Investigators reported that, in this large electronic health record analysis, people prescribed GLP-1–based medications had higher likelihoods of reversing elevated blood sugar and lower risks of progressing to Type 2 diabetes compared with those prescribed metformin. The cohort showed a reported 1.5-times greater chance of returning blood sugar to normal, a reported 31% lower risk of developing Type 2 diabetes and a reported 41% lower risk of death from any cause over the study period. These effects were observed across the group of GLP-1 agents analyzed together; the abstract did not provide agent-specific breakdowns.
The study team extracted records for 256,492 adults diagnosed with prediabetes and obesity. The mean age in the cohort was 52 years, and about three quarters were women. Data came from the TriNetX global multi-center electronic health record network, representing 104 institutions and covering 2010 through 2023. Patients were followed for an average of roughly three years. The source did not report details on medication dosing or duration beyond prescription records captured in the EHR.
Researchers paired records of patients in the GLP-1 group and the metformin group on demographic and clinical characteristics. Matching variables included age and several comorbid conditions such as high blood pressure, sleep apnea and nicotine dependence, along with concurrent medications. The matching approach aimed to reduce bias so that outcome differences would be more likely associated with the medication exposure rather than baseline differences between groups. The analysis relied on diagnosis codes and prescription entries; the team did not have direct measures of medication adherence or precise dosing, according to the abstract.
Data indicated that adults taking GLP-1 medications were about 1.5 times more likely to revert elevated glucose to normal levels compared with those taking metformin. The cohort taking GLP-1s also had a lower risk of progressing from prediabetes to Type 2 diabetes — a 31% relative reduction as reported by the investigators. The abstract noted that most prior GLP-1 research focused on patients with established Type 2 diabetes, whereas this analysis targeted the earlier prediabetes stage to capture potential preventive effects.
The investigators reported cardiovascular outcome differences favoring GLP-1 exposure. Compared with metformin, GLP-1 use was associated with 15% fewer heart attacks, 20% fewer strokes and 25% fewer heart-failure events. A smaller reduction (8%) was reported for atrial fibrillation. All-cause mortality was lower in the GLP-1 group, with a reported 41% relative risk reduction over the follow-up period.
Commentators in the release noted that weight loss likely contributed to many of the observed benefits. American Heart Association–quoted experts said losing weight can reduce the risk of progression from prediabetes to Type 2 diabetes, which in turn lowers cardiovascular risk. The study author argued that these results may inform shared decision-making between clinicians and patients considering treatment options. The abstract reiterated that metformin remains a first-line medication: it is affordable, widely used and has a long safety track record. The source also noted potential barriers to GLP-1 use, such as cost and access, but did not provide economic or utilization data beyond citing a survey that roughly 1 in 10 U.S. adults may be taking GLP-1 medications for blood-sugar control or weight loss.
The authors and the American Heart Association emphasized that this analysis is observational and therefore cannot establish causality. Limitations documented in the abstract included reliance on diagnostic coding (which can be incomplete), use of prescription records without direct measures of adherence or dosing, and grouping of different GLP-1 agents together without assessing agent-specific effects. The investigators used matching to address confounding, but residual confounding remains possible in any nonrandomized design. Finally, the findings were presented as a meeting abstract at the American Heart Association Scientific Sessions 2026 and are considered preliminary until reported in a full, peer-reviewed manuscript, as stated in the source.