Semaglutide vs. liraglutide and incidence of diabetes and cardiovascular disease: A target trial emulation using real-world data

In a real-world analysis of 57,456 matched patients without prior diabetes or cardiovascular disease, semaglutide was associated with a 12% lower risk of developing diabetes over one year compared with liraglutide. The lower risk appeared only after the first six months, and no clear difference in cardiovascular events was observed.

Journal article — Active comparator, new user target trial emulation. Population: MarketScan enrolees free of diabetes and cardiovascular disease prescribed semaglutide or liraglutide during 2021-2023. Sample size: 57 456 GLP-1 RA users. Follow-up: 1-year mean follow-up. Interventions: Semaglutide.

After 1:1 matching, 57,456 GLP-1 RA users were analysed. Over a mean follow-up of 1 year, there were 1104 diabetes events and 57 CVD events. After propensity-score adjustment, semaglutide users had a 12% lower risk of diabetes (HR 0.88; 95% CI 0.78-0.99), but the proportional hazards assumption was violated. In the first 6 months, the HR was 0.99 (95% CI 0.82-1.18); after 6 months, it was 0.80 (95% CI 0.68-0.94). The hazard ratio for CVD events was 1.25 (95% CI 0.74-2.11), indicating no detected difference.

Researchers studying semaglutide would care because this paper compares it head-to-head with liraglutide for diabetes prevention in routine care, a question not answered by placebo-controlled trials. It does not establish causality, and the cardiovascular comparison remains inconclusive. It provides no information on dosing, weight outcomes, or longer-term safety.

Key findings

Semaglutide users had a 12% lower risk of incident diabetes versus liraglutide users (HR 0.88; 95% CI 0.78-0.99).
In the first 6 months, semaglutide was not associated with diabetes risk (HR 0.99; 95% CI 0.82-1.18).
After 6 months, semaglutide was associated with a 20% lower diabetes risk (HR 0.80; 95% CI 0.68-0.94).
No difference in cardiovascular events was detected (HR 1.25; 95% CI 0.74-2.11).
The matched cohort included 57,456 GLP-1 RA users (mean age 45; 83% female) with 1104 diabetes events and 57 CVD events over a mean follow-up of 1 year.

Limitations

Observational claims-based design cannot exclude residual confounding despite propensity score adjustment.
Mean follow-up of 1 year is short for diabetes and cardiovascular outcomes.
Only 57 CVD events were observed, yielding an imprecise estimate.
Proportional hazards assumption was violated for the diabetes endpoint, requiring stratification by follow-up time.

The record

Journal: British journal of clinical pharmacology
Published: 2026-07-16
PMID: 42464447 — DOI: 10.1002/bcp.70711
Funding: NHLBI NIH HHS; American Heart Association; NIH Clinical and Translational Science; NHLBI NIH HHS
Author countries: USA

Peptide profiles: Semaglutide.

All indexed evidence: Semaglutide trials & papers.