Glucose-Lowering Medication Classes and Cardiovascular Outcomes in Patients With Type 2 Diabetes
Authors' conclusion
In this study, MACE risk varied significantly by medication class, with most protection achieved with sustained treatment with GLP-1RAs followed by SGLT2is, sulfonylureas, and DPP4is. The magnitude of benefit of GLP-1RAs over SGLT2is depended on baseline age, ASCVD, HF, and kidney impairment. These results, along with consideration of cost, availability, and collateral clinical benefits, may inform treatment decisions for adults with T2D.
Study context
- Population
- US adults with T2D in 6 large US health care delivery systems initiating 1 of 4 medication classes, 2014–2021; 4-arm cohort mean (SD) age 57.2 (12.9) years, 54.3% male241,981 in 4-arm emulation
- Participants
- 296,676 adults
- Comparison
- Sustained exposure to sulfonylureas vs DPP4is vs SGLT2is vs GLP-1RAs
- Outcome
- MACE: nonfatal MI, nonfatal stroke or cardiovascular death
- Follow-up
- 2.5 years
Methodological detailstarget trial emulation (4-arm), targeted learning with machine learning, 6 US health care delivery systems, prespecified subgroups
- Target trial emulation (4-arm)
- Targeted learning with machine learning
- 6 US health care delivery systems
- Prespecified subgroups
Main finding
As reported in the abstract
MACE cumulative risk
Cumulative risk difference (95% CI)
DPP4isvs Sulfonylureas
1.9%(1.1%–2.7%)
SGLT2isvs GLP-1RAs
1.5%(1.1%–1.9%)
Further findings
MACE risk differences by atherosclerotic cardiovascular disease (ASCVD) status · 2.5 years
- Risk differences similar in direction with vs without ASCVD but typically much smaller without ASCVD
Benefit of GLP-1RAs over SGLT2is by subgroup · 2.5 years
- Most pronounced with baseline ASCVD or heart failure (HF), age ≥65 years, or low to moderate kidney impairment; not found in patients <50 years
Original publication
Glucose-Lowering Medication Classes and Cardiovascular Outcomes in Patients With Type 2 Diabetes. JAMA Netw Open. Published 1 October 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →