Across 97,464 people who started semaglutide in Denmark, Sweden and Norway, the drug carried no higher risk of pancreatic cancer than older diabetes medicines. The number worth holding onto is not the reassurance. It is how briefly anyone was watched.
Semaglutide ↗ is the molecule sold as Ozempic and Wegovy. It mimics a gut hormone that acts on the GLP-1 receptor ↗, telling the brain you are full and nudging the pancreas to release insulin. Because part of its job happens in and around the pancreas, pancreatic safety has shadowed the whole drug class since the first GLP-1 medicines arrived over a decade ago. European regulators took the worry seriously enough to order the maker, Novo Nordisk, to run a dedicated study when semaglutide was approved. This is that study, presented at the European Association for the Study of Diabetes meeting ↗ in Milan.
The design leaned on something the Nordic countries have and most places do not: nationwide health registries that track every prescription and every cancer diagnosis. Researchers matched semaglutide starters against people who started other, non-incretin diabetes drugs, the sulfonylureas, SGLT-2 inhibitors and insulin that predate the GLP-1 era. To keep a cancer that was already brewing from being blamed on a drug someone just started, they ignored the first year on treatment and only counted diagnoses after that.
The result was flat in all three countries. Semaglutide users were no more likely to develop pancreatic cancer than the comparison group, with hazard ratios of 1.00 in Denmark, 0.82 in Sweden and 0.91 in Norway. A hazard ratio of 1.0 means no difference; all three confidence intervals comfortably crossed it, and the pooled estimate landed at 0.91 (95% confidence interval 0.71 to 1.16), a hair under even odds. In raw counts, 131 pancreatic cancers turned up among semaglutide users and 123 among the comparators. Taking more of the drug, or staying on it longer, did not push the risk up.
Here is the part the headline number cannot carry. After that one-year exclusion window, people were followed for a mean of only 1.40 to 1.85 years. Pancreatic cancer is a slow, quiet disease. It usually takes years to form and is typically caught late, which is exactly why it is so feared. A drug started this decade could raise a risk that does not surface until five or ten years out, and a study that watches for under two years will see none of it. This analysis is strong evidence that semaglutide does not cause pancreatic cancer quickly. It is silent, by construction, on whether it does so slowly.
Two more things belong in the frame. The study was funded by Novo Nordisk, the company that sells the drug and that regulators required to answer the question. And it was an observational comparison, not a randomized trial, so matching can balance the factors researchers can see and nothing about the ones they cannot. The work was led by Oystein Karlstad of the Norwegian Institute of Public Health and presented by Anton Pottegard of the University of Southern Denmark, both established pharmacoepidemiologists whose registry methods are about as good as this kind of study gets.
For the tens of millions of people now on a GLP-1 drug, the practical read is narrow and real: the near-term pancreatic cancer scare that has trailed this class does not show up in the largest, cleanest dataset assembled to look for it. What the study cannot do is close the long-latency question, and no one presenting it claimed otherwise. That answer only arrives with a longer clock. Semaglutide and its receptor sit at the center of peptidemodel's GLP-1 receptor ↗ map; this is the safety readout the class has been waiting for, with the asterisk that matters printed in the follow-up column.