A large cohort study found three popular GLP-1-based diabetes drugs — semaglutide, dulaglutide, and tirzepatide — carry similar risks for serious adverse GI events, with a rate of about 12 per 1,000 person-years.The risk of those events was lower with the SGLT-2 inhibitor class of diabetes medications than in the group of GLP-1 drugs.The authors say the findings should give clinicians confidence that safety differences are not a major factor when choosing among these three GLP-1-related drugs for patients with type 2 diabetes.
Three popular GLP-1 receptor agonists carried similar risks for serious gastrointestinal (GI) events among adults with type 2 diabetes, although all higher than seen with sodium-glucose cotransporter 2 (SGLT-2) inhibitors, according to a large, real-world analysis.
In the new-user, active-comparator cohort study of more than 200,000 medical records, rates of GI adverse events requiring medical attention came in at around 10 to 12 per 1,000 person-years across all three type 2 diabetes medications — subcutaneous semaglutide (Ozempic), dulaglutide (Trulicity), and tirzepatide (Mounjaro), reported Elisabetta Patorno, MD, DrPH, of Brigham and Women’s Hospital and Harvard Medical School in Boston, and colleagues in Annals of Internal Medicine.
Semaglutide and dulaglutide are GLP-1 receptor agonists, while tirzepatide is a dual GLP-1/glucose-dependent insulinotropic polypeptide (GIP) agonist. Head-to-head comparisons showed each drug carrying roughly the same risk of serious pancreatitis, gastroparesis, or bowel obstruction that required emergency department visits or hospitalization:
Semaglutide versus dulaglutide: HR 0.96 (95% CI 0.87-1.06)Tirzepatide versus dulaglutide: HR 0.96 (95% CI 0.77-1.20)Tirzepatide versus semaglutide: HR 1.07 (95% CI 0.90-1.26)
“Clinicians can consider these agents comparable in this regard, while keeping in mind that GLP-1–based therapies as a class appear to carry a higher gastrointestinal risk compared with other diabetes drug classes,” Patorno and co-authors Wajd Alkabbani, PhD, and Salvatore Crisafulli, PhD, both of Brigham and Women’s Hospital in Boston, told MedPage Today in a joint email. Alkabbani is also affiliated with the University of Waterloo in Ontario, and Crisafulli with the University of Verona in Italy.
Compared with SGLT-2 inhibitors, however, each GLP-1 drug carried a significantly higher risk for adverse GI events: semaglutide (HR 1.22, 95% CI 1.09-1.37), dulaglutide (HR 1.36, 95% CI 1.21-1.53), and tirzepatide (HR 1.53, 95% CI 1.21-1.93). The excess risk versus SGLT-2 inhibitors was driven mainly by motility-related problems, such as gastroparesis and bowel obstruction.
“Although confirming this increase in risk when compared with [SGLT-2 inhibitors]our study did not find evidence of a difference in the risk for severe gastrointestinal [adverse events] when dulaglutide, subcutaneous semaglutide, and tirzepatide were compared with each other,” the investigators wrote.
The study did not directly evaluate dose-response relationships, but the authors identified these as an area for future investigation. While higher doses or rapid dose titration may increase mild GI intolerance early in treatment, it remains unclear whether those factors translate into more severe events, they added. The analysis also excluded milder symptoms like nausea and vomiting, which are difficult to capture in claims data.
“Our analysis focused on serious gastrointestinal outcomes requiring medical attention [because] these events are clinically relevant, and can be reliably captured using administrative claims,” the researchers said in the email.
The study used Optum’s Clinformatics Data Mart database to identify adults with type 2 diabetes who initiated one of the three GLP-1-based therapies between January 2019 and August 2024. After 1:1 propensity-score matching, the cohorts included 65,238 semaglutide-dulaglutide pairs, 20,893 tirzepatide-dulaglutide pairs, and 46,620 tirzepatide-semaglutide pairs. Mean age of these groups ranged from 60.2 to 62.8 years, and 49.3-53.2% were women. Median follow-up was 163-183 days.
The investigators cautioned that diagnostic-code misclassification and residual confounding from incomplete data on BMI and hemoglobin A1c were possible. Because median follow-up was only about 5 to 6 months, they said longer-term studies are needed to determine whether risks persist with continued use.
