Frequently Asked Questions:
- What did this study compare?
The retrospective TriNetX analysis compared GLP-1 RAs (semaglutide, tirzepatide, liraglutide, or dulaglutide) with metformin in 256,492 propensity-matched adults with prediabetes and obesity.
- Were GLP-1 RAs linked to less progression to diabetes than metformin?
Yes. Progression to diabetes occurred in 25.1% of GLP-1 RA recipients vs 37.0% of metformin users (HR, 0.69; 95% CI, 0.68–0.70), and normoglycemia reversion was more frequent (37.3% vs 29.3%).
- Do these results prove GLP-1 RAs are superior to metformin in prediabetes?
No. The study is observational and has not been peer reviewed, and investigators said prospective randomized trials are needed to confirm causality.
Study design and glycemic outcomes with GLP-1 RAs vs metformin
The investigators used the TriNetX electronic health record network, spanning 104 institutions, to identify adults with prediabetes and obesity who initiated semaglutide, tirzepatide, liraglutide, dulaglutide, or metformin from January 2010 through December 2023. Patients with prior diabetes, stroke, heart failure, atrial fibrillation, or ischemic heart disease were excluded. Cohorts underwent 1:1 propensity score matching for demographics, comorbidities, medications, and laboratory values.1
Primary outcomes were reversion to normoglycemia and progression to diabetes, and secondary outcomes included all-cause mortality and major cardiovascular events, analyzed with Cox proportional hazards models. After matching, 256,492 patients were included, with a mean age of 52.4 years and 76.1% female. Mean body mass index was 39.5 kg/m² in the GLP-1 RA cohort and 38.5 kg/m² in the metformin cohort, and patients were followed for an average of roughly 3 years.1,2
Cardiovascular outcomes and limitations of the GLP-1 RA analysis
Compared with metformin, GLP-1 RA therapy was associated with lower risk of acute heart failure (HR, 0.74), ischemic stroke (HR, 0.80), acute myocardial infarction (HR, 0.84), cardiac arrest (HR, 0.58), and atrial fibrillation (HR, 0.92; all P < .05). Ventricular tachycardia risk did not differ between groups (HR, 1.00; P = .978). The abstract did not report 95% CIs for these individual cardiovascular outcomes.1
According to the AHA, the study is observational and cannot establish causation, outcomes were identified through diagnosis codes, and prescription records do not reflect adherence or dose. GLP-1 RAs were also analyzed as a single group, without comparisons between individual agents.2 The investigators said prospective randomized trials are needed to confirm causality, and the abstract has not been peer reviewed.1,2
“The main effect these findings could have is on the shared decision-making discussion between doctors and patients exploring various treatment options, and the pros and cons for each,” said Serra. “Metformin is still a first-line treatment choice. It is safe, inexpensive and has a long track record. This is less about replacing one medication with another and more about discussing additional options with specific patients.”2
Metformin, alongside lifestyle modification, remains the standard first-line pharmacologic option for prediabetes, and most prior GLP-1 RA research has focused on patients with established type 2 diabetes, according to the AHA.2 The investigators said the findings support an expanding role for GLP-1 RAs in diabetes prevention and cardiovascular risk reduction.1
References
de Sousa Barzon Serra V, Ibrahim R, Zanco J, et al. Glucagon-like peptide-1 receptor agonists versus metformin for glycemic control and cardiovascular risk reduction in adults with prediabetes and obesity. Presented at: American Heart Association Scientific Sessions 2026; November 6-9, 2026; Chicago, IL. Abstract MDP1524.
