People who took a GLP-1 drug after weight-loss surgery developed kidney cancer slightly less often over the next five years than people who took other diabetes medicines. The gap was real and consistent across every comparison, and it was also very small.
A retrospective study published August 26 in Nephrology Dialysis Transplantation ↗ worked from TriNetX, a network of de-identified US medical records. The team, led by researchers at the University of Texas Southwestern Medical Center in Dallas, looked at adults with type 2 diabetes and obesity who had undergone metabolic and bariatric surgery between 2006 and 2025, then started a glucose-lowering drug. They compared people who began a GLP-1 receptor agonist, the class that includes semaglutide ↗ and tirzepatide ↗, against people who began one of four alternatives: metformin, insulin, an SGLT2 inhibitor, or a sulfonylurea.
The design borrows from a randomized trial without being one. It is called target trial emulation. You pick a clear start point, here the first prescription after surgery following a year with no such drug, match the two groups one-to-one on baseline traits, exclude anyone with a prior cancer, then follow them. After matching there were tens of thousands of pairs, up to 25,025 in the GLP-1-versus-insulin comparison.
The result pointed the same way in every matchup. Over five years, GLP-1 starters had a lower absolute risk of kidney cancer than each comparator: 0.18 percentage points lower than metformin, 0.22 lower than insulin, 0.19 lower than an SGLT2 inhibitor, 0.20 lower than a sulfonylurea. In plainer terms, that is roughly two fewer kidney cancers per thousand people over five years. The direction held across the authors' sensitivity checks.
Real, small, and not yet causal
Kidney cancer is an obesity-linked tumor whose incidence has been climbing, which is why the question is worth asking after an operation that itself lowers some cancer risks. But a difference of about a fifth of a percentage point is small in absolute terms, because kidney cancer is uncommon to begin with. A consistent signal across four comparators is more convincing than any single comparison. Consistency is still not the same as cause.
Two problems sit under any records-database result like this. The first is that people who get and stay on GLP-1 drugs tend to be healthier and more engaged with care than people on older or cheaper regimens, and matching on recorded traits cannot erase what the records never captured. That healthy-user tilt is the same reason a string of GLP-1 database studies keep finding the drugs associated with fewer of almost everything. This section flagged it when GLP-1 drugs showed no melanoma signal ↗ in a similar emulation, and the caution runs the same way when the number lands in the drug's favor.
The second problem is subtler. When you compare two active drugs, a gap can open because your drug protects or because the comparator harms. Insulin and sulfonylureas both carry their own debated links to cancer risk, so a lower number in the GLP-1 arm does not by itself locate the effect in GLP-1.
What it adds
The study lands on reassurance more than discovery. It is another line in the growing file saying GLP-1 drugs do not appear to raise cancer risk, and a hint, not proof, that they might sit on the protective side for at least one obesity-linked tumor. GLP-1 use after bariatric surgery is now common enough to study at this scale. The same population was the subject of a recent finding that GLP-1 drugs reclaimed about half the weight ↗ people regain after the operation. Whether the kidney-cancer edge is the drug, the weight, or the company GLP-1 users keep in a database is the question only a prospective study can settle.