A network meta-analysis of 94 randomised trials, published online in Heart Rhythm on 26 October 2026, found that people with type 2 diabetes or obesity who were given subcutaneous semaglutide developed new atrial fibrillation (AF) less often than those given placebo. Dulaglutide pointed the other way. For anyone already taking one of these drugs, nothing here changes a prescription, but it adds a heart-rhythm signal to a body of evidence that has so far centred on weight, blood sugar and heart attacks.
What the analysis pooled
The authors searched the literature from its beginning to December 2025 and included randomised controlled trials in people with diabetes or obesity who received a GLP-1 receptor agonist or a co-agonist. That produced 94 trials, 155,989 patients and 12 different drugs. Few trials compare two of these drugs directly, so a network meta-analysis links them through shared comparators, usually placebo, and estimates the comparisons nobody ran.
Against placebo, subcutaneous semaglutide had a risk ratio for new AF of 0.82 (95% confidence interval 0.68 to 0.99). That is about 18% fewer cases relative to placebo, but the interval runs from a 32% reduction down to a 1% reduction, so the true effect could be small. Against tirzepatide the risk ratio was 0.57 (0.37 to 0.88): a 43% lower relative risk, with a plausible range of 12% to 63%.
The abstract also reports that dulaglutide was associated with an increased risk of AF and that lixisenatide was associated with fewer atrioventricular blocks, a different conduction problem, but it does not give the figures for either. The authors ranked the drugs by p-score, a measure from 0 to 1 of how likely each is to be the best option for a given outcome. For AF, semaglutide scored 0.79, tirzepatide 0.18 and dulaglutide 0.18.
How it fits with what came before
Excess weight and type 2 diabetes are both established risk factors for AF, so a drug that lowers weight and glucose might be expected to lower AF too. The tirzepatide result does not fit that simple story. In trials tirzepatide generally produces more weight loss than semaglutide (see our Wegovy versus Zepbound comparison), yet here it ranked near the bottom for AF. That could reflect a genuine difference between the drugs, or simply differences in the trials behind each estimate, which the abstract does not break down.
It is worth being precise about what the tirzepatide figure says. It compares tirzepatide with semaglutide; the abstract gives no estimate for tirzepatide against placebo, so it does not show that tirzepatide raises the risk of AF. Dulaglutide is the only drug the abstract describes as associated with more AF.
If the signal holds, it runs in the protective direction for semaglutide. It is not a new safety warning to add to the effects covered on our semaglutide side-effects page.
What is not yet known
This report is based on the published abstract; the full paper was not available to us. The abstract does not say how AF was identified in each trial. The pooled trials were designed around glucose, weight or cardiovascular events rather than arrhythmia, and the authors themselves call for studies focused on arrhythmic outcomes. Absolute event rates are not given in the abstract, so the number of cases avoided per thousand people cannot be worked out from it. The abstract does not say how much of the semaglutide-versus-tirzepatide estimate comes from head-to-head trials and how much is indirect, and the upper limit of 0.99 against placebo sits close to no effect. The search ended in December 2025, so trials reported this year are not included, and the abstract does not state funding.
Sources: Barbosa LM, Rodrigues Oliveira VM, Ruelas MG, et al. “Glucagon-Like Peptide-1 Receptor Agonists and Risk of Arrhythmias: A Network Meta-Analysis.” Heart Rhythm, published online ahead of print 26 October 2026. Journal article (DOI); PubMed record, PMID 42933519.
