Landmark Nordic Registry Study Rules Out Pancreatic Cancer Risk From Semaglutide

One of the most persistent safety questions surrounding the world’s most famous weight-loss and diabetes drug has now been put to a rigorous, population-scale test — and the answer is reassuring. A regulator-mandated post-authorisation safety study, presented at the Annual Meeting of the European Association for the Study of Diabetes (EASD) in Milan, Italy, from September 28 to October 2, has found no link between the use of semaglutide and pancreatic cancer. The research, led by Professor Anton Pottegård of the University of Southern Denmark in Odense and colleagues, drew on complete nationwide health registry data from Denmark, Sweden and Norway, comparing more than 97,000 new users of semaglutide with matched patients taking other classes of glucose-lowering medication. The study was funded by Novo, the manufacturer of semaglutide, and was required by regulators as part of the drug’s original approval process, making it one of the most comprehensive investigations of this suspected risk ever conducted.

Semaglutide, marketed under the brand names Wegovy and Ozempic, belongs to a class of medicines known as glucagon-like peptide-1 receptor agonists, or GLP-1RAs. These drugs mimic the action of the gut hormone GLP-1, which stimulates insulin secretion in a glucose-dependent manner, slows gastric emptying and reduces appetite. They have transformed the treatment landscape for type 2 diabetes mellitus and, more recently, for weight management, achieving unprecedented uptake across the globe. Yet almost from the moment incretin-mimetic drugs entered clinical practice, a shadow of concern has followed them. The pancreas sits at the centre of that concern, because GLP-1 signalling acts directly on pancreatic tissue, and the organ’s central role in both digestion and glucose regulation means that any theoretical harm to it would carry serious consequences.

The hypothesis that GLP-1 receptor agonists might raise the risk of pancreatic cancer did not emerge from clinical trials but from a patchwork of early warning signs. Initial safety signals appeared in spontaneous adverse event reporting systems, where physicians occasionally filed reports of pancreatic inflammation and other pancreatic disorders in patients using these drugs. Preclinical studies in rodents added to the anxiety, revealing cellular changes in pancreatic tissue that raised the possibility of a pathway toward malignant transformation. Early case reports also linked GLP-1 medications to acute pancreatitis, a painful and sometimes dangerous inflammatory condition of the gland. Because chronic inflammation is a well-established driver of cancer development in many tissues, researchers and regulators alike questioned whether repeated or persistent pancreatic stimulation by incretin-mimetic drugs could, over years of treatment, eventually elevate the risk of pancreatic cancer — one of the deadliest and most difficult-to-detect malignancies in human medicine.

It was against this backdrop that regulators, as part of the overall original approval process for semaglutide, mandated a post-authorisation safety study to estimate the risk of pancreatic cancer associated with the drug. The design chosen was an active comparator, new user cohort study, a methodological approach widely regarded as the gold standard for assessing drug safety in real-world populations. Rather than comparing semaglutide users with people taking no diabetes medication at all — a comparison that would be hopelessly confounded by differences in disease severity — the investigators compared new users of semaglutide for type 2 diabetes with matched new users of non-incretin antidiabetic drugs, specifically sulfonylureas, SGLT-2 inhibitors or insulin. These comparator drugs are typically used at a similar stage of disease as semaglutide, which helps ensure that the two groups are genuinely comparable in terms of diabetes duration, severity and other characteristics that might independently influence cancer risk.

The scale of the data infrastructure behind the study is one of its defining strengths. Denmark, Sweden and Norway each maintain comprehensive national health registries that capture essentially every prescription filled, every hospital diagnosis and every cancer diagnosis for their entire populations, linked at the individual level through unique personal identifiers. By pooling these registries, the researchers assembled a cohort of 97,464 semaglutide users. Eligible patients had to initiate semaglutide or one of the active comparators and fill at least two prescriptions within the first year of treatment, a requirement designed to ensure that participants were genuinely exposed to their assigned therapy. Follow-up began one year after initiation, an intentional lag period that guards against protopathic bias — the possibility that early symptoms of an undiagnosed pancreatic tumour might prompt a physician to prescribe a new diabetes drug, creating a spurious association between the medication and a cancer that was already developing.

After matching, patient characteristics were well balanced between the semaglutide group and the active comparator cohort. Among the semaglutide users, median age ranged from 59 to 62 years across the three countries, and mean follow-up after the one-year lag period was between 1.40 and 1.85 years. During the observation window, 131 incident pancreatic cancer events occurred among semaglutide users and 123 among active comparator users, accumulated over 167,399 and 140,306 person-years of follow-up respectively. The incidence rates were strikingly similar between the two groups in every country, ranging from 0.64 to 0.91 per 1,000 people per year among semaglutide users and from 0.80 to 0.98 per 1,000 people per year among active comparator users. In other words, the raw cancer rates in the two treatment groups overlapped almost completely before any statistical modelling was even applied.

The formal statistical analysis confirmed what the raw numbers suggested. The primary outcome, incident pancreatic cancer, was analysed using country-specific hazard ratios and incidence rate differences, which were then pooled across the three countries using a fixed-effects meta-analysis. In Denmark, the hazard ratio for pancreatic cancer with semaglutide versus active comparators was 1.00, with a 95 percent confidence interval of 0.66 to 1.51 — a value of exactly one indicating no difference in risk whatsoever. In Sweden, the hazard ratio was 0.82 (95 percent confidence interval 0.56 to 1.22), and in Norway it was 0.91 (95 percent confidence interval 0.55 to 1.51). The pooled estimate across all three countries showed no increased risk, with a hazard ratio of 0.91 (95 percent confidence interval 0.71 to 1.16) and an incidence rate difference — the absolute change in risk — of minus 0.10 per 1,000 person-years, with a confidence interval spanning from minus 0.31 to plus 0.10. Every confidence interval comfortably included the null value of no effect.

Just as important as the headline result was the pattern of secondary and sensitivity findings. If a drug truly caused cancer, one would generally expect to see a dose-response relationship, with risk climbing as cumulative exposure increases, and a duration effect, with risk rising the longer patients remain on therapy. Neither pattern appeared. The pooled hazard ratio for high-dose semaglutide use was 0.74 (95 percent confidence interval 0.40 to 1.40), and the pooled hazard ratio for long duration of use was 0.79 (95 percent confidence interval 0.37 to 1.69) — both point estimates below one and both statistically compatible with no effect. The researchers also performed extensive sensitivity and supplementary analyses to probe the robustness of their conclusions, and these consistently showed no increased risk, regardless of how the analyses were configured. This internal consistency across countries, doses, durations and analytical approaches substantially strengthens the credibility of the null finding.

Professor Pottegård summarised the implications of the work in his presentation, stating that the comprehensive study, using nationwide registry data from three countries, found no increased risk of pancreatic cancer with the use of semaglutide compared with other glucose-lowering drugs used at a similar stage in the treatment of type 2 diabetes. He emphasised that the conclusion was supported by the absence of any cumulative dose-response or dose-duration effect, by consistent findings in the individual countries, and by a wide range of supplementary and sensitivity analyses. He further noted that the results align with the current external evidence base, thereby strengthening the overall argument that semaglutide does not increase the risk of pancreatic cancer. For the millions of patients now taking the drug worldwide, and for the clinicians who prescribe it, this convergence of registry-scale real-world data with prior evidence offers a measure of clarity that early case reports and rodent studies could never provide.

Some caveats are worth bearing in mind. The research was presented as an oral presentation, abstract LBA 35, at the EASD meeting in Milan, and no full article is available at this stage; the work has not yet been submitted to a medical journal for peer-reviewed publication, so the findings should be regarded as preliminary until they appear in the scientific literature. The mean follow-up of roughly one and a half to two years, while appropriate for the question of newly emerging cancers, cannot rule out effects that might manifest only after many more years of exposure, and continued pharmacovigilance will remain essential as the treated population ages. The study was funded by Novo, the manufacturer of semaglutide, a fact that regulators explicitly anticipated by mandating the study design and comparator choices as a condition of approval. Nevertheless, the sheer scale of the Nordic registry infrastructure, the active-comparator design, the balanced cohorts and the consistency of every analysis performed make this one of the most definitive real-world assessments to date of a safety concern that has shadowed GLP-1 receptor agonists since their introduction — and that concern, at least for pancreatic cancer, now appears substantially diminished.

Subject of Research: Post-authorisation safety study of semaglutide and pancreatic cancer risk in type 2 diabetes patients

Article Title: Regulator-mandated study finds no link between semaglutide use and pancreatic cancer using nationwide registry data from Denmark, Sweden and Norway

Article References: Regulator-mandated study finds no link between semaglutide use and pancreatic cancer using nationwide registry data from Denmark, Sweden and Norway. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: semaglutide, pancreatic cancer, GLP-1 receptor agonists, type 2 diabetes, drug safety, registry study, Denmark, Sweden, Norway, EASD, pharmacovigilance, Novo