Tirzepatide and Atrial Fibrillation Risk

A meta-analysis published in the Journal of the American Heart Association on September 9, 2026 pooled ten randomized controlled trials of tirzepatide, the dual GIP and GLP-1 receptor agonist Eli Lilly sells as Mounjaro and Zepbound, and asked a narrow question: does the drug raise the odds of atrial fibrillation in adults with overweight or obesity^[1]. The answer the authors reached is a qualified no. The point estimate for atrial fibrillation leaned upward, but the uncertainty around it was wide enough that the researchers concluded tirzepatide was not associated with atrial fibrillation or atrial arrhythmia.

The numbers behind the headline

The team, led by Mansouri and colleagues from institutions in Iran, the United States and Brazil, searched PubMed, Embase and the Cochrane Library for randomized trials that compared tirzepatide with placebo in adults with overweight or obesity. Ten trials qualified, covering 6,515 participants, of whom 4,491 were assigned to tirzepatide. Because atrial fibrillation is rare in trial populations selected for weight or glycaemic outcomes rather than cardiac ones, the authors used a Bayesian binomial-normal hierarchical model, a statistical approach built for pooling low-event-count data without the instability that standard fixed-effect methods produce when many trials report zero events.

Three results came out of that model, and they do not all point the same way. The odds ratio for atrial fibrillation specifically was 2.20, with a 95% credible interval running from 0.81 to 6.75^[1]. Because that interval crosses 1, the finding is not statistically significant: the true effect could be a doubling of risk, no effect at all, or something in between. Atrial arrhythmia more broadly came out similarly uncertain, at an odds ratio of 2.16 (0.90 to 5.87). The one result that did clear the bar for significance was the broadest category, any arrhythmia, at an odds ratio of 1.84 with a credible interval of 1.04 to 3.90.

Why a wide credible interval matters more than the headline number

An odds ratio of 2.20 sounds alarming read on its own. Read next to its credible interval, it says something different: with this few events across these ten trials, the data cannot distinguish between tirzepatide roughly doubling atrial fibrillation risk and tirzepatide having no effect on it at all. That is the honest reading of a wide interval, and it is also why the authors did not call the result a safety signal. The one endpoint that did reach significance, any arrhythmia, is a composite that lumps atrial fibrillation together with other rhythm disturbances, including the sinus tachycardia already known from tirzepatide's trial programme. A composite can cross the significance threshold even when none of its individual components do, because pooling more events narrows the interval. That is a statistical property, not proof that every rhythm category inside the composite is independently elevated.

What the drug label already says about heart rate

Mounjaro's prescribing information already documents a heart rate signal, separate from this new meta-analysis. The label reports a mean increase in heart rate of 2 to 4 beats per minute on tirzepatide compared with 1 beat per minute on placebo, and it records episodes of sinus tachycardia, a fast but regular heartbeat originating from the heart's normal pacemaker, in 4.3% to 10% of trial participants depending on dose^[2]. The label states plainly that the clinical relevance of that heart rate increase is uncertain. What the label does not do is single out atrial fibrillation as a listed risk. That gap between a documented heart rate effect and an undocumented atrial fibrillation risk is exactly the space the new meta-analysis tried, and largely failed, to fill with a definitive answer.

How this compares with semaglutide

The picture looks different for a related but distinct drug. A 2025 systematic review and meta-regression in the European Journal of Preventive Cardiology pooled 26 randomized trials of semaglutide, covering 48,583 participants and 541 new-onset atrial fibrillation cases, and found a 17% reduction in new-onset atrial fibrillation on semaglutide compared with controls, with an odds ratio of 0.83 (95% confidence interval 0.70 to 0.98)^[3]. That interval sits entirely below 1, so unlike the tirzepatide result, it counts as a genuine finding rather than a coin flip. The two drugs share a GLP-1 mechanism, but tirzepatide adds a GIP receptor arm that semaglutide does not have, the trial populations differ, and the semaglutide analysis pooled roughly four times as many participants and more than twice as many trials. Different molecules, different evidence bases, and for now, different answers on this specific question.

What this is not

This is not a change to tirzepatide's approved label, and it is not a finding that overturns the weight loss and glycaemic control data that supported Mounjaro's and Zepbound's approvals in the first place. The 72-week SURMOUNT-1 trial that anchored those approvals, for instance, recorded body-weight reductions of up to 20.9% on the highest dose in 2,539 adults with obesity^[4]. None of the ten trials pooled in the atrial fibrillation meta-analysis was designed with atrial fibrillation as a primary endpoint, which means the event counts feeding the model were small and the researchers were, by their own description, working at the edge of what a rare-event analysis can resolve. It is also not evidence that tirzepatide protects against arrhythmia the way the Cesaro meta-analysis found for semaglutide. The honest summary sits between those two poles: an uncertain signal that needs a larger, purpose-built study before anyone can call it settled.

Why this matters

If you have a personal or family history of atrial fibrillation and are considering or already taking tirzepatide, this is a reasonable thing to raise with the clinician managing your care, alongside the heart rate changes already on the label. It is not, on this evidence, a reason to conclude the drug causes atrial fibrillation. Regulators track signals like this one over time as more trial and real-world data accumulate, and a single meta-analysis with a wide credible interval is the beginning of that tracking, not the end of it. For the current state of tirzepatide's approved indications and regulatory status, see the regulation section of the tirzepatide page.