GLP-1 and Atrial Fibrillation: SELECT, FLOW, and Anticoagulation Stack

Atrial fibrillation affects roughly 6 million US adults and is projected to reach 12 million by 2030. The Framingham analysis (Wang 2004^[1]) and the Women’s Health Study (Tedrow 2010^[2]) established obesity as one of the strongest modifiable drivers of new AF. The LEGACY cohort (Pathak 2015^[3]) and ARREST-AF (Pathak 2014^[4]) then showed that sustained weight loss of 10% or more cuts AF burden and improves ablation outcomes. With SELECT (Lincoff 2023^[5]), STEP-HFpEF (Verma 2024^[6]), and SUMMIT (Packer 2025^[7]) now adding GLP-1 and GIP/GLP-1 data, the obesity-medicine and electrophysiology pathways are starting to converge. This article walks through what the published evidence says, how GLP-1 therapy stacks with anticoagulation, and the practical clinic workflow.

The honest summary

• Obesity is causal, not just associated. Framingham (Wang 2004^[1]) reported a 4–5% increase in AF risk per unit of BMI; the Women’s Health Study (Tedrow 2010^[2]) replicated the dose-response in women. Modifiable risk factors account for roughly half to seventy percent of new AF cases.
• Weight loss reduces AF burden. LEGACY (Pathak 2015^[3]) followed 355 AF patients in a structured weight-management program for 5 years; sustained loss of ≥ 10% body weight was associated with a 6-fold higher rate of arrhythmia-free survival vs < 3% loss. ARREST-AF (Pathak 2014^[4]) showed the same pattern improved ablation success.
• GLP-1 evidence is accumulating. SELECT (Lincoff 2023^[5]) followed 17,604 non-diabetic adults with obesity on semaglutide 2.4 mg for a median 3.3 years and reported a 20% reduction in 3-point MACE. STEP-HFpEF (Verma 2024^[6]) reported that baseline AF did not blunt semaglutide’s heart-failure benefit. SUMMIT (Packer 2025^[7]) showed similar structural and symptom benefits with tirzepatide in HFpEF with obesity.
• Anticoagulation stays the same. The 2024 ESC AF guideline (Van Gelder 2024^[8]) keeps the DOACs (apixaban, rivaroxaban, dabigatran, edoxaban) as first-line for stroke prevention based on CHA2DS2-VA score. Warfarin remains the choice for mechanical valves or moderate-to-severe mitral stenosis. There is no guideline-level reason to change anticoagulation when starting a GLP-1.

Why obesity drives atrial fibrillation

The mechanism is structural. Excess adiposity, especially epicardial adipose tissue surrounding the left atrium, drives chronic low-grade inflammation, fibrosis, and progressive left-atrial enlargement. Over years the atrium remodels into a substrate that supports re-entrant arrhythmia. Obstructive sleep apnea (often co-traveling with obesity) adds intermittent hypoxia and sympathetic surges. Hypertension adds afterload. By the time AF declares itself, the substrate has been forming for a decade or more. That is why short bursts of weight loss do not reliably reverse AF, while sustained loss of 10% or more — the threshold LEGACY identified^[3] — can shrink the left atrium and reduce arrhythmia burden.

LEGACY and ARREST-AF: weight loss as rhythm therapy

The LEGACY trial (Pathak 2015, J Am Coll Cardiol^[3]) enrolled 355 overweight or obese patients with symptomatic AF into a structured weight-management program and followed them for 5 years. Patients who sustained ≥ 10% body-weight loss had a roughly 6-fold higher arrhythmia-free survival than those who lost < 3%. Patients in the highest weight-loss tier also had measurable reductions in left-atrial volume, BP, lipids, and inflammatory markers. ARREST-AF (Pathak 2014^[4]) extended the finding to the post-ablation population: aggressive risk-factor reduction (including weight loss) improved single-procedure arrhythmia-free survival from 32% to 87% over 5 years.

These cohorts predate the GLP-1 era and used diet, exercise, and lifestyle counseling. The relevance for GLP-1 medicine is that they established a magnitude-of-loss threshold: ≥ 10% sustained loss is the dose that moves AF outcomes. That threshold is now routinely achievable with semaglutide and tirzepatide.

SELECT, STEP-HFpEF, and SUMMIT

SELECT (Lincoff 2023, NEJM^[5]) is the largest cardiovascular outcomes trial in non-diabetic obesity: 17,604 adults with established cardiovascular disease and BMI ≥ 27 randomized to semaglutide 2.4 mg weekly or placebo for a median 3.3 years. The primary 3-point MACE endpoint (CV death, non-fatal MI, non-fatal stroke) dropped 20%. AF was not the primary endpoint, but adjudicated AF events were reported as a secondary safety outcome, and the signal was neutral-to-favorable rather than harmful.

STEP-HFpEF and STEP-HFpEF DM were trials of semaglutide 2.4 mg in patients with heart failure with preserved ejection fraction and obesity. A pooled analysis by Verma 2024^[6] (J Am Coll Cardiol) looked specifically at the AF subgroup and found that baseline AF did not blunt the semaglutide effect on KCCQ symptom score, 6-minute walk distance, or weight loss. The data argue that semaglutide is an appropriate add-on in obesity-related HFpEF whether or not AF is present.

SUMMIT (Packer 2025, NEJM^[7]) extended the evidence to tirzepatide. Patients with HFpEF and obesity on tirzepatide had improved KCCQ scores, reduced N-terminal pro-BNP, and lower rates of worsening heart failure events. The implication for AF is indirect but real: tirzepatide delivers more weight loss than semaglutide in head-to-head analyses, and weight loss is the lever LEGACY identified.

The mechanism: why GLP-1 may modify AF substrate

Three plausible pathways:

• Weight loss reduces epicardial adipose tissue. EAT volume around the left atrium correlates with AF incidence and recurrence after ablation. Weight loss of ≥ 10% measurably reduces EAT, and the GLP-1 trials consistently deliver this magnitude.
• Anti-inflammatory effect. Semaglutide and tirzepatide both reduce hsCRP and IL-6 in clinical trials. Chronic atrial inflammation is part of the substrate that supports re-entry.
• Blood-pressure reduction. Mean systolic drops of 4–6 mm Hg are typical on full-dose semaglutide and tirzepatide. Lower afterload reduces left-atrial wall stress over time. See our companion article on GLP-1 + diuretics for the BP and electrolyte side.

Whether GLP-1 receptor agonists exert any direct anti-arrhythmic effect independent of weight loss is unproven. The honest interpretation of the 2026 evidence base is that the dominant mechanism is weight loss; any direct effect is a hypothesis, not a finding.

The anticoagulation stack

Patients with AF and a CHA2DS2-VA score ≥ 2 (men) or ≥ 3 (women, before the 2024 sex-neutral simplification) need chronic anticoagulation. The 2024 ESC AF guideline (Van Gelder 2024^[8]) keeps the DOACs (apixaban, rivaroxaban, dabigatran, edoxaban) as first-line. Warfarin is reserved for mechanical valves or moderate-to-severe mitral stenosis. There is no guideline reason to change anticoagulation when starting a GLP-1.

That said, the GLP-1 gastric-emptying delay raises a pharmacokinetic question for the orally absorbed DOACs. The published data (reviewed in our companion article on GLP-1 + warfarin, Eliquis, and Xarelto) is that DOAC absorption is largely preserved because the agents are peptide-resistant small molecules; warfarin INR variability has been observed in case reports but is manageable with standard anticoagulation-clinic monitoring. The practical pattern: continue the DOAC, do not change the dose at GLP-1 initiation, and check INR more often on warfarin during dose-escalation weeks.

Antiarrhythmics and rate-control drugs

Patients with AF on a GLP-1 may also be on rhythm- or rate-control medications. The published interactions worth knowing:

• Amiodarone: long half-life (weeks) dominates any short-term gastric-emptying effect. No dose change needed.
• Sotalol: QT-prolonging; standard ECG monitoring continues. No published GLP-1 interaction.
• Flecainide: requires structural-disease screening before use; not GLP-1 specific.
• Dronedarone: contraindicated in symptomatic HF; the HFpEF benefit of GLP-1 therapy may shift the risk-benefit toward stopping dronedarone in favor of weight loss plus other rhythm strategies.
• Beta-blockers: the small resting heart-rate rise seen with GLP-1 medications is partly offset by ongoing beta-blockade. See our companion article on GLP-1 and heart rate.
• Diltiazem and verapamil: watch for additive bradycardia in rate-control patients; routine pulse checks during titration.
• Digoxin: narrow therapeutic window; a digoxin level at baseline and 4–6 weeks into GLP-1 titration is reasonable, given gastric-emptying-related absorption changes.

Magnitude: AF burden change at 12 months by intervention

Rate vs rhythm control: where GLP-1 changes the calculus

AFFIRM (Wyse 2002, NEJM^[9]) famously found no mortality difference between rate and rhythm control in older AF patients. The field has since evolved; EAST-AFNET 4 and CASTLE-AF have shifted practice toward earlier rhythm control in younger patients with HF or symptomatic AF. CASTLE-AF (Marrouche 2018, NEJM^[10]) found catheter ablation reduced death and HF hospitalization in AF with HFrEF. The GLP-1 angle is upstream: in obese patients with new or paroxysmal AF, sustained ≥ 10% loss can shift the risk-benefit toward attempted rhythm restoration after weight stabilizes, because the substrate has partly regressed. The practical sequence is weight loss first, then re-evaluate symptom burden, then decide on ablation with cardiology and electrophysiology input.

The practical clinic pathway

1. New AF on a GLP-1 dose-escalation week: rule out hyperthyroidism (TSH, free T4), check electrolytes (K, Mg) given the gastric-emptying-related shifts, hold the next dose increase, and refer for cardiology workup.
2. Stable AF on a DOAC starting a GLP-1: continue the DOAC unchanged. Routine titration. Monitor weight, BP, and resting HR. No INR change because there is no INR.
3. Stable AF on warfarin starting a GLP-1: continue warfarin; check INR weekly during the first month of GLP-1 dose-escalation. Counsel on appetite changes affecting vitamin-K intake (variable green-leafy consumption is a bigger driver of INR drift than the GLP-1 itself).
4. Sustained 10%+ loss achieved: revisit AF burden and symptom score with cardiology. Discuss whether a rhythm-restoration attempt (cardioversion or ablation) is now appropriate given the regressed substrate.
5. Pre-cardioversion: standard 3-week therapeutic DOAC or TEE-guided. GLP-1 medications are not held for cardioversion itself; if general anesthesia is planned, follow the ASA 2023 guidance on holding GLP-1 the day of the procedure (clear liquids only) at the anesthesiologist’s discretion.
6. Pre-ablation: the LEGACY and ARREST-AF evidence supports delaying ablation, where clinically safe, until weight has stabilized at the new lower set point. Ablation success is meaningfully higher in patients who have lost ≥ 10%.

Insurance and cost

AF care is universally covered: anticoagulation, cardioversion, catheter ablation, and electrophysiology consultation all sit inside standard cardiology benefits. DOACs run $400–500 per month at full retail and $20–50 per month with most commercial insurance copay cards (apixaban and rivaroxaban both have manufacturer programs). Catheter ablation runs $30,000–$50,000 case-rate but is generally a covered benefit when symptomatic AF persists on medical therapy. GLP-1 coverage for AF-related weight management depends on the underlying obesity coverage; SELECT-style CV indications are increasingly recognized by commercial payors.

Related research and tools

• GLP-1 + warfarin, Eliquis, and Xarelto — the anticoagulation interaction in depth
• GLP-1 and heart rate — resting HR, palpitations, and the sympathetic mechanism
• STEP-HFpEF and SUMMIT — the heart-failure trials and the AF subgroup
• GLP-1 + diuretics — BP, potassium, and the electrolyte stack
• SELECT trial cardiovascular benefits — the 17,604-patient MACE outcome in detail
• GLP-1 and long COVID — post-acute cardiac sequelae and the dysautonomia overlap

Important disclaimer. This article is educational and does not constitute medical advice. Atrial fibrillation requires individualized stroke-risk assessment and ongoing cardiology input. Anticoagulation choice should be made with the prescribing clinician based on CHA2DS2-VA score, bleeding risk, comorbidities, and renal function. GLP-1 initiation, dose escalation, and procedural holds should be coordinated with the obesity-medicine clinician, cardiologist, and anesthesia team as appropriate. PMIDs were verified live against the PubMed E-utilities API on 2026-05-29.

Last verified: 2026-05-29. Next review: every 12 months, or sooner if new prospective trial data on GLP-1 therapy and AF burden, ablation outcomes, or anticoagulation interactions is published.