Independent · Evidence-led · We don't sell peptides
EU / NordicsUpdated weeklyEN

Explore this article's sources with AI

ChatGPTClaudePerplexityGeminiGrokGoogle AI

Follow PeptideMethods on Google

First published

GLP-1 vs SGLT-2: What Real-World Data Found

A target trial emulation compared three GLP-1 medicines with SGLT-2 inhibitors for weight and HbA1c in type 2 diabetes and obesity.

Why we wrote this. Target trial emulation can sound more definitive than it is. This article explains both the comparative estimates and the remaining observational uncertainty.

In this article (5 sections)
  1. What a target trial emulation is
  2. What the study found
  3. How this relates to randomized evidence
  4. Why HbA1c and weight need separate interpretation
  5. What we don't yet know

An observational study emulating target trials found that, among people with type 2 diabetes and obesity, semaglutide was associated with greater reductions in weight and HbA1c than SGLT-2 inhibitors at 18 months[1]. The comparison is useful real-world evidence, but it was not a randomized trial.

The researchers separately compared three GLP-1 receptor agonists with SGLT-2 inhibitors: semaglutide, liraglutide and dulaglutide. Follow-up extended to three years for semaglutide and five years for the other two medicines[1]. The results should not be collapsed into one class-wide ranking because each analysis involved different initiators and follow-up.

What a target trial emulation is

A target trial emulation starts by specifying the randomized trial researchers would ideally run, including eligibility, treatment strategies, follow-up and outcomes. Investigators then apply that plan to observational records. The structure can prevent common design errors, such as comparing people at mismatched points in treatment.

This study used medical records from Clalit Health Services and inverse-probability weighting to adjust for measured differences between treatment groups[1]. Weighting attempts to make comparison groups more similar on recorded characteristics. It cannot balance factors that were not captured, were measured poorly or were omitted from the model.

The word emulation therefore does not mean that records became randomized. Clinicians and patients chose treatment in ordinary care. Kidney function, cardiovascular disease, cost, availability, expected adverse effects and patient preference may have influenced those choices. Some of those influences may also affect later weight or HbA1c.

What the study found

The eligible groups included 4,296 people initiating liraglutide, 11,159 initiating dulaglutide and 5,041 initiating semaglutide. Each GLP-1 receptor agonist was compared with SGLT-2 inhibitor initiation in a separate emulated trial[1].

At 18 months, semaglutide was associated with 2.6 kg more weight loss than SGLT-2 inhibitors, with a 95% confidence interval from 1.2 to 4.0 kg. The estimated HbA1c reduction was 0.6 percentage points greater, with a 95% confidence interval from 0.5 to 0.8[1].

For liraglutide and dulaglutide, long-term weight loss was similar to SGLT-2 inhibitors, with reported losses of 6.4 to 7.3 kg at five years. HbA1c reductions at 18 months were 0.4 percentage points greater for liraglutide and 0.7 points greater for dulaglutide than SGLT-2 inhibitors, and the advantage was maintained during follow-up[1].

These are average estimates. They do not predict one person's response, and a difference in one outcome does not make a medicine the better choice overall. Treatment decisions also involve cardiovascular and kidney indications, adverse effects, contraindications, route of administration, access and the person's priorities.

How this relates to randomized evidence

SUSTAIN 8 directly randomized adults with type 2 diabetes taking metformin to weekly semaglutide or daily canagliflozin for 52 weeks. It found larger average reductions in HbA1c and body weight with semaglutide[2]. That trial supports the direction of the semaglutide comparison, but it tested one SGLT-2 inhibitor under a defined protocol and shorter follow-up.

The new observational analysis extends the question into routine practice and longer periods[1]. That adds relevance because adherence, switching and clinical follow-up differ outside trials. It also adds uncertainty because treatment assignment and missing data are not controlled in the same way.

The two designs answer related questions. A randomized trial estimates effects under its protocol and eligibility rules. A well-designed observational emulation estimates effectiveness in the records and care setting it captures. Agreement raises confidence, but it does not erase the limitations of either source.

Why HbA1c and weight need separate interpretation

HbA1c reflects average blood glucose over the preceding months. Weight is a separate outcome. A medicine may have clinically important effects on glucose even when weight differences are modest, and vice versa. Baseline values, other medicines and treatment changes during follow-up influence both measures.

The study compared GLP-1 receptor agonists as diabetes treatments in people who also had obesity. It does not establish results for people without diabetes or for products and doses used solely for chronic weight management. It also does not compare semaglutide with tirzepatide.

What we don't yet know

The abstract does not provide all details about treatment discontinuation, switching, missing measurements or the balance achieved for every baseline factor. Those details matter more as follow-up lengthens. People remaining observable at five years may differ from those who stop treatment or leave the health system.

Residual confounding remains possible. Direct long-term randomized comparisons would give stronger evidence about relative effects, while pragmatic trials could preserve more of routine care. Future analyses should report adverse events, kidney and cardiovascular outcomes, persistence and how results vary across baseline risk groups.

Do not change diabetes medicine based on an average comparison. A qualified healthcare professional can weigh glucose control, kidney and cardiovascular health, adverse effects and access in the context of an individual's history.

This article is for educational and journalistic purposes only and does not constitute medical advice. Peptides discussed may be classified as prescription medicines or research chemicals depending on your jurisdiction. Always consult a qualified healthcare professional before using any peptide product. PeptideMethods.com does not sell, distribute, or facilitate the sale of any peptide product.

Frequently asked

Was this a randomized trial?

No. It was an observational analysis structured to emulate target trials and used weighting to adjust for measured differences between treatment groups.

How did semaglutide compare with SGLT-2 inhibitors?

At 18 months, semaglutide was associated with 2.6 kg greater weight loss and a 0.6 percentage-point greater HbA1c reduction.

Did all GLP-1 medicines produce more weight loss?

No. Long-term weight loss with liraglutide and dulaglutide was similar to SGLT-2 inhibitors, although estimated HbA1c reductions were greater.

Do these results apply to people without diabetes?

Not directly. The study population had both type 2 diabetes and obesity, so it cannot establish comparative effects in people without diabetes.

Sources

  1. [1]GLP-1 RA and SGLT-2 inhibitor target trial emulation, PubMed PMID 43003406Tier 1 · primary↩
  2. [2]SUSTAIN 8 randomized semaglutide versus canagliflozin trial, PubMed PMID 31540867Tier 1 · primary↩

No revisions yet. First published .

About the editorial team

PeptideMethods is written and edited by the PeptideMethods Editorial Team and published by Digital Compass Group Ltd. The team is not made up of medical professionals; every health, regulatory or dosage claim on the site is tied to a primary source and is not a substitute for advice from a qualified clinician.

See our editorial policy and methodology for how we research, source and verify.