A new-user cohort study compared dulaglutide, subcutaneous semaglutide, and tirzepatide in patients with type 2 diabetes using Optum Clinformatics claims and a target-trial emulation. Three cohorts were constructed: semaglutide vs dulaglutide (2019 to 2024), tirzepatide vs dulaglutide (2022 to 2024), and tirzepatide vs semaglutide (2022 to 2024).
Patients aged at least 18 years with type 2 diabetes, 365-day continuous enrollment, and no baseline glucagon-like peptide-1 receptor agonist (GLP-1 RA) use were eligible; those with prior study outcomes and standard diabetes-related exclusions were removed, plus additional exclusions for prior bariatric/abdominal surgery, metoclopramide or erythromycin use, or nursing-home admission during the baseline year.
The primary outcome was a composite of severe gastrointestinal events (acute pancreatitis, biliary disease, bowel obstruction, gastroparesis, severe constipation) requiring inpatient or emergency department care; secondary outcomes grouped pancreaticobiliary and motility-related components. Follow-up began the day after initiation and ended at outcome, disenrollment, discontinuation (supply plus 60 days), switch, death, or study end (August 30, 2024).
After 1:1 propensity-score matching, there were 65,238; 20,893; and 46,620 pairs across the three cohorts. Primary-outcome rates were similar across comparisons (hazard ratios ≈1.0; no statistically significant differences). Secondary components were also similar. Subgroup and extensive sensitivity analyses were consistent with the primary analysis.
Using sodium-glucose cotransporter-2 (SGLT-2) inhibitors as an external comparator, risk for severe gastrointestinal events was higher with GLP-1 RAs and tirzepatide, driven by motility-related outcomes; pancreaticobiliary risks did not differ.
Limitations include claims-based outcome misclassification, residual confounding (e.g., body mass index/hemoglobin A1c), inability to assess dose, limited power for some subgroups, and short median follow-up (~5-6 months); findings may not generalize beyond patients with type 2 diabetes.
Source: Annals of Internal Medicine
Funding/Disclosures: Supported by the National Institute of Diabetes and Digestive and Kidney Diseases (R01DK138036). Disclosures available with the article.