Stopping GLP-1 Drugs Linked to Double the Heart Attack, Stroke or Death Risk in Type 2 Diabetes

On 9 November 2026, The Lancet Regional Health Europe published an analysis of records from 301 diabetes clinics across Italy. It followed what happened to people with type 2 diabetes after they started, and then in many cases stopped, a GLP-1 receptor agonist. The headline finding: those who came off the drug had roughly twice the risk of a heart attack, a stroke or death as those who stayed on it.

The data covered the years 2004 to 2024 and included 80,302 patients. All had begun a GLP-1 receptor agonist and had no history of a major cardiovascular event when they did. By their last recorded visit, 39.5% (31,687 people) were still being treated. The remaining 60.5% (48,615 people) had stopped, which the researchers defined as having no GLP-1 prescription at two or more consecutive visits.

To make the groups comparable, the team matched 28,713 patients in each. They matched on age, BMI, HbA1c (a measure of blood glucose) and how long each person had lived with diabetes. In the matched groups, the rate of non-fatal heart attack, non-fatal stroke or death from any cause was 3.94 per 100 years among those who stopped. Among those who continued it was 1.96. Death from any cause, examined separately as a secondary endpoint, showed a wider gap: 2.23 per 100 person-years in the discontinuation group against 0.48 in the continuation group.

The association appeared stronger in people with a BMI of 30 or above, with a hazard ratio of 2.27, than in those below 30, where it was 1.61. The authors also noted that earlier studies have found 20% to 50% of people stop GLP-1 treatment within the first year, mostly because of gastrointestinal side effects, cost or inconvenience.

The authors were careful about what the numbers can carry. The study is observational and cannot show cause and effect, and they did not know why individual patients stopped. A gap this large, they wrote, is unlikely to be explained solely by withdrawal of the drug. Illness itself may have prompted some people to stop, so discontinuation may signal a higher underlying risk rather than directly causing one.

They also checked whether ordinary risk factors could account for the difference. Comparing weight, blood pressure and cholesterol from the start of treatment to 12 and 24 months, they looked at people who stopped and went on to have an event, and people who stopped and did not. The differences were small, suggesting the higher risk may not be fully explained by a worsening of traditional risk factors alone.

There are limits on who the results describe. They may not apply to people taking GLP-1 drugs for obesity without type 2 diabetes. Nor do they cover tirzepatide, which was not yet available during the study period.

Faye Riley, research communications lead at the charity Diabetes UK, said the findings should be interpreted with caution. Because the study is observational, she said, it cannot tell whether stopping raised the risk directly or whether those who stopped were already experiencing health problems. She added that the findings reinforce the importance of supporting people to stay on treatments that work for them, and that anyone with type 2 diabetes considering stopping should talk to their healthcare team first.

What to watch for is whether later research can separate the effect of stopping from the reasons for stopping, since that distinction is the one this study could not make.