High-dose semaglutide shows a colorectal tumour signal in pooled trial data

GLP-1 receptor agonists and SGLT2 inhibitors have become two of the most widely used drug classes for metabolic disorders, so even small safety signals matter at population scale. Recent evidence has raised concerns about possible cancer-promoting effects, particularly in the gastrointestinal tract. Colorectal tumours are the third most diagnosed cancer globally, and the people who typically take these drugs, those with metabolic disorders and obesity, already carry an elevated baseline risk. Existing evidence on whether the drugs add to that risk has been inconsistent, which is the gap this analysis set out to address.

The researchers ran a confirmatory network meta-analysis following Cochrane methodological recommendations, pooling 68 randomised controlled trials with a combined 207 200 participants. The primary outcome was the incidence of colorectal tumours reported as an adverse effect across trials of GLP-1 receptor agonists and SGLT2 inhibitors. Safety profiles were assessed through dropout rates as a secondary outcome. The team also stratified the analysis by dose and regimen, and looked separately at trials conducted in participants with obesity.

Only one drug stood out. Semaglutide was associated with an increased incidence of colorectal tumours compared with controls, while no other GLP-1 receptor agonist and no SGLT2 inhibitor showed a significant association. The dose-stratified analysis narrowed the signal further: high-dose injectable semaglutide at 2.4 mg per week was the only regimen linked to increased incidence. When the analysis was restricted to trials in participants with obesity, the association with semaglutide persisted, a pattern the authors describe as dose-dependent and specific to a single drug rather than a class-wide effect.

The authors present this as the first comprehensive network meta-analysis of colorectal tumour incidence across individual GLP-1 receptor agonists and SGLT2 inhibitors, in a literature that had previously been inconsistent. They are careful to frame the result as a risk signal requiring further validation, not a demonstrated harm. Interpretation is complicated by the fact that the populations in these trials already face elevated colorectal tumour risk, especially those with obesity. Long-term follow-up studies are needed to characterise the mechanisms and clinical implications, and the null finding for every other drug in both classes is itself part of the picture.

For anyone using semaglutide, particularly the high-dose weekly injection, this is information worth having rather than a reason for alarm. The signal comes from pooled trial data, has not been validated, and sits against a backdrop where the people studied already had higher colorectal risk. It is equally worth knowing what did not show a link: every other GLP-1 receptor agonist and every SGLT2 inhibitor examined came out clean here. Questions about individual risk belong in a conversation with a prescriber, not in a decision made from a single analysis.