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

Semaglutide vs Tirzepatide Heart Risks

A critical letter followed a large semaglutide-versus-tirzepatide target trial emulation. Here is what the observational evidence can and cannot show.

Why we wrote this. A methodological letter can be mistaken for a new trial result. We separated the critique's limited indexed record from the original observational study and its stated limitations.

In this article (5 sections)
  1. What the original study did
  2. What the analysis reported
  3. Why matching does not settle the comparison
  4. What the methodological letter adds
  5. What we do not know yet

A 2026 target trial emulation reported that tirzepatide initiation was associated with lower risks of several cardiovascular and cerebrovascular outcomes than semaglutide initiation in a large TriNetX dataset. This was an observational comparison, not a randomized head-to-head cardiovascular outcomes trial, so the findings do not establish that tirzepatide caused the differences[2]. A later letter by Yan and Zeng raised methodological concerns, but its PubMed record contains no abstract. The specific criticisms should not be inferred from the letter's title alone[1].

The useful reading is therefore narrower: the study found comparative associations after matching measured baseline factors. It also documented important limits involving residual confounding, treatment changes, diagnostic coding, follow-up and generalizability[2]. Those limits matter before applying the estimates to an individual choosing between semaglutide and tirzepatide.

What the original study did

The investigators used the TriNetX Research Network to identify 217,920 adults who initiated semaglutide or tirzepatide. They separated people with type 2 diabetes from those without documented diabetes, then performed one-to-one propensity score matching within each stratum. The matched analysis included 48,507 pairs with type 2 diabetes and 60,453 pairs without diabetes[2].

Matching covered recorded demographics, body measurements, vital signs, laboratory values, comorbidities and prior medicine use. The main outcomes were atrial fibrillation, heart failure and acute myocardial infarction. Secondary outcomes included ischemic stroke or transient ischemic attack, hemorrhagic stroke and peripheral artery disease. Outcomes were assessed in fixed one-year and three-year windows[2].

The study described this design as a target trial emulation. That framework asks researchers to specify the observational analysis as if it were trying to mirror a hypothetical randomized trial. It can make design choices clearer and reduce some avoidable biases. It does not randomize treatment, recover unrecorded information or guarantee that the groups are otherwise exchangeable.

What the analysis reported

The paper reported lower risks with tirzepatide across all evaluated outcomes at one and three years. For heart failure at one year, the risk ratio was 0.82 in participants with type 2 diabetes and 0.60 in participants without diabetes. The corresponding absolute risk differences were 1.5 and 0.9 percentage points[2]. Relative and absolute estimates answer different questions, so both are needed to understand the size of an association.

The authors also reported evidence that diabetes status modified the one-year associations for atrial fibrillation, heart failure and acute myocardial infarction. They wrote that the risk-reduction associations appeared to strengthen over three years[2]. That pattern remains observational. It should not be rewritten as proof that longer exposure progressively produces greater cardiovascular protection.

Why matching does not settle the comparison

Propensity score matching can balance the variables that were measured and entered into the model. It cannot balance factors that were missing, poorly recorded or unknown. Prescribing may differ according to disease severity, insurance coverage, prior treatment response, access, clinician preference and the timing of market uptake. Any of those factors could influence both medicine choice and later outcomes.

The original authors explicitly acknowledged residual confounding and selection bias. They also noted that medication adherence and persistence were not captured. Participants were analyzed according to the medicine they started, regardless of later discontinuation, switching or other treatment changes[2]. In an observational database, that approach cannot be assumed to behave exactly like intention-to-treat analysis in a randomized trial.

Outcome identification relied on diagnostic codes, creating a risk of misclassification. TriNetX used fixed analysis windows rather than continuous time-to-event tracking, and the authors could not report a traditional median follow-up. The network was primarily based in the United States, which may limit transportability to other health systems and populations[2].

What the methodological letter adds

PubMed indexes the Yan and Zeng publication as a letter titled "Cardiovascular and Cerebrovascular Outcomes Risk Reduction Associated With Semaglutide vs Tirzepatide: Methodological Concerns." The record has no abstract, and its publisher page was not openly readable during our review. Crossref identifies the Azzam target trial emulation as the letter's sole cited reference[1].

That record establishes that a methodological critique was published. It does not disclose the critique's individual arguments. We therefore do not attribute particular objections to Yan and Zeng. The concerns discussed above come from the original paper's methods and stated limitations, not from an assumed reconstruction of the letter.

This distinction is important. A commentary can identify weaknesses, request clarification or dispute an interpretation, but it is not a new trial and does not supply new patient outcomes. Conversely, the existence of a critical letter does not by itself invalidate the original estimates. Readers need the full exchange, including any author response, to judge disputed technical points.

What we do not know yet

The analysis does not tell us whether tirzepatide would outperform semaglutide on cardiovascular events in a randomized population assigned directly to one medicine or the other. It also cannot show how switching, discontinuation, dose escalation or adherence influenced the recorded associations. Three years is short for some cardiovascular questions, while fixed windows obscure the distribution of actual follow-up time.

The study did not turn its database estimates into a treatment rule for an individual patient. Cardiovascular history, diabetes status, kidney function, other medicines, contraindications, tolerability, access and approved indications all matter. A clinician should interpret those factors rather than selecting treatment from observational risk ratios alone.

The next useful evidence would include transparent responses to the published methodological critique, analyses that test unmeasured-confounding assumptions and randomized comparative cardiovascular outcomes. Until then, the paper is evidence of an association in a matched real-world dataset, not proof of comparative cardiovascular superiority.

Medical disclaimer: 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 of semaglutide versus tirzepatide?

No. It was a retrospective observational analysis designed as a target trial emulation. Propensity score matching balanced recorded baseline factors, but treatment was not randomly assigned.

What did the study report for heart failure?

At one year, the reported heart-failure risk ratio for tirzepatide versus semaglutide was 0.82 among matched participants with type 2 diabetes and 0.60 among those without documented diabetes. The absolute risk differences were 1.5 and 0.9 percentage points.

What methodological concerns did the later letter raise?

The indexed record identifies the publication as a methodological critique, but PubMed provides no abstract and the full letter was not openly readable during review. We do not infer or attribute specific criticisms that we could not verify.

Does the study prove that tirzepatide prevents more cardiovascular events?

No. It reports associations after matching measured factors. Residual confounding, treatment changes, coding error and selection effects remain possible, so randomized comparative evidence would be needed for a stronger causal conclusion.

Sources

  1. [1]Yan J, Zeng G. Methodological concerns regarding semaglutide versus tirzepatide cardiovascular outcomes. JACC: Advances. PMID 42962433.Tier 1 · primary↩
  2. [2]Azzam AY et al. Semaglutide versus tirzepatide cardiovascular and cerebrovascular outcomes: a target trial emulation. JACC: Advances. PMC13400122.Tier 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.

Read the pillars