Researchers at UT Southwestern Medical Center used electronic health records from 359,595 adults seen there between January 2010 and July 2025. Among those with type 2 diabetes and no previous obesity-related cancer, they compared people who started a GLP-1 receptor agonist with people who started metformin, insulin, an SGLT2 inhibitor, a DPP-4 inhibitor or a sulfonylurea, in five separate analyses designed to mimic head-to-head trials. The groups were weighted to balance age, health conditions, prior treatment and other factors, and the analysis was repeated by how advanced each person’s diabetes was.[1]
What they found at three years
- Against insulin: 2.01% of GLP-1 starters were diagnosed with an obesity-related cancer, against 3.26% of insulin starters, a hazard ratio of 0.62 (95% CI 0.49 to 0.77).[1]
- Against sulfonylureas: hazard ratio 0.67 (95% CI 0.54 to 0.84).[1]
- Against metformin, SGLT2 inhibitors and DPP-4 inhibitors: the authors write that “estimates were closer to null” for these comparisons. The hazard ratios were 0.78 (0.58 to 1.05) against metformin and 0.81 (0.52 to 1.25) against SGLT2 inhibitors, and both intervals include no difference.[1]
- At five years, the hazard ratio against insulin was similar, at 0.61 (0.50 to 0.75).[1]
Why the insulin result needs caution
The authors point out that insulin is often started later in type 2 diabetes, and that insulin starters in their data generally had more advanced complications. Differences in how long and how badly blood sugar had been raised, in access to care and in cancer screening could remain even after weighting. They list single-system records, residual confounding and sparse events in some subgroups as limitations, and say the study identified where associations were clearest rather than proving that any drug is better for cancer outcomes.[1]
Their conclusion is that the findings are broadly reassuring on cancer safety but do not support choosing a diabetes drug to prevent cancer; the established heart, kidney and blood-sugar benefits should still guide treatment.[1] The work was funded by the National Cancer Institute, the National Institutes of Health and other public grants, and one author reported personal fees from several drugmakers, including Eli Lilly and Novo Nordisk.[1]
For two cancers where the GLP-1 evidence has been reviewed in more depth, see our pages on GLP-1 drugs, colorectal cancer risk and screening and on GLP-1 drugs and endometrial cancer. The separate thyroid-cancer warning on GLP-1 labels is covered in our review of GLP-1 drugs and thyroid cancer.