Study finds similar gastrointestinal safety for semaglutide, tirzepatide and dulaglutide

A large population-based study found that three widely used diabetes medicines — dulaglutide, subcutaneous semaglutide (sold as Ozempic and Wegovy) and tirzepatide (sold as Mounjaro and Zepbound) — carried similar risks of serious gastrointestinal problems in adults with type 2 diabetes ^[1].

The study, published in Annals of Internal Medicine, followed adults with type 2 diabetes who started one of the three drugs between January 1, 2019 and August 30, 2024 ^[1]. It was a new-user, active-comparator cohort study, meaning researchers compared people starting one drug against people starting another, rather than assigning treatments at random ^[1]. Patients were matched one-to-one using propensity scores, so each comparison weighed two groups that looked alike on measured characteristics ^[1].

The main outcome was a combined measure of five serious events: acute pancreatitis, biliary disease (gallbladder and bile duct problems), bowel obstruction, gastroparesis (slow stomach emptying) and severe constipation ^[1].

The results, by comparison:

• Semaglutide versus dulaglutide: 65,238 matched pairs, hazard ratio 0.96 (95% CI, 0.87 to 1.06) ^[1].
• Tirzepatide versus dulaglutide: 20,893 matched pairs, hazard ratio 0.96 (95% CI, 0.77 to 1.20) ^[1].
• Tirzepatide versus semaglutide: 46,620 matched pairs, hazard ratio 1.07 (95% CI, 0.90 to 1.26) ^[1].

Because the matching was one-to-one, the tirzepatide-versus-semaglutide comparison covered 46,620 people on each drug ^[1]. A hazard ratio near 1.0 means the two groups had roughly the same rate of events over the study period, and every confidence interval included 1.0, so the results are consistent with no difference between the drugs ^[1].

The authors concluded that dulaglutide, semaglutide and tirzepatide "have similar gastrointestinal safety profiles in adults with T2D" and wrote that the study gives clinicians evidence to weigh benefits and risks ^[1]. The work was funded by the National Institute of Diabetes and Digestive and Kidney Diseases ^[1].

Two limits are worth noting. The authors list possible residual confounding by glycemic control and body mass index as a limitation, meaning they could not fully rule out that differences in blood sugar control or weight between the groups, rather than the drugs themselves, influenced the results ^[1]. In addition, the study enrolled only adults with type 2 diabetes, so it does not show whether the same pattern holds for people who take these medicines for weight loss without diabetes ^[1].

The abstract reports the combined outcome and the hazard ratios for each pairing. It says the individual conditions were secondary outcomes of interest but does not give their separate numbers, so rates for any single problem, such as gastroparesis alone, have not been disclosed in the abstract ^[1]. Common, milder side effects such as nausea were not part of the measured outcome ^[1].

Why it matters for patients

For adults with type 2 diabetes who are choosing among these drugs or already taking one, this study suggests that, on average, the risk of these five severe gastrointestinal events was not meaningfully different across dulaglutide, semaglutide and tirzepatide ^[1]. That is a comparison, not a clean bill of health: the study measured how often the events occurred in large groups, and an event rate near the same in two groups does not mean the events do not happen ^[1].

The findings are averages across tens of thousands of people, not a prediction for any one person, and they come from an observational design that can show associations but not cause and effect ^[1]. The authors' own caveat about residual confounding means some uncertainty remains about whether the drugs themselves account for the full picture ^[1].

Results also should not be extended beyond the group studied. Whether the same safety pattern applies to people using these medicines for weight loss, or to people with other health profiles, has not been reported ^[1]. Evidence on those questions is not yet known from this study.