducted between 2016 and 2025 qualified, together enrolling 38,160 patients. Both the injectable and the oral form of the drug were represented, with treatment lasting between 26 and 156 weeks. The team pooled the odds of overall, benign and malignant neoplasms, ran sensitivity analyses to check whether any single trial drove the result, and tested for publication bias using funnel plots alongside Begg's and Egger's tests.

Semaglutide use was associated with higher odds of developing any neoplasm, with a pooled odds ratio of 1.19, and higher odds of benign neoplasms specifically, at 1.28. For malignant neoplasms the pooled odds ratio was 1.12, which did not reach statistical significance, so a genuine link with cancer could not be confirmed. The signal, in other words, was concentrated in benign growths rather than malignancies. No single influential trial explained the pattern, and the authors found no evidence of publication bias, which strengthens confidence that the numbers reflect the trials as a whole.

Whether semaglutide encourages tumour formation has prompted considerable debate, and this pooled analysis adds structured randomised evidence to a conversation often dominated by single studies and anecdote. Its answer is nuanced: a modestly increased risk of overall and benign neoplasms, with no confirmed rise in malignant ones. The included trials ran for at most 156 weeks, which is short on the timescale over which cancers develop, so the analysis cannot rule out effects that only emerge over many years. The authors themselves call for further long-term randomised trials to clarify what the signal means clinically.

If you take semaglutide or are considering it, this is context rather than cause for alarm. The clearest signal in the pooled data concerned benign growths, and the association with malignant tumours fell short of statistical significance. At the same time, the finding is a reminder that even very popular drugs carry open safety questions, and the honest answer on long-term risk is that trials have not yet run long enough to give one. It is the sort of information worth having in mind for a conversation with a prescriber, not a reason to change anything on your own.