Semaglutide vs. liraglutide: Head-to-head study reveals a diabetes risk difference

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Jul 20 2026

A large head-to-head analysis reveals when outcomes began to diverge and why questions about the drugs’ cardiovascular effects remain unresolved.

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

In a recent study published in the British Journal of Clinical Pharmacology, a group of researchers compared semaglutide and liraglutide with respect to the incidence of diabetes and cardiovascular disease among adults without a history of diabetes or cardiovascular disease who initiated formulations indicated for chronic weight management.

Background

Obesity affects about 40% of the adult population in the United States, increasing their risk of diabetes and heart-related illnesses. Prescriptions for glucagon-like peptide-1 receptor agonists (GLP-1 RAs), such as liraglutide and semaglutide, are increasing, making it important to understand how well these medications compare in relation to long-term complications. Both drugs have reduced diabetes risk, improved blood glucose control, promoted weight loss, and decreased cardiovascular events compared with placebo in selected populations, although the populations, doses, and indications differed. However, head-to-head data comparing the drugs in routine practice remain limited, particularly among people without diabetes or established cardiovascular disease.

About the study

The study was conducted using the Merative™ MarketScan® Commercial and Medicare Research Databases covering 2018 through 2023. Adults aged 18–99 years who filled a first prescription for weekly semaglutide or daily liraglutide formulations indicated for chronic weight management between 2021 and 2023 were included in the study. Adults with recorded evidence of prior type 1 or type 2 diabetes, myocardial infarction, stroke, or heart failure were excluded from the study. The participants were required to have prior insurance enrollment for a period of at least six months before treatment initiation.

An active-comparator, new-user design was used to emulate a target trial. Each liraglutide initiator was matched 1:1 to a semaglutide initiator using the criteria of age and sex along with enrollment date and treatment start date. In the primary analysis, participants remained classified according to the drug they initially received, regardless of discontinuation or switching. The main outcomes were the incidence of diabetes and cardiovascular disease, where the latter included hospitalized myocardial infarction, stroke, or heart failure. The secondary composite expanded the cardiovascular outcome to include unstable angina and coronary revascularization. Cox proportional hazards regression was used to assess hazard ratios (HRs) while accounting for age and a propensity score reflecting initial comorbidities and drug use. Additional sensitivity analyses were conducted to check if the outcomes remained stable under other enrollment criteria, matching strategies, statistical weighting approaches, and a per-protocol approach that censored follow-up at treatment discontinuation or switching.

Study results

From the MarketScan® databases, the researchers identified 33,920,450 unique enrolled people from 2021 to 2023, of whom 142,642 adults qualified to be matched, including 103,533 semaglutide-treated individuals and 39,109 liraglutide-treated individuals. After this 1:1 matching, the final study population included 57,456 participants, with 28,728 individuals in each treatment group. The average age was 45 years, 83% of participants were female, and around 40% were diagnosed with hypertension. The baseline characteristics of the two groups were similar, with standardized mean differences of less than 0.1.

During a median follow-up of approximately one year, 1,104 participants had an incident diabetes event and 57 experienced cardiovascular disease events. Among the cardiovascular disease events, 30 participants experienced myocardial infarction, 21 cases of stroke, and six cases of heart failure. The total number of composite cardiovascular outcomes reached 154. Treatment discontinuation was common: 53% of patients on semaglutide discontinued treatment, including 6% who switched to liraglutide, and 79% of participants taking liraglutide discontinued treatment, including 33% who switched to semaglutide.

After adjusting for age and baseline characteristics, semaglutide was associated with a 12% lower risk of incident diabetes compared with liraglutide (HR 0.88; 95% confidence interval (CI): 0.78–0.99). Researchers then analyzed the data based on duration of follow-up because the proportional hazards assumption was violated. There was no significant difference in terms of risk for diabetes during the first six months after treatment initiation (HR 0.99; 95% CI: 0.82–1.19). However, during follow-up beyond six months, semaglutide use was associated with a significantly lower risk than liraglutide use (HR 0.80; 95% CI: 0.68–0.94). This indicates that the association favoring semaglutide emerged later in follow-up, although it does not establish an effect of continuous treatment duration.

For cardiovascular outcomes, the relatively small number of events limited statistical precision. No statistically significant difference could be established between semaglutide and liraglutide in the risk of cardiovascular disease (HR 1.25; 95% CI 0.74–2.11) or the composite cardiovascular outcome (HR 0.89; 95% CI 0.65–1.23). The researchers detected no significant interaction by sex, although the study may have been underpowered to identify sex-related differences. Most of the analytical approaches applied produced results consistent with the primary analysis, and the HR estimates for diabetes remained generally similar. However, the per-protocol analysis produced an attenuated and less precise association with diabetes.

Study limitations

The observational claims-based design left the analysis vulnerable to residual confounding and misclassification. The databases lacked baseline weight, body mass index, and HbA1c measurements, while the short follow-up, frequent treatment changes, and predominantly female, insured cohort limited interpretation and generalizability.

Conclusion

The researchers concluded that semaglutide was associated with a lower risk of incident diabetes than liraglutide among adults without recorded diabetes or cardiovascular disease who initiated formulations indicated for chronic weight management in routine clinical practice. The difference emerged during follow-up beyond six months, and no significant difference between the two drugs was established for cardiovascular disease risk, although the limited number of cardiovascular events reduced the precision of these estimates. The results suggest that semaglutide may have greater comparative effectiveness with respect to incident diabetes under real-world conditions, but they do not establish superior preventive efficacy. Further studies with longer follow-up are needed to compare cardiovascular outcomes.