All three GLP-1 receptor agonists showed similar kidney and cardiovascular outcomes in veterans with type 2 diabetes, though liraglutide was associated with lower all-cause mortality compared to dulaglutide.
31% lower mortality riskVeterans initiating liraglutide had a 31% lower hazard of all-cause death compared to those starting dulaglutide (HR 0.69, 95% CI 0.58–0.83).
What the researchers found
In this large comparative effectiveness study of 21,790 veterans with type 2 diabetes, liraglutide, semaglutide, and dulaglutide showed no significant differences in kidney failure or cardiovascular outcomes. Kidney failure hazard ratios were similar across all comparisons (e.g., liraglutide vs semaglutide HR 0.93, 95% CI 0.60–1.44), and the composite cardiovascular-kidney-metabolic outcome and major adverse cardiovascular events were also comparable.
The most notable finding involved all-cause mortality. Liraglutide was associated with a 31% lower hazard of death compared to dulaglutide in intent-to-treat analysis (HR 0.69, 95% CI 0.58–0.83) and a 50% lower hazard in per-protocol analysis (HR 0.50, 95% CI 0.31–0.82). Liraglutide also showed a trend toward lower mortality versus semaglutide (HR 0.83, 95% CI 0.69–0.99), though this lost statistical significance in per-protocol models. For gastrointestinal side effects, dulaglutide was associated with lower risk of gallstones (HR 0.72) and acute cholecystitis (HR 0.62) compared to semaglutide.
Why it matters
GLP-1 receptor agonists have become a cornerstone of type 2 diabetes treatment, but most evidence comes from placebo-controlled trials rather than head-to-head comparisons. Clinicians and patients need to know whether one GLP-1RA offers meaningful advantages over another. This large real-world study provides reassurance that the three most commonly used GLP-1RAs are broadly comparable for heart and kidney protection, while flagging potential mortality differences that warrant confirmation in randomized trials.
How the study worked
This was a comparative effectiveness study using a target trial-emulation design, linking national data from the VA health system, Medicare, and the US Renal Data System. Researchers identified veterans with type 2 diabetes who were new users of liraglutide, semaglutide, or dulaglutide between 2018 and 2021, all of whom were already on metformin and did not have end-stage kidney disease. They used weighted Cox regression models to compare outcomes including kidney failure, cardiovascular events, death, and gastrointestinal adverse events through March 2023.
What this study cannot tell us
As an observational study, even with careful statistical adjustments, residual confounding cannot be ruled out — differences in why doctors prescribed one drug over another could influence outcomes. The study population was predominantly male veterans (91%), which limits generalizability to women and non-veteran populations. The mortality advantage for liraglutide lost significance in some per-protocol analyses, suggesting the finding may not be robust. Additionally, the study period (2018–2021) included the early COVID-19 pandemic, which may have affected outcomes in ways that are difficult to account for.
How to read the evidence
This is a large, well-designed observational comparative effectiveness study using a target trial-emulation framework and linked national databases. While not a randomized controlled trial, the rigorous methodology, large sample size, and use of active comparators with new-user designs strengthen the evidence. However, observational design inherently limits causal conclusions.
When this study was published
Published in 2025, this is a very recent study using data through 2023. Its findings reflect current prescribing patterns and contemporary formulations of these three GLP-1RAs.
The bigger picture
As GLP-1 receptor agonists see rapidly expanding use for diabetes, obesity, and cardiorenal protection, understanding which agent to choose matters for individualized care. This study adds to a growing body of real-world evidence suggesting the class effect is largely consistent across agents for cardiovascular and kidney outcomes. The mortality signal favoring liraglutide is intriguing but could reflect residual confounding in observational data — the authors themselves call for head-to-head randomized trials to settle the question.
Questions still open
- Would a head-to-head randomized clinical trial confirm the mortality advantage seen with liraglutide over dulaglutide?
- Do these findings hold in more diverse populations, including women and non-veteran groups?
- What mechanisms might explain potential mortality differences between GLP-1RAs that share the same receptor target?
Common questions
Are liraglutide, semaglutide, and dulaglutide equally effective for heart and kidney protection?
Should I switch from dulaglutide to liraglutide based on the mortality findings?
Read the original research
Liraglutide vs Semaglutide vs Dulaglutide in Veterans With Type 2 Diabetes.
JAMA network open, 8(10), e2537297
Citation
Derington, Catherine G; Sarwal, Amara; Wei, Guo; Hartsell, Sydney E; Throolin, Michael; Singh, Ravinder; Nevers, McKenna R; Zhang, Chong; Katkam, Niharika; Takyi, Augustine; Chakravartula, Akhil R; Babu, Poorvika; Deshmukh, Vikrant G; Boucher, Robert E; Drakos, Stavros G; Greene, Tom; Shen, Jincheng; Beddhu, Srinivasan. (2025). Liraglutide vs Semaglutide vs Dulaglutide in Veterans With Type 2 Diabetes.. JAMA network open, 8(10), e2537297. https://doi.org/10.1001/jamanetworkopen.2025.37297