Atrial fibrillation: ESC 2024 versus ACC/AHA 2023

A 58-year-old man with hypertension and obesity comes to the outpatient clinic with palpitations that have come and gone for a few months. An electrocardiogram performed in the emergency department during one of the episodes showed atrial fibrillation (AF), and the rhythm returned to sinus on its own before discharge. He drinks beer almost every weekend, hardly exercises, and wants to know two things: whether he needs to take an anticoagulant and whether ablation is an option right away.

The two major guidelines that guide this consultation were published a few months apart. The American guideline from the American College of Cardiology, the American Heart Association, the American College of Clinical Pharmacy, and the Heart Rhythm Society was published at the end of 2023 (Joglar et al., 2024). The European guideline from the European Society of Cardiology (ESC), produced with the European association of cardiothoracic surgery, came out in 2024 (Van Gelder et al., 2024).

This text compares the two by the reasoning of management, without becoming a summary of each recommendation. In several points they agree on the essentials and differ in detail; in others, such as the stroke risk score and AF seen only by pacemaker, the difference changes management.

Key points

• ACC/AHA 2023 replaces the classification based only on duration with stages that go from AF risk to permanent AF (Joglar et al., 2024).
• ESC 2024 organizes care in AF-CARE: comorbidities and risk factors, avoiding stroke and thromboembolism, reducing symptoms through rate and rhythm control, and dynamic reassessment (Van Gelder et al., 2024).
• ESC removed sex from the score and began using CHA2DS2-VA; ACC/AHA keeps CHA2DS2-VASc with different thresholds by sex and accepts other validated scores.
• In AF detected only by device, ACC/AHA uses episode duration to grade the recommendation, and ESC speaks of considering direct oral anticoagulant in high thromboembolic risk.
• Both treat weight, alcohol, and exercise as part of arrhythmia treatment.
• Both place catheter ablation as first-line in selected patients with symptomatic paroxysmal AF, with class 1 and level A.

Summary

• Two ways of organizing the same disease
• Diagnosis and screening
• Risk scores: CHA2DS2-VA versus CHA2DS2-VASc
• Anticoagulation: what the two recommend
• AF detected by device
• Risk factors as part of treatment
• Rhythm control, ablation, and heart failure
• How NEXO helps train this reasoning
• Frequently asked questions
• Take-home messages

Two ways of organizing the same disease

ACC/AHA 2023 states that the old classification, based only on arrhythmia duration, tended to emphasize AF after diagnosis and therapeutic interventions (Joglar et al., 2024). The new proposal uses stages and treats AF as a progressive disease, with different strategies at each phase, from prevention to screening and to rate and rhythm control (Joglar et al., 2024). The guideline notes that the stages overlap: management of risk factors, for example, crosses several of them (Joglar et al., 2024).

ESC 2024 chose another entry point. It keeps the categories by duration, such as paroxysmal AF, which ends within up to seven days, and organizes care in the acronym AF-CARE (Van Gelder et al., 2024). The letter C comes first and deals with comorbidities and risk factors; A deals with avoiding stroke and thromboembolism; R deals with reducing symptoms and morbidity through rate and rhythm control; E deals with dynamic assessment and reassessment (Van Gelder et al., 2024).

What the two structures have in common

ACC/AHA describes care as a building whose foundation is the treatment of comorbidities and risk factors, with three processes on top: assessment and treatment of stroke risk, optimization of modifiable factors, and symptom control by rate and rhythm (Joglar et al., 2024). ESC puts the same content in another order and insists that comorbidities come first, because they affect all other steps (Van Gelder et al., 2024).

In practice, anyone who reads both finds the same message: AF is no longer just a rhythm to correct and is now treated as a marker of an atrial and cardiometabolic disease. The difference lies in how each teaches this to the clinician, by disease stages in the American case and by care tasks in the European one.

Diagnosis and screening

The ESC requires electrocardiographic confirmation to diagnose AF, with class I and level A, and accepts a single-lead tracing of 30 seconds or more, including from a wearable device, when the rhythm is analyzable (Van Gelder et al., 2024). This rule has a direct consequence for those who arrive with a watch alert: the alert alone raises suspicion, but the decision to anticoagulate depends on the tracing.

For population screening, the ESC suggests considering AF screening in people aged 75 or older, or 65 or older with additional risk factors, with class IIa and level B (Van Gelder et al., 2024). The ACC/AHA gives screening its own place within the stages, as part of the care of those at risk before diagnosis (Joglar et al., 2024).

Back to the case

The patient from the opening has an electrocardiogram during the episode, which meets the confirmation requirement of both guidelines. The next step, in both, is to estimate stroke risk and bleeding risk, and to look for the factors that drive the arrhythmia. The blog has already discussed why AF matters in people who cannot suddenly lose consciousness, such as pilots, in the text on aviation cardiology and the risk of sudden incapacitation.

Risk scores: CHA2DS2-VA versus CHA2DS2-VASc

The most visible change in ESC 2024 is the removal of female sex from the score. The guideline recommends CHA2DS2-VA to estimate thromboembolic risk, with oral anticoagulation recommended at a score of 2 or more (class I, level C) and to be considered at a score of 1 (class IIa, level C) (Van Gelder et al., 2024).

The ACC/AHA keeps CHA2DS2-VASc and starts talking about annual risk. It recommends anticoagulation when the estimated annual risk of thromboembolism is 2 in 100 or more, which corresponds to a score of 2 or more in men and 3 or more in women, with class 1 and level A (Joglar et al., 2024). With an annual risk of 1 to 2 in 100, equivalent to a score of 1 in men and 2 in women, anticoagulation is considered reasonable, with class 2a (Joglar et al., 2024).

Other scores and modifiers

The ACC/AHA gives more room to alternative scores. It recommends assessing annual risk with a validated score, such as CHA2DS2-VASc, and accepts that, in the intermediate-risk patient who remains in doubt, other scores and other variables help in shared decision-making (Joglar et al., 2024). Among the modifiers cited are AF burden, sex and blood pressure control (Joglar et al., 2024).

An Australian population-based study shows why this flexibility matters. In 224,451 patients with a first hospitalization for AF and no anticoagulant, CHA2DS2-VA had the lowest discrimination among the scores tested, with a C statistic of 0.61 (Zubrzycki et al., 2026). The ATRIA score had the best discrimination, with a C statistic of 0.66, and patients at intermediate risk by CHA2DS2-VA had a stroke incidence below the treatment threshold used in the study (Zubrzycki et al., 2026). The authors call for the treatment threshold to be revised, and the message for the clinician is that any score is a rough estimate.

Anticoagulation: what the two recommend

In both guidelines, direct oral anticoagulants are preferred over warfarin in those who are candidates for anticoagulation. The ACC/AHA recommends direct oral anticoagulants over warfarin, except in moderate to severe rheumatic mitral stenosis or a mechanical valve prosthesis, with class 1 and level A (Joglar et al., 2024). The ESC gives the same preference, also for cardioversion, with class I and level A (Van Gelder et al., 2024).

Acetylsalicylic acid lost ground in both. The ACC/AHA classifies as harmful the use of aspirin, alone or with clopidogrel, as an alternative to anticoagulation in those who are candidates for it (Joglar et al., 2024). The ESC advises against adding an antiplatelet agent to the anticoagulant with the aim of preventing stroke, with class III and level B (Van Gelder et al., 2024).

Bleeding risk

The ESC now recommends assessing modifiable bleeding risk factors with class I and level B, and advises against using bleeding scores to deny anticoagulation, with class III (Van Gelder et al., 2024). The ACC/AHA asks, with class 1, that factors indicating higher bleeding risk be sought, such as prior bleeding and use of medications that increase it, so that what can be corrected is corrected (Joglar et al., 2024).

The ESC also takes a position against reduced-dose direct oral anticoagulant regimens outside the specific criteria for each drug, with class III and level B (Van Gelder et al., 2024). This point matters in the polymedicated older adult, in whom dose reduction out of fear of bleeding is common. The blog addresses this decision in deprescribing: how to decide what to stop with the patient.

Left atrial appendage occlusion

The ACC/AHA raised percutaneous left atrial appendage occlusion to class 2a in people with a CHA2DS2-VASc of 2 or more and a long-term, non-reversible contraindication to oral anticoagulation (Joglar et al., 2024). In the patient at high bleeding risk on an anticoagulant, it places the procedure as a possible alternative, with class 2b, noting that the evidence for anticoagulation is more extensive (Joglar et al., 2024).

Table: ESC 2024 and ACC/AHA 2023 side by side

Topic	ESC 2024	ACC/AHA 2023
Organization of care	AF-CARE, with comorbidities first	Disease stages, from the risk phase to permanent AF
Stroke score	CHA2DS2-VA, without sex	CHA2DS2-VASc or another validated score, by annual risk
Anticoagulate	Score of 2 or more (I C); score of 1, consider (IIa C)	Annual risk of 2 in 100 or more (1 A); 1 to 2 in 100, reasonable (2a A)
Device-detected AF	Consider direct oral anticoagulant if risk is raised, with high bleeding risk excluded (IIb B)	Episode of 24 h or more with score of 2 or more (2a); 5 min to 24 h with score of 3 or more (2b); under 5 min, against (3)
Weight	Loss of 10 in 100 or more of body weight (I B)	Loss of at least 10 in 100 of body weight, with BMI above 27 (1 B-R)
Alcohol	Up to three standard drinks per week (I B)	Reduce or eliminate in those seeking rhythm control (1 B-R)
Ablation as first line in paroxysmal AF	I A, within shared decision-making	1 A in selected patients, generally younger and with few comorbidities

Table sources: Van Gelder et al., 2024 and Joglar et al., 2024.

Device-detected AF

Pacemakers, defibrillators and implantable monitors record rapid atrial episodes that the patient does not feel. The ESC defines these high atrial rate episodes as lasting more than 5 minutes with a rate of 170 beats per minute or more, and treats subclinical AF as a device finding (Van Gelder et al., 2024). It suggests that a direct oral anticoagulant may be considered in subclinical AF with raised thromboembolic risk, excluding patients at high bleeding risk, with class IIb and level B (Van Gelder et al., 2024).

The ACC/AHA grades the recommendation by episode duration. With an episode of 24 hours or more and a CHA2DS2-VASc of 2 or more, starting anticoagulation is reasonable within shared decision-making, with class 2a (Joglar et al., 2024). With an episode between 5 minutes and 24 hours and a score of 3 or more, it may be reasonable, with class 2b, and with an episode under 5 minutes, the guideline advises against anticoagulating if the person has no other indication (Joglar et al., 2024).

What the trials showed

The two trials that underpin the topic, NOAH-AFNET 6 and ARTESiA, reached results that call for caution. A meta-analysis of the two studies estimated a relative reduction in stroke of close to 32 in 100 with anticoagulation, alongside a relative increase in major bleeding of close to 62 in 100, with no effect on mortality (Scott, 2026). Scott argues that a single duration threshold is an unstable tool for deciding and proposes looking at the patient's phenotype, such as vascular disease and absolute risk, and using AF burden to define urgency and monitoring (Scott, 2026).

In people who have already had a stroke or a transient ischemic attack, the evidence is even more limited. A 2026 review shows that, in this subgroup, apixaban from ARTESiA reduced ischemic stroke with an increase in major bleeding, while edoxaban from NOAH-AFNET 6 had a small reduction without statistical significance, with a significant increase in bleeding (Liu et al., 2026). The authors attribute part of the disagreement to differences in inclusion: ARTESiA accepted episodes of at least 6 minutes, generally shorter than 24 hours (Liu et al., 2026).

A Finnish economic analysis adds another angle. In a simulated cohort of 20 000 people, anticoagulating all patients with device-detected AF yielded 0.016 quality-adjusted life year more per patient over ten years, and the authors concluded that routine anticoagulation of everyone probably represents low-value care (Winstén et al., 2026). The blog post about stroke mimics reminds us that not every acute deficit is vascular, which matters when discussing stroke in these patients.

Risk factors as part of treatment

Both guidelines treat risk factors as part of arrhythmia treatment, not as a footnote recommendation. ACC/AHA says lifestyle change and risk factor control are a pillar of AF management, to prevent onset, progression and adverse outcomes (Joglar et al., 2024). ESC puts the same idea in the C of AF-CARE and gives class I to several of these measures (Van Gelder et al., 2024).

Weight

In people who are overweight or obese with a BMI above 27, ACC/AHA recommends weight loss with an ideal target of at least 10 in every 100 of body weight, to reduce symptoms, burden, recurrence and progression to persistent AF, with class 1 (Joglar et al., 2024). ESC recommends losing 10 in every 100 or more of body weight as part of comorbidity control, with class I and level B (Van Gelder et al., 2024).

A 2026 review on recurrence after ablation shows that how the intervention is delivered matters. A structured, physician-led program reduced recurrence at 12 months with a hazard ratio of 0.53, while nurse-led care that improved adherence to guidelines without an equivalent structure left the primary outcome unchanged (Fu et al., 2026). The same review describes a benefit of sodium-glucose cotransporter 2 inhibitors on recurrence, especially in people with type 2 diabetes and heart failure, and inconsistent data with GLP-1 receptor agonists (Fu et al., 2026). The blog discussed the relationship between obesity and sleep in tirzepatide, obesity and sleep apnea.

Alcohol and exercise

For alcohol, ESC recommends reducing consumption to three standard drinks or fewer per week, with class I and level B (Van Gelder et al., 2024). ACC/AHA asks that those seeking rhythm control reduce alcohol to a minimum or eliminate it to lower recurrence and AF burden, with class 1 (Joglar et al., 2024). The blog has a post about the patient who arrives drunk at the emergency department, in alcohol beyond withdrawal.

ACC/AHA recommends moderate to vigorous intensity training with a target of 210 minutes per week, except in AF linked to excessive training, with class 1 (Joglar et al., 2024). ESC recommends an individualized exercise program, with class I and level B (Van Gelder et al., 2024).

For the patient in this case, that means the conversation about beer, weight and physical activity is part of arrhythmia treatment, with the same strength of recommendation given to many drugs.

Rhythm control, ablation and heart failure

ACC/AHA emphasizes early and continuous management, with a focus on maintaining sinus rhythm and reducing AF burden (Joglar et al., 2024). It cites EAST-AFNET 4, which randomized patients with recent AF to rhythm or rate control, and the observation that rhythm control is more likely to work when started early (Joglar et al., 2024). ESC suggests considering rhythm control in the first 12 months after diagnosis in selected patients at risk of events, with class IIa and level B (Van Gelder et al., 2024).

In the acute crisis, ESC suggests the strategy of waiting up to 48 hours for spontaneous reversion as an alternative to immediate cardioversion in selected patients, with class IIa and level B (Van Gelder et al., 2024). The approach at first contact recalls the reasoning in the blog text about syncope in the emergency department: separate what is unstable from what can wait.

Ablation as first line

Here the two guidelines converge. ACC/AHA gives class 1 and level A to catheter ablation as first line in selected patients, generally younger and with few comorbidities, with symptomatic paroxysmal AF in whom rhythm control is desired (Joglar et al., 2024). For the other patients with symptomatic paroxysmal or persistent AF in rhythm control, ablation as first line may be useful, with class 2a (Joglar et al., 2024).

ESC recommends catheter ablation as a first-line option in paroxysmal AF, within shared decision-making, with class I and level A; in the 2020 edition, the same indication had class IIa (Van Gelder et al., 2024). For the patient in the case, this means that ablation can enter the conversation right away, alongside antiarrhythmics, without requiring prior drug failure.

Anticoagulation after ablation

Successful ablation does not by itself change the indication for anticoagulation in the guidelines, which follow the risk score. Two recent trials, ALONE-AF and OCEAN, tested stopping the anticoagulant after ablation and showed no increase in stroke with stopping (Dowell and Patel, 2026). The authors suggest that stopping may be considered in those with little or no evidence of recurrence, with risk-guided monitoring, and this topic still awaits guideline updates (Dowell and Patel, 2026).

AF with heart failure with reduced ejection fraction

ACC/AHA gives class 1 and level A to catheter ablation in appropriate patients with AF and heart failure with reduced ejection fraction, already on optimized medical therapy and with a reasonable expectation of benefit, to improve symptoms, quality of life, ventricular function and cardiovascular outcomes (Joglar et al., 2024). It further recommends suspecting arrhythmia-induced cardiomyopathy in those who present with a new diagnosis of both conditions, with an early and aggressive rhythm approach (Joglar et al., 2024).

A 2026 meta-analysis of 13 studies and 131,492 participants with ejection fraction of 50 in every 100 or less associated ablation with lower overall mortality, with a relative risk of 0.61, and with a gain in ejection fraction compared with medical therapy (Malik et al., 2026). The same analysis showed that cardiovascular hospitalizations were statistically equal, and most participants came from observational studies (Malik et al., 2026).

ESC recommends sodium-glucose cotransporter 2 inhibitors in patients with heart failure and AF, whatever the ejection fraction, with class I and level A (Van Gelder et al., 2024). The blog covered heart failure with preserved ejection fraction in finerenone in heart failure with preserved ejection fraction.

AF after stroke

Among people with ischemic stroke, a 2026 review estimates that 20 to 30 in every 100 have underlying AF (Wasser et al., 2026). The authors note that prolonged monitoring after stroke increases AF detection, but that it is still uncertain whether this reduces recurrence (Wasser et al., 2026). This uncertainty is the same one that appears in device-detected AF: finding more AF is different from preventing more strokes. When the doubt is diagnostic, the text about diagnostic error in primary care helps to think about where the reasoning fails.

How NEXO helps train this reasoning

In NEXO, the case of the day is fictional, with a real clinical presentation and up to six progressive clues, and you propose the diagnosis by searching for the ICD-10 code.

The limits on using the game are described on the Economy page of NEXO. If you research the teaching of clinical reasoning, the page for Researchers explains how to get in touch. Professionals who write or review cases find information on Curator and on Creators, and institutions can learn about the proposal on the partnerships page. Choosing the right ICD-10 code completes this training.

Frequently asked questions

What changes with the removal of sex from the score?

The 2024 ESC moved to CHA2DS2-VA, which drops female sex and simplifies the calculation for men and women (Van Gelder et al., 2024). The ACC/AHA keeps CHA2DS2-VASc with different thresholds for men and women (Joglar et al., 2024).

With a score of 1, should I anticoagulate?

Both guidelines leave the decision open. The ESC says anticoagulation should be considered with a CHA2DS2-VA of 1, class IIa (Van Gelder et al., 2024). The ACC/AHA considers anticoagulation reasonable with an annual risk of 1 to 2 per 100 and suggests using risk modifiers in shared decision-making (Joglar et al., 2024).

Is a smartwatch alert enough to diagnose AF?

The ESC requires electrocardiographic confirmation and accepts a single-lead tracing of 30 seconds or more when the rhythm is analyzable (Van Gelder et al., 2024). An alert without a tracing is a reason to investigate.

Does a short pacemaker episode require an anticoagulant?

According to the ACC/AHA, with an episode under 5 minutes and no other indication, the recommendation is against anticoagulation (Joglar et al., 2024). Between 5 minutes and 24 hours, it calls for a score of 3 or more and shared decision-making (Joglar et al., 2024).

Can ablation come before antiarrhythmics?

In symptomatic paroxysmal AF, both guidelines give ablation as first-line therapy class 1, level A in selected patients (Joglar et al., 2024; Van Gelder et al., 2024).

Is aspirin an option for those afraid of anticoagulants?

The ACC/AHA classifies using aspirin, with or without clopidogrel, as harmful when used as an alternative to anticoagulation in those who have an indication for it (Joglar et al., 2024). For those with a definitive contraindication to anticoagulants, left atrial appendage occlusion is the alternative discussed, with class 2a (Joglar et al., 2024).

Take-home messages

• The 2023 ACC/AHA organizes AF into stages of progressive disease; the 2024 ESC organizes care in AF-CARE, with comorbidities first.
• The ESC uses CHA2DS2-VA, without sex; the ACC/AHA uses CHA2DS2-VASc with sex-based thresholds and accepts other validated scores.
• Direct oral anticoagulants are preferred over warfarin, except in moderate to severe rheumatic mitral stenosis and in mechanical prostheses, and aspirin is out of stroke prevention in AF.
• In device-detected AF, the ACC/AHA grades management by episode duration and the ESC speaks of considering anticoagulation in high risk; the trials show fewer strokes and more bleeding.
• Weight, alcohol and exercise have class 1 recommendations in both guidelines.
• Catheter ablation is first-line, with class 1 and level A, in symptomatic paroxysmal AF in selected patients, and the ACC/AHA also recommends it in reduced ejection fraction.

References

1. Joglar JA, Chung MK, Armbruster AL, et al. 2023 ACC/AHA/ACCP/HRS Guideline for the Diagnosis and Management of Atrial Fibrillation: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2024;149(1):e1-e156. PMID 38033089. DOI 10.1161/CIR.0000000000001193.
2. Van Gelder IC, Rienstra M, Bunting KV, et al. 2024 ESC Guidelines for the management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS). Eur Heart J. 2024;45(36):3314-3414. PMID 39210723. DOI 10.1093/eurheartj/ehae176.
3. Zubrzycki N, Hyun K, Teber E, et al. Comparative performance of stroke risk scores in patients with atrial fibrillation with low stroke risk. Open Heart. 2026;13(2):e004232. PMID 42469001. DOI 10.1136/openhrt-2026-004232.
4. Scott J. Device-detected Atrial Fibrillation: How Much is Too Much? Arrhythm Electrophysiol Rev. 2026;15:e20. PMID 42170358. DOI 10.15420/aer.2026.04.
5. Liu H, Li Z, Chen X, Zhou J. Discrepant Outcomes of Direct Oral Anticoagulant Trials in Device-Detected Atrial Fibrillation after Stroke: Insights from NOAH-AFNET 6 and ARTESIA. Clin Appl Thromb Hemost. 2026;32:10760296261438235. PMID 41877615. DOI 10.1177/10760296261438235.
6. Winstén AK, Langén V, Airaksinen KEJ, Teppo K. Cost-Effectiveness of Anticoagulation Treatment for Subclinical Device-Detected Atrial Fibrillation. JAMA Netw Open. 2026;9(6):e2617213. PMID 42258212. DOI 10.1001/jamanetworkopen.2026.17213.
7. Fu S, Li S, Zhang X, Yu R, Sun L. The Effect of Weight Loss and Metabolic Interventions on Recurrence After Atrial Fibrillation Ablation. J Clin Med. 2026;15(16):6493. PMID 42652896. DOI 10.3390/jcm15166493.
8. Malik AA, Saeed N, Oparaugo G, et al. Catheter Ablation Versus Medical Therapy in Atrial Fibrillation With Heart Failure With Reduced or Mildly Reduced Ejection Fraction: A GRADE-Assessed Systematic Review and Meta-Analysis. J Arrhythm. 2026;42(5):e70464. PMID 42769399. DOI 10.1002/joa3.70464.
9. Dowell JS, Patel N. Discontinuing Oral Anticoagulation after Atrial Fibrillation Ablation: Time for a Paradigm Shift? US Cardiol. 2026;20:e17. PMID 42775132. DOI 10.15420/usc.2025.23.
10. Wasser K, Bochert D, Leyhe JR, et al. Atrial fibrillation in stroke patients: facts, numbers and current challenges. Neurol Res Pract. 2026;8(1):85. PMID 42897704. DOI 10.1186/s42466-026-00538-9.