Atrial fibrillation

Atrial fibrillation (AF) is an irregular, often rapid heart rhythm in which the two upper chambers of the heart, the atria, stop contracting in a coordinated way and instead quiver with disorganized electrical signals. It is the most common sustained cardiac arrhythmia in clinical practice and a major concern in cardiovascular disease and public health. Its importance lies less in the palpitations it causes than in its consequences: because blood can pool and clot in the atria, AF substantially raises the risk of ischaemic stroke and systemic thromboembolism, and it is associated with heart failure, cognitive decline and higher mortality. Its prevalence rises steeply with age and with comorbidities such as hypertension and obesity, which makes it a growing burden for health systems.

This entry explains AF from its electrophysiological basis to its clinical management. It begins with the definition and the temporal patterns used for classification, then describes the mechanisms of triggers and self-sustaining circuits, and the causes and genetic background that promote them. It then covers symptoms, diagnosis and monitoring, followed by the complications that make the condition clinically important. The management sections address anticoagulation for stroke prevention, rate and rhythm control including catheter ablation, and the comorbidity-focused approach that current guidelines emphasize. The entry closes with special clinical situations, epidemiology and the history of the field, and the questions that remain open. Throughout, the entry relies on the medical research base, including randomized controlled trials and evidence-based guidelines, and it distinguishes established knowledge from areas still under study.

Definition and classification

Basic definition and ECG features

Atrial fibrillation (AF) is a supraventricular arrhythmia in which the atria lose coordinated electrical activity. Instead of one impulse from the sinoatrial node spreading across the atria in an orderly sequence, many disorganized wavelets of action potential activity circulate through atrial tissue, and the atria stop contracting as a coordinated unit. Atrial activation rates are commonly estimated at 400 to 600 beats per minute, far faster than the atrioventricular (AV) node can conduct. The AV node passes only some of these impulses to the ventricles, and because its refractoriness varies from beat to beat, the ventricular rhythm is irregular.

On a 12-lead electrocardiogram (ECG), the core findings are the absence of distinct P waves, their replacement by fibrillatory (f) waves, and R-R intervals with no repeating pattern, usually described as "irregularly irregular." The f-waves may be coarse or fine, and occasionally no atrial activity can be seen. Interpretation has pitfalls. Fine f-waves can be mistaken for baseline noise, and muscle activity can imitate them. A rhythm that looks irregular may, on closer inspection, show grouped or regular R-R patterns, which points toward another diagnosis such as atrial flutter with variable AV conduction or multifocal atrial tachycardia. The pulse can suggest AF at the bedside, but the ECG is needed to confirm it.

AF is distinguished from atrial flutter, which shows an organized sawtooth pattern of flutter waves with a much more regular atrial rate. Atrial tachycardia arises from a focal or re-entrant source and keeps organized atrial activation. AV nodal re-entrant tachycardia is typically regular. Pre-excited AF in Wolff-Parkinson-White syndrome produces rapid, irregular, wide QRS complexes that can resemble ventricular arrhythmias.

Temporal patterns

Guidelines classify AF by how long episodes last and whether they end without treatment:

• First-diagnosed AF: the first documented episode, whatever its duration or symptoms. It says nothing about the future course.
• Paroxysmal AF: recurrent episodes that stop on their own, classically within seven days, although some end only after intervention.
• Persistent AF: continuous AF lasting more than seven days, or episodes that require pharmacological or electrical cardioversion.
• Long-standing persistent AF: continuous AF lasting more than 12 months.
• Permanent AF: AF that the patient and clinician have jointly accepted, with no further attempt at rhythm control. It is a treatment decision rather than a fixed physiological state, which is why the term is avoided once ablation is under consideration.

These categories guide treatment because episode duration is a rough marker of atrial remodeling and of the likelihood that AF will stop spontaneously. They are imperfect, however. Short-term monitoring often misclassifies paroxysmal and persistent AF, and when both patterns occur in one patient, the predominant pattern is used. The 2023 ACC/AHA/ACCP/HRS guideline keeps the duration categories but adds a staged model: stage 1 (at risk), stage 2 (pre-AF, with structural or electrical predisposition), stage 3A to 3D (documented AF, with 3D denoting successful ablation), and stage 4 (permanent AF). It also encourages viewing AF as a continuum rather than a set of fixed classes.

Valvular and nonvalvular terminology

For roughly two decades, guidelines divided AF into "valvular" and "nonvalvular" forms, mainly to determine anticoagulant eligibility. The 2001 ACC/AHA/ESC guidelines defined nonvalvular AF as AF without rheumatic mitral disease or a prosthetic valve. The 2014 ACC/AHA/HRS guideline excluded rheumatic mitral stenosis, mechanical or bioprosthetic valves, and mitral repair. The 2012 ESC update instead defined valvular AF as AF linked to rheumatic disease, mainly mitral stenosis, or to prosthetic valves. Later guidance abandoned the nonvalvular label. The 2016 ESC guideline kept only moderate-to-severe rheumatic mitral stenosis and mechanical valves as exclusions for direct oral anticoagulants (DOACs), and the 2019 AHA update recommended dropping the term. Current usage therefore reserves "valvular AF" for moderate-to-severe mitral stenosis or a mechanical heart valve, and it is not a synonym for any valve disease. Critics have argued that the older definitions were confusing and inconsistently applied across trials.

Subclinical AF

Implanted devices such as pacemakers, defibrillators and loop recorders, along with wearable sensors, can detect atrial high-rate episodes or irregular rhythms in people who have no symptoms. Detection thresholds differ between devices and studies, and algorithms can flag artefacts or frequent atrial ectopic beats, so device alerts usually need clinical or ECG confirmation. The debate centres on classification. The duration categories above assume documented, usually symptomatic, episodes, and it is unclear whether brief silent episodes belong to the same disease or represent a separate, earlier stage. Some researchers regard them as a marker of underlying atrial disease, while others regard them as AF of low burden. Whether they carry stroke risk comparable to clinical AF is discussed in later sections.

Electrophysiological mechanisms

Triggers

Paroxysmal AF is usually initiated by rapid, repetitive firing from a focus outside the normal sinus-node pathway. The key observation came in 1998, when Haïssaguerre and colleagues documented spontaneous AF episodes in 37 patients and found that most initiating ectopic beats arose in the pulmonary veins (PVs), which accounted for about 94% of the triggering foci in that series. This location makes anatomical sense: the PVs are sheathed by extensions of left atrial muscle, a mixture of working cardiomyocytes and pacemaker-like cells, and these sleeves can generate fast, abnormal impulses that spread into the atrium. The cellular basis is still debated. Proposed mechanisms include enhanced automaticity and triggered activity, the latter arising from early or delayed afterdepolarizations driven by calcium handling abnormalities in the cell. Mechanical stretch of the venous sleeves may also favour triggered firing.

The substrate that sustains AF

A trigger can start an episode, but sustained AF depends on a substrate in which activation circulates. The classical multiple-wavelet hypothesis, proposed by Moe in 1962, describes many chaotic wavefronts that split, collide and regenerate. Later work emphasised rotors, spiral waves that turn around a core and act as sources of fibrillatory activity, and a double-layer model in which the epicardial and endocardial layers become electrically dissociated in advanced disease. The relative importance of these mechanisms remains controversial, particularly in persistent AF. Re-entry is favoured when conduction slows, because reduced coupling between cells lengthens the time needed for a wavefront to pass. A circuit can persist when the wavelength of the impulse, roughly the product of conduction velocity and refractory period, is shorter than the atrial path that the wave must travel, allowing several re-entrant wavelets to coexist. In general terms, paroxysmal AF is dominated by focal triggers and drivers near the PVs, whereas persistent and permanent AF increasingly reflect re-entrant substrates that are first functional and later structural.

Electrical and structural remodeling

AF alters the atrium that sustains it. Electrical remodeling, which develops within the first few days of an episode, shortens the action potential and the effective refractory period, largely through changes in the expression and function of ion channels. Because shorter wavelengths let more wavelets coexist, the rhythm becomes more stable and more easily maintained. Calcium handling is also disturbed: calcium overload, altered release from the sarcoplasmic reticulum and changes in calcium-dependent currents increase the likelihood of triggered beats. Over months and years, structural remodeling adds cellular hypertrophy, loss of intercellular contacts and interstitial fibrosis. Fibroblasts proliferate and transform into myofibroblasts, which disrupt the coupling between myocytes and hinder impulse propagation. Electrical and contractile changes are generally regarded as potentially reversible, whereas structural changes are more persistent. The interaction between these processes underlies the principle that AF tends to promote further AF: each episode leaves the atrium more prone to the next.

Autonomic and ionic modulation

Autonomic tone shapes both the triggers and the substrate. Acetylcholine released by the vagus nerve activates M2 muscarinic receptors and opens the acetylcholine-activated potassium current, I_KACh, which shortens action potentials and refractory periods unevenly across the atrium. This regional heterogeneity favours re-entry, and vagal stimulation can precipitate atrial flutter or fibrillation in experimental models. In humans, AF onset shows a diurnal peak at night, and heart rate variability often suggests a rise in vagal drive shortly before episodes. Sympathetic activity acts mainly through catecholamines, which shorten the atrial refractory period slightly, increase L-type calcium current and enlarge calcium transients, thereby promoting triggered activity. Simultaneous vagal and sympathetic activation appears particularly arrhythmogenic, since the abbreviated action potential coexists with a prolonged calcium transient that can produce late phase-3 early afterdepolarizations. Adrenergic signalling also affects the slow delayed-rectifier potassium current I_Ks and the inward rectifier I_K1. Ion-channel remodeling adds further variation: atrial myocytes from patients with chronic AF show reduced I_KACh channel expression, which may partly counteract action-potential shortening. Because autonomic measurements differ considerably between studies, the precise contribution of these pathways in an individual patient remains uncertain.

Causes, risk factors and genetics

Acquired risk factors

Age is the most consistent acquired determinant. The likelihood of AF rises with each decade of life and climbs most steeply after about 65 years of age Your risk for AFib goes up as you get older, and especially after age 65. Alongside age, AF shares many of its risk factors with other cardiovascular diseases. Coronary artery disease, valve disease and heart failure are all linked to a higher risk, and so are high blood pressure, obesity, hyperthyroidism, chronic kidney disease and obstructive sleep apnoea These include coronary artery disease, valve disease and heart failure. ... These include high blood pressure, overweight, obesity, hyperthyroidism, chronic kidney disease and obstructive sleep apnea.

The size of each effect varies. In one pooled estimate, hypertension carried a relative risk of about 1.5, and some analyses attribute more than a fifth of AF cases to it Hypertension 1.50 (1.42–1.58). ... Likewise, over 20 % percent of AF cases can be attributed to hypertension. Body weight acts in a graded fashion: each 5 kg/m² rise in body mass index has been associated with roughly 19–29% additional risk Any 5 kg/m^2 increase in BMI confers 19–29 % additional risk for developing AF. Diabetes and diabetes mellitus are also consistently reported in cohort studies A number of cohort studies have shown clinical risk factors for the development of AF in the general population, including advanced age, hypertension, diabetes and heart failure.

Alcohol shows a dose-related pattern. Each additional daily drink has been associated with a modest increase in risk, and intake above roughly three drinks a day is linked to a clearly higher risk Alcohol per 1 drink/day 1.08 (1.06–1.10). ... AF risk is increased when consuming more than 3 alcoholic drinks per day. Binge drinking is particularly relevant. Thyroid dysfunction, specifically an overactive thyroid, is a recognised cause, through the effects of excess thyroid hormone on the heart.

Sleep-disordered breathing deserves separate mention because it is common among patients with AF. Estimates of its prevalence in AF populations range widely, from about 21% to 74% Obstructive sleep apnoea (OSA) is prevalent in 21–74 % of AF patients. In one analysis, OSA carried a hazard ratio of about 2.18 for AF in unadjusted models. Its clinical relevance is therefore linked to the physiological effects described in obstructive sleep apnea.

How these conditions create a substrate

Most of these risk factors act through a common route. Obstructive sleep apnoea, for example, repeatedly drops oxygen levels and raises carbon dioxide, causes swings in intrathoracic pressure, and increases sympathetic activity and inflammation. Over time these changes contribute to hypertension, diastolic dysfunction, left atrial enlargement and atrial fibrosis. Similar processes follow from hypertension, valve disease and heart failure, which raise pressure or volume in the atria. The resulting stretch of atrial muscle, together with inflammation and the deposition of collagen in the atrial wall, produces the structural and electrical fibrosis that allows AF to start and persist. Since many of these factors coexist, their effects are cumulative rather than separate These lifestyle-dependent factors rarely exist in isolation, but rather exist together, exerting a complex influence on the development of AF.

Several of these factors are modifiable. Several risk factors for AF may be reversible, including obesity, hypertension, diabetes mellitus (DM), sleep apnea, coronary artery disease, heart failure (HF), and alcohol or tobacco use. Whether reducing them lowers AF incidence in a given patient, however, depends on the setting and is discussed in the comorbidity section.

Heritable susceptibility

Family studies indicate that AF has a genetic component. Heritability estimates vary by method and population. One review gives an upper estimate of about 22% in people of European ancestry There is a significant heritability to AF, estimated to be up to 22% in people of European ancestry., whereas a large Japanese genome-wide study estimated a liability-scale figure of about 12% the liability-scale h2 was estimated at 11.7% (s.e.m. 2.6%) using linkage disequilibrium (LD)-score regression.

The first AF genome-wide association study (GWAS) was published in 2007 by an Icelandic group, and its strongest signal lay on chromosome 4q25, in an intergenic region with the nearest gene, PITX2, about 150 kilobases away The first large GWAS for AF, including 550 AF cases, was performed on Iceland in 2007 (20), identifying a genetic locus associated with AF on chromosome 4q25. The most significant variant was located in an intergenic region, with the nearest gene, PITX2, 150 kilobases (kb) away.. Later genome-wide association studies have greatly expanded the list. A multi-ethnic meta-analysis reported 97 susceptibility loci In total, we identified 97 loci significantly associated with AF, including 67 that were novel in a combined-ancestry analysis, and 3 that were novel in a European-specific analysis., and a later meta-analysis of more than 180,000 cases reported over 350 loci, with candidate genes at many of them relating to muscle contractility, cardiac development and cell–cell communication This large-scale meta-analysis of genome-wide association studies increased the power to detect single-nucleotide variant associations and found more than 350 AF-associated genetic loci. We identified candidate genes related to muscle contractility, cardiac muscle development and cell–cell communication at 139 loci.. A cross-ancestry analysis highlighted an immune-related gene, IL6R, as a putative causal candidate Transcriptome-wide association analysis identified IL6R as a putative causal gene, suggesting the involvement of immune responses..

These common variants explain only part of the risk, and each has a small individual effect. Their presence signals susceptibility rather than proving causation Although common variants provide an estimate of disease susceptibility, they are not necessarily causative..

Rare, high-impact variants form the second layer. Family-based and sequencing studies have implicated mutations in genes encoding ion channels and related proteins in familial or early-onset AF linkage analysis and whole exome sequencing have implicated rare mutations encoding ion channels, signaling molecules, transcription factors, and, recently, structural proteins linked with familial or early-onset AF.. One early example was a gene on chromosome 11p15.5 encoding the KCNQ1 potassium channel. The S140G variant in that family produced a gain of function, which contrasts with the loss-of-function effects seen in long QT syndrome Functional analysis of the S140G mutant revealed a gain-of-function effect on the KCNQ1/KCNE1 and the KCNQ1/KCNE2 currents, which contrasts with the dominant negative or loss-of-function effects of the KCNQ1 mutations previously identified in patients with long QT syndrome.. This matters because both gain and loss of function in the same current have been linked to AF, and gain-of-function changes tend to shorten the atrial action potential, which favours re-entry Both gain of function and loss of function mutations in I_Ks have been associated with AF. Mutations in KCNQ1 and KCNE2 increase the current I_Ks, which is predicted to shorten the action potential (dashed line) in cardiac myocytes and render atrial myocytes susceptible to reentrant arrhythmias.. A sodium channel variant, SCN5A M1875T, was identified in a three-generation family in which persistent AF began as early as age 27 A missense SCN5A point mutation, Met1875Thr (M1875T), located in the C-terminus of the channel protein, was linked to autosomal dominant familial AF that spanned three generations. Atrial ectopy was evident in mutation carriers in adolescence, and persistent AF occurred as early as 27 years of age.. In a screen of 30 patients with juvenile-onset AF, three variants (about 10%) were identified, in SCN5A, KCNJ2 and KCNQ1 In 30 consecutive juvenile-onset AF patients (onset age o 50 years), we screened AF-related genes (KCNQ1, KCNH2, KCNE1-3, KCNE5, KCNJ2, SCN5A). ... three mutations (10%): SCN5A-M1875T, KCNJ2-M301K, and KCNQ1-G229D..

Endurance exercise and other lifestyle exposures

Physical activity is a paradox in AF research. Low cardiorespiratory fitness is itself listed among risk factors Poor cardiovascular endurance, yet many studies of long-term endurance athletes have reported a higher frequency of AF than in sedentary controls, sometimes with a pattern suggesting that cumulative training volume matters. Proposed explanations include atrial dilatation, changes in autonomic tone and inflammation. However, the evidence is complicated by confounding (age, sex, prior cardiac disease and drug use), by detection bias from intensive screening in athletes, and by possible reverse causation, in which early atrial changes may reduce exercise tolerance or drive increased training. Current data therefore support an association with heavy, prolonged endurance training in some populations, but they do not establish that exercise itself causes AF, and moderate activity is generally linked to lower risk through its effects on weight, blood pressure and fitness.

Other lifestyle exposures are also tangled together. Modifiable factors include obesity, hypertension, hypercholesterolemia, diabetes mellitus, smoking, chronic obstructive pulmonary disease, alcohol consumption, exercise, sedentary lifestyle and obstructive sleep apnoea. Disentangling their individual causal contributions remains a major challenge for observational research.

Symptoms, diagnosis and monitoring

Symptoms and their variability

Atrial fibrillation is often silent. Estimates suggest that roughly one third of people with AF report no symptoms, while up to a quarter have severe complaints. The most frequent complaints are palpitations, which patients describe as fluttering, skipping or pounding, together with fatigue, breathlessness on exertion, dizziness, near-fainting and chest discomfort. Symptoms tend to be more pronounced when the ventricular rate is fast, and they are more often reported by people with heart disease. Symptom burden does not track closely with the amount of AF, and psychological distress, particularly anxiety and depression, is common among symptomatic patients. Older adults frequently present with fatigue or breathlessness rather than palpitations. Asymptomatic AF is more common in men, in older people and in permanent AF, and it is often discovered incidentally on a routine examination or electrocardiogram, or when a patient presents with a stroke or transient ischaemic attack.

Electrocardiographic diagnosis and ambulatory monitoring

The 12-lead electrocardiogram (see electrocardiography) is the reference test. AF is identified by the absence of organised atrial P waves and by irregularly irregular R–R intervals. A resting tracing is diagnostic when AF happens to be present during the recording, but paroxysmal AF is episodic and easily missed, which is why longer recording is often needed. The Holter monitor is an external device that records continuously and is usually worn for up to 48 hours; it is unsuitable for symptoms that occur less often than about every two days. Patch-type recorders adhered to the chest extend monitoring to days or weeks and are useful when symptoms are infrequent, and continuous recordings of around one to two weeks are often preferred when feasible. Implantable loop recorders are small subcutaneous devices that can monitor for months to years. Some models can trigger recording automatically, which allows them to detect asymptomatic AF; they are used, for example, after a cryptogenic stroke, as tested in a randomised trial published in the Lancet in 2021. Monitoring selection depends mainly on how often symptoms occur and how much clinical risk the arrhythmia poses.

Initial evaluation

Once AF is documented, the initial work-up aims to classify the rhythm pattern, assess for structural heart disease and identify reversible contributors. Transthoracic echocardiography is the standard test for estimating left atrial size, left ventricular function and valve disease, all of which affect stroke risk and treatment choice. Routine blood tests generally include thyroid function, because hyperthyroidism is a recognised cause of AF (see hormones), and kidney function, which is needed for anticoagulant dosing and bleeding-risk assessment (see chronic kidney disease). Electrolytes, glucose control and a full blood count are also commonly checked. Clinicians then screen for other drivers, such as excess alcohol, obstructive sleep apnoea, acute infection, lung disease, and medications or stimulants that may provoke the rhythm. Ischaemia is considered when chest pain is present.

Smartphone applications and smartwatches

Consumer devices detect AF either by photoplethysmography, which infers irregular pulse from light signals at the wrist or fingertip, or, in some newer watches, by a single-lead electrocardiogram recorded through electrodes on the case. The Apple Heart Study, published in the New England Journal of Medicine in 2019, found that about one third of participants who received an irregular-pulse notification had AF on subsequent ECG patch recordings, and that the notifications were concordant with AF in 84% of cases. Fitbit's large heart study reported irregular-rhythm notifications in about 1% of participants, and among those with analysable patch data, AF was confirmed in roughly 32%, though the authors' headline positive predictive value came from a much smaller subgroup and has been criticised for that selection. Smartphone camera applications show high sensitivity and specificity in diagnostic studies, but a meta-analysis estimated that their positive predictive value in an asymptomatic older population with hypertension would be only around 20% to 40%, while the negative predictive value approaches 100%. Because AF prevalence in general populations is low, even accurate tests produce many false positives; a randomised study in older stroke survivors reported a positive predictive value near 43% using a system designed to limit false alarms. Researchers are therefore examining how tools perform across real-world populations, how inconclusive or poor-quality recordings are handled, and whether false alerts cause anxiety or unnecessary testing. Whether population screening reduces stroke has not been established, and the US Preventive Services Task Force has concluded that the evidence is insufficient to assess screening of asymptomatic adults older than 50.

Complications and prognosis

Thrombus formation and stroke

Most thrombi that form in AF arise in the left atrial appendage (LAA), a small pouch projecting from the left atrium. In AF the atrial muscle no longer contracts in a coordinated way, so the LAA's sluggish, low-flow interior allows blood to pool and platelets and clotting factors to aggregate. A fragment that breaks away can lodge in a cerebral artery. Because emboli from the left heart usually reach large brain vessels, AF-related ischaemic strokes are commonly described as more disabling and more often fatal than strokes of other origin, although severity also depends on age, comorbidity and how quickly treatment starts. The LAA is therefore the principal target of both anticoagulation and mechanical occlusion, which are discussed in later sections.

Thrombus detection itself is a marker of disease severity as well as of embolic risk. In a 2024 prospective study of 500 anticoagulated patients admitted for electrical cardioversion, transoesophageal echocardiography showed a thrombus in 13%. Stroke and systemic embolism did not differ significantly between patients with and without thrombus over a median follow-up of more than five years, whereas cardiovascular death and heart-failure admissions were more frequent in the thrombus group. The authors interpreted thrombus as a sign of a more advanced atrial and ventricular disease process, though this single-centre observational study cannot establish causation.

Tachycardia-mediated cardiomyopathy

Prolonged rapid heart rates can by themselves weaken the left ventricle. This condition, tachycardia-mediated cardiomyopathy, is a dilated cardiomyopathy without ventricular hypertrophy that is at least partly reversible when the arrhythmia is controlled. Atrial fibrillation is its most common cause. Proposed mechanisms include loss of atrial contribution to filling, an irregular rhythm, high ventricular rates, neurohumoral activation, intracellular calcium overload, mitochondrial injury and myocardial ischaemia. Its onset is variable: it may appear within a week of the tachycardia or take years to declare itself. Once the rate or rhythm is brought under control, left ventricular function often recovers within weeks to months, which makes recognition important, because the condition is easily mistaken for idiopathic cardiomyopathy.

Bidirectional relationship with heart failure

AF and heart failure frequently coexist and share many risk factors, including hypertension, ageing, valve disease, obesity and coronary disease. Heart failure raises filling pressures and stretches the atria, creating a substrate for AF, while AF in turn impairs ventricular function through the mechanisms above and can worsen congestion in a patient whose ventricle was already compromised. Three components are often distinguished in AF-induced cardiomyopathy: tachycardia-related dysfunction, the harm caused by irregular rhythm, and AF-induced atrial myopathy. A common clinical sequence is a patient with subclinical cardiomyopathy who develops AF with a rapid ventricular response, which may become permanent and then precipitate overt heart failure.

Cognitive impairment, quality of life, hospitalization and mortality

Observational data link AF with several adverse outcomes beyond stroke. A 2013 meta-analysis of 21 studies reported a relative risk of about 2.7 for cognitive impairment among patients with prior stroke and about 1.4 in broader populations, with or without stroke. A later synthesis of 15 general-population studies involving about 2.8 million people found a hazard ratio of 1.39 for cognitive impairment. Proposed mechanisms include silent cerebral infarcts, cerebral small-vessel disease and chronic cerebral hypoperfusion during arrhythmia, but the evidence for these pathways is incomplete, and the pooled estimates show considerable heterogeneity. Patients with AF also report reduced quality of life, exercise intolerance and more frequent hospital admissions, and cohort studies associate AF with higher all-cause and cardiovascular mortality. Some hospitalizations, however, reflect procedures such as cardioversion or ablation rather than the natural course of the disease.

The central interpretive problem is that AF rarely occurs in isolation. Its patients are typically older and carry hypertension, diabetes, kidney disease, heart failure or previous stroke, all of which independently affect cognition, quality of life and survival. Residual confounding is therefore unavoidable in observational work, and reverse causation is also possible, since acute or chronic illness can itself provoke AF. A detected arrhythmia may thus be a marker of vulnerability rather than a cause of harm. Whether treating AF itself changes these outcomes is a separate question that depends on randomized evidence rather than association alone.

Variation with pattern, rate and comorbidity

The burden of complications is not uniform across AF patients. Persistent and permanent AF are associated with more structural remodelling of the left atrium and, by extension, a higher thrombotic risk than brief paroxysmal episodes, although paroxysmal AF still carries meaningful stroke risk, so anticoagulation decisions rely on clinical risk scores rather than pattern alone. The ventricular rate also matters. One proposed model describes a J-shaped relationship between rate and outcome, with a best-observed range of roughly 60 to 80 beats per minute, and with both persistently fast and very slow rates linked to worse function and prognosis; this model remains a hypothesis rather than an established rule. Finally, the presence of other cardiovascular disease, particularly reduced ejection fraction, left atrial enlargement and impaired kidney function, strongly shapes prognosis. In the Polish cohort, greater left atrial diameter, lower ejection fraction and lower estimated glomerular filtration rate were all predictors of poor outcome, which suggests that the surrounding disease, rather than the rhythm alone, drives much of the risk.

Stroke prevention and anticoagulation

Risk stratification with CHA₂DS₂-VA

Decisions about oral anticoagulation rest on a points-based estimate of thromboembolic risk. The CHA₂DS₂-VA score awards points for heart failure, hypertension, diabetes, vascular disease and age, with extra weight for age 75 and above, and for prior stroke or transient ischaemic attack, which counts double. CHA₂DS₂-VA=Congestive heart failure, Hypertension, Age ≥75 years (doubled), Diabetes, Stroke (doubled), Vascular disease, Age 65–74 years

The older CHA₂DS₂-VASc score also gave a point to female sex. The 2024 ESC/EACTS guideline dropped that term and adopted CHA₂DS₂-VA as the default tool. The rationale is that female sex appears to modify risk mainly in older patients who already qualify for treatment, rather than to add risk independently. The CHA2DS2-VA score excludes gender as a risk factor based on the premise that female sex does not independently contribute to anticoagulation choices but rather acts as a risk modifier, mainly in older individuals who would already meet criteria for anticoagulation. In a UK cohort of nearly 196,000 patients, removing the sex term left the score's ability to discriminate thromboembolic events essentially intact. Removing female sex from the CHA2DS2VASc score does not compromise its discrimination performance for thromboembolic events.

The change is not uncontested. Post-stroke cohorts have found that women who fall below the treatment threshold after recalibration may still have worse outcomes, which raises concern about under-treatment. A clinically meaningful subgroup of women now falls below treatment thresholds, raising concern for under-treatment. Under the current guideline, anticoagulation is now recommended based on a CHA₂DS₂-VA score of 2 or higher, and can be considered for a score of 1, irrespective of sex.

Choice of agent: DOACs and vitamin K antagonists

Direct oral anticoagulants (DOACs) are the default choice in most patients. Direct oral anticoagulants (DOACs) (apixaban, dabigatran, edoxaban, and rivaroxaban) are preferred over vitamin K antagonists, except in patients with mechanical heart valves and mitral stenosis. Patients who remain on a vitamin K antagonist such as warfarin should aim for an INR of 2.0 to 3.0. Those whose time in therapeutic range stays below 70% are candidates for switching to a DOAC. Target an INR of 2.0–3.0 for patients with AF prescribed VKAs for stroke prevention to ensure both safety and effectiveness, and switch eligible patients who do not maintain an adequate TTR (<70%) on a VKA to a DOAC to prevent thromboembolism.

Two groups are exceptions. Patients with rheumatic or moderate-to-severe mitral stenosis, and those with mechanical valves, are treated with a vitamin K antagonist. For mitral stenosis, the ACC/AHA guidance states that to prevent cardiovascular events in patients with a history of AF and rheumatic mitral stenosis or moderate or severe mitral stenosis, long-term anticoagulation with warfarin is recommended over DOACs. For mechanical valves, the evidence summarised in guideline tables favours warfarin over rivaroxaban. Guidance also warns against unnecessary dose reduction: a reduced dose of DOAC therapy is not recommended, unless patients meet DOAC-specific criteria, to prevent underdosing and avoidable thromboembolic events. Some guidance additionally suggests that patients over 75 with stable INR control may remain on warfarin, because switching has been linked to more major bleeding in that group. It is recommended that patients over 75 years and those on multiple medications with stable INR levels remain on VKAs rather than switching to DOACs, as the latter have been associated with increased major bleeding in this demographic.

Bleeding risk and antithrombotic combinations

Bleeding risk is assessed separately from stroke risk. Tools such as HAS-BLED estimates risk of major bleeding for patients on anticoagulation to assess risk-benefit in atrial fibrillation care. Its components include uncontrolled hypertension, abnormal renal or liver function, prior stroke, prior bleeding, labile INR, age over 65, concomitant antiplatelet drugs or NSAIDs, and heavy alcohol use. The practical emphasis is on correcting the modifiable items, such as blood pressure control, stopping unnecessary antiplatelet drugs or NSAIDs, reducing alcohol and stabilising INR values, rather than using a high score alone as a reason to withhold anticoagulation.

Antiplatelet drugs are not a substitute for anticoagulation. Avoid antiplatelet drugs alone (aspirin or aspirin combined with clopidogrel) for stroke prevention in AF, as they are not recommended. Adding antiplatelet therapy to an anticoagulant is also discouraged for stroke prevention. Adding antiplatelet treatment to oral anticoagulation is not recommended in AF patients for the goal of preventing ischemic stroke or thromboembolism. Anticoagulants and antiplatelet agents should not be combined, unless the patient has an acute vascular event or needs interim treatment for procedures.

Left atrial appendage occlusion

Left atrial appendage occlusion (LAAO) is considered for selected patients who cannot take long-term anticoagulation, or who have a stroke despite treatment. The rationale is that the appendage is a major site of thrombus formation in AF. The left atrial appendage accounts for approximately 90% of atrial thrombus location (it accounts for 60% among those with valvular AF, however). Occlusion can be performed percutaneously, using a catheter-delivered device that is placed in the appendage, or surgically, by clipping, suturing or stapling the appendage during another operation.

The strongest randomised evidence for the surgical approach comes from the LAAOS III trial. Patients undergoing cardiac surgery for other reasons were randomised to occlusion or no occlusion. The trial enrolled 4811 participants with AF and a CHA_2_DS_2_-VASc score of at least 2 who were undergoing cardiac surgery for other indications. Surgical occlusion of the left atrial appendage reduces the risk of ischaemic stroke or systemic embolism by 33% in patients with atrial fibrillation. Importantly, most participants continued anticoagulation, so the trial tested an add-on benefit rather than a replacement for anticoagulation. The additive benefit of surgical LAAO on top of blood thinners has now been proven. Evidence for percutaneous devices as alternatives to anticoagulation comes from earlier trials that compared devices with warfarin, and the 2024 guideline addresses the indications for both percutaneous and surgical exclusion as a distinct clinical question.

Rate and rhythm control

Rate control and rhythm control are the two strategies for managing symptomatic AF. Rate control accepts that AF persists and aims to slow the ventricular response so that the ventricles fill and pump adequately. Beta blockers and non-dihydropyridine calcium channel blockers such as diltiazem and verapamil slow conduction through the atrioventricular (AV) node, but the calcium channel blockers are avoided in reduced ejection fraction. Digoxin also slows AV conduction, mainly through vagal effects, and is used alone or as an add-on, especially in heart failure and in sedentary patients. Its narrow therapeutic window and renal clearance require careful dosing and monitoring. The RACE II trial (2010) supported lenient targets: a resting ventricular rate below 110 beats per minute was not inferior to stricter control for major outcomes.

Rhythm restoration is achieved either electrically or pharmacologically. Electrical cardioversion delivers a synchronized shock through electrodes on the chest, timed to the R wave to avoid provoking ventricular fibrillation, usually under brief sedation. Pharmacological cardioversion uses agents such as flecainide, propafenone, ibutilide, or intravenous amiodarone. Flecainide and propafenone are generally restricted to patients without structural or ischaemic heart disease. When AF has lasted more than 48 hours or its duration is unknown, guidelines call for at least three weeks of effective anticoagulation before cardioversion, or transoesophageal echocardiography to exclude left atrial appendage thrombus. Anticoagulation is then continued for at least four weeks afterwards, because thromboembolic risk stays elevated for some time after sinus rhythm returns.

Antiarrhythmic drugs act mainly on ion channels that govern the cardiac action potential. Sodium channel blockers (class I) include flecainide and propafenone; the CAST trial (1989) showed excess mortality with class IC agents after myocardial infarction. Potassium channel blockers (class III) include sotalol and dofetilide, which prolong the QT interval and can cause torsades de pointes; initiation is usually done in hospital with ECG monitoring and renal dose adjustment. Amiodarone is highly effective but has a long list of toxicities, including thyroid dysfunction (see thyroid hormones), pulmonary toxicity, hepatotoxicity, corneal deposits and photosensitivity. It inhibits CYP3A4 and P-glycoprotein, raising concentrations of warfarin (increasing the INR), digoxin and some statins. Dronedarone is less toxic but is contraindicated in decompensated heart failure and permanent AF. The choice of agent depends on the presence of structural heart disease, left ventricular function, renal and hepatic function, QT interval, age, concomitant medicines and patient preference.

Catheter ablation most commonly isolates the pulmonary veins by creating circumferential lesions at their ostia with radiofrequency or cryoballoon energy, guided by three-dimensional electroanatomic mapping. It is used for symptomatic paroxysmal AF and, in selected cases, persistent AF. Complications include pericardial effusion and tamponade, pulmonary vein stenosis, phrenic nerve injury (particularly with cryoballoons), vascular injury, stroke and, rarely but fatally, atrio-oesophageal fistula. The randomized CABANA trial (2019) found no statistically significant difference between ablation and drug therapy in its primary composite endpoint. The surgical maze procedure, developed by James Cox and colleagues, uses a pattern of incisions and lesions to compartmentalize the atria so that re-entrant wavelets cannot sustain AF. It is usually performed alongside other cardiac operations, such as mitral valve surgery, and modern versions often use energy sources instead of cut-and-sew techniques. For refractory, symptomatic AF with an uncontrolled ventricular rate, AV node ablation combined with permanent pacing produces a regular ventricular rhythm. This creates pacemaker dependence and does not treat the atrial arrhythmia itself, so anticoagulation must continue.

Comorbidity and lifestyle management

The AF-CARE framework

The 2024 ESC/EACTS guideline reorganised AF care around four linked pillars: comorbidity and risk factor management, avoiding stroke and thromboembolism, reducing symptoms by rate and rhythm control, and evaluation with dynamic reassessment. The letters C, A, R and E give the framework its name. Placing comorbidity first is the main structural change from the earlier ABC approach. The guideline's rationale is that AF therapies are most effective, and recurrence least likely, when associated health conditions have been effectively managed. Separate patient pathways were also written for first-diagnosed, paroxysmal, persistent and permanent AF, each beginning with the identification and treatment of risk factors.

Hypertension, obesity, diabetes and sleep apnoea

Obesity is the risk factor with the most consistent evidence behind weight-loss interventions. One review estimates that every 1 kg/m² increase in body mass index raises AF risk by 50%. In the LEGACY study, sustained weight loss of at least 10% was associated with a sixfold greater likelihood of freedom from atrial arrhythmias than less weight loss or weight gain, and weight fluctuation greater than 5% was independently associated with AF recurrence. Because these data are largely observational, they show association rather than proof that weight loss itself prevents recurrence.

For diabetes, the guideline's suggestion is that metformin or SGLT2 inhibitors should be considered for individuals needing pharmacological management of diabetes mellitus to prevent AF. Sleep apnoea has a mechanistic link to AF: intermittent hypoxia and sympathetic surges destabilise the atrial substrate, and CPAP therapy reduces remodeling and halves recurrence after cardioversion or ablation. Control of hypertension is listed among the same modifiable factors, and treatment of these comorbidities is generally considered part of the same effort, though the evidence for blood pressure and glycaemic targets specifically is less direct than for weight and alcohol.

Alcohol and physical activity

Alcohol acts in a dose-dependent manner, and abstinence reduces risk of recurrence. In a randomised trial of 140 Australian patients with symptomatic paroxysmal or persistent AF who drank regularly, those assigned to abstinence had a longer time to first recurrence and a 58% lower burden of AF during 6-month follow-up. In a separate cohort of 122 patients treated with ablation, one-year freedom from recurrence was markedly lower among heavy drinkers than among non-drinkers. Physical activity follows a U-shaped pattern. Moderate exercise is associated with lower risk, while endurance extremes and inactivity foster atrial stretch, fibrosis, and ectopic triggers. Practical targets used in lifestyle programmes include an initial 10% weight loss and a modest rise in fitness, progressing toward a body mass index of 25 kg/m² or less and about 200 minutes per week of moderate-intensity activity.

Education, shared decision-making and reassessment

Ongoing care depends on structured patient education, so that people understand why lifestyle change matters and what the realistic options are. Treatment choices, including whether to pursue rhythm control, catheter ablation or anticoagulation, are best reached through shared decision-making in which the clinician and patient weigh symptoms, preferences and the risks of each option. Informed consent is part of this process.

The E pillar of AF-CARE requires that stroke and bleeding risk be recalculated as patients age, as blood pressure, kidney function or heart failure status change, or as comorbidities are treated. A patient who was low risk at diagnosis may later require anticoagulation, and the reverse can also occur. Reassessment therefore extends across the whole course of AF rather than occurring once at diagnosis.

Special clinical situations

Valve disease

Guidelines still treat AF differently depending on valve status. Warfarin remains the recommended anticoagulant for patients with moderate or severe mitral stenosis or a mechanical heart valve, and for patients with mechanical heart valves, whether or not they have AF, long-term anticoagulation should be managed with VKAs to avoid valve thrombosis. These patients were excluded from the landmark AF trials, which is why the evidence base for direct oral anticoagulants (DOACs) does not extend to them. For mitral stenosis, the 2023 ACC/AHA/ACCP/HRS guideline favours warfarin over DOACs in patients with moderate or greater stenosis and a history of AF, regardless of the CHA₂DS₂-VASc score. Bioprosthetic valves occupy a middle ground: limited research supports using the CHA₂DS₂-VASc score to determine risk in patients with bioprosthetic valves. In mechanical mitral valve replacement, older reviews cite a higher warfarin target (INR 2.5 to 3.5) than the standard 2.0 to 3.0 range.

Device-detected subclinical AF and stroke despite anticoagulation

Subclinical AF found by pacemakers, defibrillators or implanted monitors has been studied in two large randomized trials. The NOAH-AFNET 6 trial tested edoxaban, and ARTESiA tested apixaban against aspirin. A combined analysis found that without vascular disease, outcomes were equal with and without anticoagulation. Across both trials, the pooled estimate was that anticoagulation reduces the relative risk of stroke by approximately 32% and increases the relative risk of major bleeding by approximately 62%. Because of this trade-off, the duration threshold for starting anticoagulation remains contested: in one European survey, 34% of responders still use a >24 h threshold, while 15% consider episodes as short as 6 min. Event rates also differ by episode length; progression to clinical AF is more common when episodes exceed 24 hours.

When ischaemic stroke occurs despite anticoagulation, the first step is to check adherence, dose, drug interactions and renal function, and to consider whether the event has a different mechanism, such as large-artery or small-vessel disease. Evidence for switching agents or adding antiplatelet therapy in this setting is limited, so decisions are generally individualized.

Renal failure, dialysis, extreme body weight and prior haemorrhage

Patients with end-stage renal disease present a particular problem. The 2019 US guidance advised that dabigatran, rivaroxaban, and edoxaban should be avoided in patients with AF and ESRD or on dialysis, because data were lacking to show that benefits outweighed harms. Warfarin and apixaban have therefore been the agents most often considered in dialysis, although the evidence there is weaker than in the general AF population. Drug dosing in severe renal impairment usually depends on creatinine clearance, and trial populations often excluded the most severely impaired patients.

Patients at the extremes of body weight are also underrepresented in trials. Those with very high body mass index or weight have limited trial data for DOACs, so dose selection and agent choice depend partly on expert consensus. A prior intracranial haemorrhage raises bleeding risk substantially; decisions about resuming anticoagulation, or considering left atrial appendage occlusion, are made case by case, weighing the site and cause of the bleed.

Hypertrophic cardiomyopathy and cardiac amyloidosis carry a higher thromboembolic risk than the general AF population, and European guidance generally favours anticoagulation in these patients even when conventional scores would suggest a lower risk. Amyloid-related cardiomyopathy in particular is linked to atrial thrombus formation, which makes stroke-risk scoring less reliable.

Postoperative AF and older or frail patients

Postoperative AF is the most common complication after cardiac surgery. It occurs in 20%-55% of patients who have cardiac surgery, and typically appears between postoperative days 2 and 4. Most episodes are self-limiting. Beta-blockade is the main preventive measure; all guidelines recommend β-blockade for prevention of AF after cardiac surgery. Anticoagulation decisions are more contentious. Canadian guidance has advised a weak recommendation against early anticoagulation, recommending that it be held during the first 72 hours postoperatively. The stroke benefit of anticoagulation is also less clear for these patients, and bleeding risk may outweigh the benefits of anticoagulation in patients with advanced age, uncontrolled hypertension, or a history of bleeding. Postoperative AF is associated with higher short-term mortality, but whether it causes that excess risk is uncertain.

After noncardiac surgery, management usually focuses on rate control and on deciding whether anticoagulation is needed once the acute period has passed. In older adults with frailty or several coexisting conditions, decisions are shaped by polypharmacy, falls, cognitive status and life expectancy, and the trials that underpin standard recommendations included few such patients. Shared decision-making and regular reassessment are therefore central to care.

Epidemiology, burden and history of study

Atrial fibrillation (AF) is strongly age-dependent. In one population-based analysis, prevalence roughly doubled with each advancing decade of age, from 0.5% at ages 50–59 to almost 9% at ages 80–89. Sex differences are also consistent. Men have a higher age-standardised incidence, about 1.49 times that of women in a 3.4-million-person study, yet because there were more women in the older age groups, the estimated total number of incident cases in 2017 was only 18% higher in men. Among people over 75, women make up most cases; one US analysis estimated about 60% of people with AF in that age range are women. In a 2019 US primary care cohort, the proportion with AF rose from 6.4% at ages 65–69 to 28.5% at 85 and older, and incidence showed the same age gradient.

Global totals have grown with the population. The Global Burden of Disease (GBD) 2010 study estimated 33.5 million people with AF worldwide in 2010. The GBD 2019 analysis counted 59.7 million cases of AF/AFL in 2019, and the number of deaths rose from about 117,000 in 1990 to about 315,000. Trends in age-standardised rates are less uniform. A 3.4-million-person population study found that AF incidence had increased by 30% in 2017 compared to 1998, and the authors suggested that factors other than an ageing population are contributing. GBD analyses instead attribute most of the rise in case numbers to demographic change: one reported that population growth contributed 52% to the rise in prevalent cases, surpassing aging, while another concluded that the burden has an overall increasing trend from 1990 to 2021, primarily driven by population growth and aging. Projections also differ on rates: one GBD-based model expects rising incidence rates, whereas another projects declines in incidence and rising mortality to 2050, so forecasts depend on the metric chosen. Data are uneven; corresponding data are limited in Africa, Asia, and South America, and some observers suspect that prevalence is underestimated even in high-income countries.

Early recognition and changing understanding of mechanism

Irregular pulses were described long before AF was understood. Jean Baptiste Sénac connected 'rebellious palpitation' and mitral stenosis in 1783, and in 1903 Hering described "pulsus irregularis perpetuus" as a distinct arrhythmia. The pulse was not linked to a rhythm mechanism until the electrocardiogram arrived. James Mackenzie described the loss of the 'a' wave in the venous pulse using his clinical polygraph in 1905, and Willem Einthoven published the first electrocardiogram showing atrial fibrillation in 1906. The clinical link was made soon afterwards. Thomas Lewis used the electrocardiogram in 1909 to show that 'delirium cordis', a clinical diagnosis of irregular heartbeat, resulted from atrial fibrillation, and Rothberger and Winterberg reached a similar conclusion in Vienna.

Mechanistic understanding then went through a debate that is still reflected in current thinking. Rothberger and Winterberg initially postulated that AF was the result of a single electrical focus. Lewis instead proposed that its basic perpetuating mechanism is circus movement of electrical impulse (re-entry). The multiple-wavelet picture was accepted for decades, but the idea of a single focus was soon abandoned, and in 1950 Prinzmetal and colleagues rejected Lewis' concept after filming the atrial surface. Later technology revived the debate: advanced technology has confirmed that the turbulent electrical activity seen by electrogram recordings of the atria may in some cases be explained by a single or a small number of rapidly spinning sources. Subsequent 20th-century contributions came from figures such as Karel Frederik Wenckebach, Gordon Moe, Bernhard Lown and Maurits Allessie.

Surgical treatment and the shift to catheter ablation

Surgery was the first attempt at a cure. The maze concept, in which the surgeon creates a pattern of atrial incisions and scar lines so that the sinus impulse is channelled and cannot sustain fibrillatory wavelets, was developed by James Cox and colleagues in the late 1980s. Early maze procedures were effective but required open-heart surgery, often alongside other cardiac operations, and were reserved for patients already undergoing cardiac surgery. Later modifications replaced cut-and-sew lines with energy-based lesions. The conceptual turning point for catheter-based treatment came in 1998, when Haissaguerre and colleagues reported that AF was often triggered by rapid firing from muscle sleeves extending into the pulmonary veins, and that isolating these veins could eliminate the trigger. Pulmonary vein isolation using radiofrequency energy delivered through catheters then became the dominant ablation strategy, and it expanded through the 2000s as mapping systems and energy sources improved.

Clinical trials and guideline change

Anticoagulation changed first. Randomised trials of warfarin in the 1990s showed large reductions in stroke compared with placebo or aspirin, and vitamin K antagonists became the default for decades, with dose control driven by INR monitoring. The early 2000s brought the AFFIRM trial, which compared rate and rhythm control strategies and found no survival advantage for rhythm control in that population, a result that shaped subsequent debate. From 2009 onwards, phase 3 trials of direct oral anticoagulants (RE-LY with dabigatran, ROCKET AF with rivaroxaban, ARISTOTLE with apixaban, and ENGAGE AF-TIMI 48 with edoxaban) generally found stroke prevention comparable to or better than warfarin, with less intracranial bleeding, and guidelines gradually moved to make these agents preferred for most eligible patients. Stroke risk stratification was refined in parallel, with CHA₂DS₂-VASc widely adopted after 2010.

Rhythm control was reassessed by later trials. CABANA (2019) compared catheter ablation with antiarrhythmic drugs and found no overall mortality difference, while EAST-AFNET 4 (2020) showed that early rhythm control in recently diagnosed patients reduced cardiovascular outcomes. These results, together with newer guidelines from 2023 and 2024, have pushed earlier rhythm control and comorbidity management into mainstream practice.

Open questions and research directions

Anticoagulation for subclinical and device-detected AF

The central question is whether anticoagulation helps patients whose AF is found only by pacemakers, defibrillators or implanted monitors. The two large randomized trials disagree in their details. NOAH-AFNET 6 failed to show a reduction in stroke with edoxaban versus placebo, but it was stopped early and so was underpowered. ARTESiA, which compared apixaban with aspirin, found that apixaban was better at preventing stroke and systemic embolism than aspirin, but there was a higher risk of major bleeding. A pooled analysis reported a relative risk reduction of stroke of approximately 32% and a relative increase in major bleeding of approximately 62%. Because absolute event rates are low, the balance is sensitive to how risk is framed. Estimates from the pooled data suggest that anticoagulation prevented approximately 0.3 strokes per 100 patient-years and induced 0.7 to 1.6 major bleeding events per 100 patient-years.

Risk thresholds are equally contested. An NOAH-AFNET 6 substudy found that the burden of subclinical AF didn't predict stroke or other embolic events, or the impact of direct oral anticoagulation. Duration may matter more than burden: patients with episodes of 24 hours or more were more likely to develop clinical AF over time than those with shorter durations. Benefit also appears to depend on baseline vascular disease. In a combined analysis, the effect of anticoagulation on the composite primary outcome was reported as 0.75 (95% C.I. 0.61-0.92) in the vascular disease group compared with 1.01 (95% C.I. 0.78-1.30) in patients without vascular disease. Whether these subgroup signals should set thresholds, and at what level, remains unresolved.

Catheter ablation versus antiarrhythmic drugs

The CABANA trial enrolled 2204 patients with symptomatic AF and compared catheter ablation with drug therapy. In the intention-to-treat analysis, the primary end point occurred in 8.0% of patients in the ablation group versus 9.2% in the drug therapy group (hazard ratio 0.86; 95% CI 0.65-1.15; P = .30). The trial therefore did not show a significant reduction in death, disabling stroke, serious bleeding or cardiac arrest. Its benefits lay elsewhere: ablation improved the quality of life and freedom from atrial fibrillation recurrence, and in the heart failure subgroup, ablation improved both survival and quality of life. Crossover complicated interpretation, since 301 (27.5%) of the patients assigned to drug therapy ultimately received catheter ablation. Post hoc analyses have since tried to separate the effect of sinus rhythm itself from the randomized strategy, a question that the trial's design could not fully settle.

Whether rhythm control improves long-term prognosis is therefore still open. Outcome comparisons are also affected by how recurrence is measured. Recurrence is hard to detect when it is asymptomatic, which complicates comparisons across trials and devices.

Machine learning and AI on ECGs and wearable data

Deep neural networks are increasingly applied to standard and single-lead ECGs, as well as to wearable signals. Accuracy figures in early reports are often high, but they depend heavily on the reference standard and on which tracings are counted. In one study, 1 in 7 tracings was deemed as unreadable by the smartwatch, and 17% of the total ECGs recorded by the smartwatch could not be algorithmically classified as either AFib or sinus rhythm. A deep-learning platform reported a diagnosis on 99% of ECGs, while the Apple app was able to classify only 78% of strips. Inconclusive rates are therefore a central validation metric, alongside sensitivity and specificity, which a pooled analysis summarized as 86% and 94%, respectively for algorithmic smartwatch readings.

Other challenges are specific to AI. Many studies are retrospective or single-centre, and models trained in one population may not transfer. One risk-prediction model that estimates the probability of paroxysmal or incident AF from sinus-rhythm ECGs was validated in both internal and external cohorts, which illustrates the need for internal and interethnic external validations. Since AF prevalence in screening populations is often low, positive predictive value can be poor even when specificity looks high, and ground truth is usually a physician-read 12-lead ECG rather than long-term clinical outcomes.

Atrial cardiomyopathy and precision approaches

Atrial cardiomyopathy is a proposed framework in which structural, electrical and fibrotic changes in the atrium are regarded as a disease in their own right, capable of producing both AF and thromboembolism, possibly before AF is detected. Its operational definition and its markers remain under debate, and no single test is accepted as its reference standard. If validated, such markers could help explain why some patients with device-detected AF have a higher or lower stroke risk than their clinical scores predict.

Genetic risk prediction follows a similar pattern. Genome-wide studies have identified many loci associated with AF, and polygenic scores have been proposed to refine risk beyond clinical factors. Their incremental predictive value, their performance across ancestries and their clinical utility are all still being evaluated. Precision approaches would need to show that tailoring treatment on such scores changes outcomes.

Underlying these questions is one that has not been resolved: whether restoring sinus rhythm can be achieved without losing the benefits of other treatments. Stroke prevention, risk-factor control and rhythm control interact, and anticoagulation decisions after successful ablation still rest mainly on stroke-risk scores rather than on rhythm status.

References

1. Atrial Fibrillation Topic Reviewhealio.com
2. Atrial Fibrillation ECG Reviewhealio.com
3. What is Atrial Fibrillation?aclsmedicaltraining.com
4. atrial fibrillation (af) - atriyal fibrilasyon - ekg - ecg - ankara kardiyoloji - kalp hastalıkları - mete alpaslan - doktorekg.commetealpaslan.com
5. ECG How to Diagnose Atrial Fibrillation - Top 5 Tipsacadoodle.com
6. Atrial Fibrillation ECG Interpretation #312practicalclinicalskills.com
7. Atrial Fibrillation Guideline Introduces Stage-Based Classification, Stresses Early Rhythm Controlconsultqd.clevelandclinic.org
8. New atrial fibrillation guideline: Modify risk, control rhythm, prevent progressionccjm.org
9. Classification of Atrial Fibrillation - Decision Point - Medscapedecisionpoint.medscape.com
10. CONVERGE - Epi/Endo Ablation For Treatment of Persistent Atrial Fibrillation(AF)cdn.clinicaltrials.gov
11. Apixaban for Early Prevention of Recurrent Embolic Stroke and Hemorrhagic Transformationcdn.clinicaltrials.gov
12. 2023 Guideline for Diagnosis and Management of Atrial Fibrillation: Key Perspectives - American College of Cardiologyacc.org
13. Heart Failure and Sudden Cardiac Death Japan Registrycdn.clinicaltrials.gov
14. Valvular Atrial Fibrillation: A Confusing and Obsolete Definitionjacc.org
15. Use of Direct Oral Anticoagulants in Patients With Atrial Fibrillation and Valvular Heart Lesionsahajournals.org
16. ACC/AHA Clarify When AF Patients With Mitral Stenosis Should—And Should Not—Take DOACsajmc.com
17. Non-Valvular Atrial Fibrillation: The Overlooked Concepts and Challenges of Novel Oral Anticoagulantsjscimedcentral.com
18. Nonvalvular atrial fibrillation: Causes and treatmentmedicalnewstoday.com
19. Paroxysmal Atrial Fibrillation - an overviewsciencedirect.com
20. Spontaneous Initiation of Atrial Fibrillation by Ectopic Beats Originating in the Pulmonary Veinsnejm.org
21. Mechanoelectrical feedback in pulmonary vein arrhythmogenesis: Clinical challenges and therapeutic opportunitiesncbi.nlm.nih.gov
22. Atrial Fibrillation Pathophysiologyahajournals.org
23. Electrical, contractile and structural remodeling during atrial fibrillationresearchgate.net
24. Multiple wavelets, rotors, and snakes in atrial fibrillation - a computer simulation studyresearchgate.net
25. Mechanisms of Atrial Fibrillationresearchgate.net
26. Commentary: Atrial Fibrillation Dynamics and Ionic Block Effects in Six Heterogeneous Human 3D Virtual Atria with Distinct Repolarization Dynamicsncbi.nlm.nih.gov
27. Frontiersfrontiersin.org
28. Autonomic modulation preceding the onset of atrial fibrillation*jacc.org
29. Autonomic Nervous Dysfunction and Ultra-Short-Term Heart Rate Variability in Atrial Fibrillation: Recent Advances in Early Detection - PMCpmc.ncbi.nlm.nih.gov
30. Use of autonomic nervous system neurotransmitters inhibition and atrial parasympathetic fibers ablation for the treatment of atrial arrhythmias and to preserve drug effectsimage-ppubs.uspto.gov
31. Autonomic Dysregulation and Atrial Arrhythmias - PMCpmc.ncbi.nlm.nih.gov
32. Role of the Autonomic Nervous System in Atrial Fibrillationahajournals.org
33. Interrelationships Between the Autonomic Nervous System and Atrial Fibrillation - ScienceDirectsciencedirect.com
34. Atrial Fibrillation (AFib): Symptoms & Treatmentmy.clevelandclinic.org
35. Impact of lifestyle risk factors on atrial fibrillation: Mechanisms and prevention approachespmc.ncbi.nlm.nih.gov
36. Distinct risk factors of atrial fibrillation in patients with and without coronary artery disease: a cross-sectional analysis of the BOREAS-CAG Registry data - PMCpmc.ncbi.nlm.nih.gov
37. Obstructive Sleep Apnoea and Atrial Fibrillation - PMC - NIHpmc.ncbi.nlm.nih.gov
38. Lifestyle and Risk Factor Modification for Reduction of Atrial Fibrillation: A Scientific Statement From the American Heart Associationahajournals.org
39. Genetics and Pharmacogenetics of Atrial Fibrillation: A Mechanistic Perspectivejacc.org
40. Cross-ancestry genome-wide analysis of atrial fibrillation unveils disease biology and enables cardioembolic risk predictionncbi.nlm.nih.gov
41. Frontiersfrontiersin.org
42. Multi-ethnic genome-wide association study for atrial fibrillationnature.com
43. Meta-analysis of genome-wide associations and polygenic risk prediction for atrial fibrillation in more than 180,000 casesnature.com
44. KCNQ1 gain-of-function mutation in familial atrial fibrillation - PubMedpubmed.ncbi.nlm.nih.gov
45. Genetics of Atrial Fibrillation - PMC - NIHpmc.ncbi.nlm.nih.gov
46. Increased early sodium current causes familial atrial fibrillation and dampens effect of flecainidebiorxiv.org
47. A novel KCNQ1 missense mutation identified in a patient ...heartrhythmjournal.com
48. Modifiable Risk Factors in Atrial Fibrillation: Clinical Implications and Pathophysiological Insights - PMCpmc.ncbi.nlm.nih.gov
49. Prediction and Maintenance of Sinus Rhythm Among Atrial Fibrillation Patientscdn.clinicaltrials.gov
50. Risk Factors of Atrial Fibrillation in Patients with Heart Failurencbi.nlm.nih.gov
51. Genome-wide association study of 1 million people identifies 111 loci for atrial fibrillationbiorxiv.org
52. Insights From Atrial Fibrillation Genomics: From Bedside to Bench and Back Again - PMCpmc.ncbi.nlm.nih.gov
53. Familial atrial fibrillation mutation M1875T-SCN5A increases early sodium current and dampens the effect of flecainidencbi.nlm.nih.gov
54. Atrial Fibrillation - Cardiology - Merck Manual Professional Editionmerckmanuals.com
55. Atrial Fibrillation Symptoms: 9 to Know, Tests, and Treatmentshealthline.com
56. Atrial Fibrillation - Symptomsnhlbi.nih.gov
57. Atrial Fibrillation Symptoms and Pain Sensitizationcdn.clinicaltrials.gov
58. Patients with symptomatic permanent atrial fibrillation show quantitative signs of pain sensitisationncbi.nlm.nih.gov
59. Atrial Fibrillation: Common Questions and Answers About Diagnosis and Treatmentaafp.org
60. Smartwatch-Detected Atrial Fibrillation: The “Value” in the Positive Predictive Valueahajournals.org
61. Accuracy of Smartwatches in the Detection of Atrial Fibrillation: A Systematic Review and Diagnostic Meta-Analysisjacc.org
62. Watch Out: The Many Limitations in Smartwatch-Driven AF Detectionjacc.org
63. Accuracy of Apple Watch for Detection of Atrial Fibrillationahajournals.org
64. Accuracy, Usability, and Adherence of Smartwatches for Atrial Fibrillation Detection in Older Adults After Stroke: Randomized Controlled Trial - PMCpmc.ncbi.nlm.nih.gov
65. Large-Scale Assessment of a Smartwatch to Identify Atrial Fibrillationnejm.org
66. Accuracy of Smartphone Camera Applications for Detecting Atrial Fibrillation: A Systematic Review and Meta-analysisncbi.nlm.nih.gov
67. Ambulatory Cardiac Rhythm Monitoring 2024-07-01guidelines.carelonmedicalbenefitsmanagement.com
68. carelon.com Status: Created Effective Date: 07/01/2024 Doc ID: CAR11-0724.1guidelines.carelonmedicalbenefitsmanagement.com
69. Implantable Loop Recorders and Wearable Heart Rhythm Monitors Page 1 of 16uhcprovider.com
70. Atrial Fibrillation Detected by Continuous ECG Monitoringclinicaltrials.gov
71. Left Atrial Appendage Thrombus as a Marker of Disease Severity in 500 Patients with Atrial Fibrillation on Oral Anticoagulation: A 13-Year Follow-Up Study - PMCpmc.ncbi.nlm.nih.gov
72. Atrial fibrillation-induced tachycardiomyopathy and heart failure: an underappreciated and elusive conditionlink.springer.com
73. Arrhythmia-Induced Cardiomyopathy: JACC State-of-the-Art Reviewjacc.org
74. Tachycardia-Induced Cardiomyopathy - an overviewsciencedirect.com
75. Atrial Fibrillation-Induced Cardiomyopathy - PubMedpubmed.ncbi.nlm.nih.gov
76. Cognitive Impairment Associated With Atrial Fibrillation: A Meta-analysis: Annals of Internal Medicine: Vol 158, No 5_Part_1doi.org
77. Predictive role of atrial fibrillation in cognitive decline: a systematic review and meta-analysis of 2.8 million individualsncbi.nlm.nih.gov
78. Impact of Left Atrial or Left Atrial Appendage Thrombus on Stroke Outcome: A Matched Control Analysis - PMCpmc.ncbi.nlm.nih.gov
79. Atrial Fibrillation Management: ESC 2024 Guideline Summaryreference.medscape.com
80. Stroke Risk Stratification in Incident Atrial Fibrillation: A Sex-Specific Evaluation of CHA2DS2-VA and CHA2DS2-VAScncbi.nlm.nih.gov
81. Refining the CHA2DS2VASc risk stratification scheme: shall we drop the sex category criterion?academic.oup.com
82. Prospective Evaluation of the CHA₂DS₂-VA Score: Do Sex Differences Still Matter in Stroke?researchsquare.com
83. Clinical implications of the recalibrated CHA₂DS₂-VA score for women after ischemic stroke: a prospective cohort studyncbi.nlm.nih.gov
84. 2024 ESC Guidelines for Management of Atrial Fibrillation: Key Points - American College of Cardiologyacc.org
85. Atrial Fibrillation: AHA/ACCP/HRS Focused Update ...nursingcenter.com
86. Pocket Summary of the 2024 Atrial Fibrillation Guidelinessecce.es
87. CHA₂DS₂-VASc Score for Atrial Fibrillation Stroke Risk Calculatormdcalc.com
88. LAAOS III - Wiki Journal Clubwikijournalclub.org
89. Stroke Mechanism and Severity After Left Atrial Appendage Occlusion: Insights From the LAAOS III Randomized Clinical Trialncbi.nlm.nih.gov
90. LAAO during cardiac surgery prevents strokenature.com
91. LAAOS III: Left Atrial Appendage Occlusion Cuts Stroke Risk in Patients with Atrial Fibrillationhcplive.com
92. CHA₂DS₂-VASc Score for Atrial Fibrillation Stroke Risk [Updated]mdcalc.scholasticahq.com
93. Refining the CHA2DS2VASc risk stratification scheme: shall we drop the sex category criterion?ncbi.nlm.nih.gov
94. Spotlight on the 2024 ESC/EACTS management of atrial fibrillation guidelines: 10 novel key aspectsacademic.oup.com
95. Surgical bioprosthetic valve implantation or repair Mechanical valveacforum-excellence.org
96. Insights from the LAAOS III trial of left atrial appendage occlusion to prevent stroke in atrial fibrillationccjm.org
97. Spotlight on the 2024 ESC/EACTS management of atrial fibrillation guidelines: 10 novel key aspectsncbi.nlm.nih.gov
98. 2024 ESC Clinical Practice Guidelines for the Management of Atrial Fibrillationescardio.org
99. Modifiable Risk Factors in Atrial Fibrillation: Clinical Implications and Pathophysiological Insights - PubMedpubmed.ncbi.nlm.nih.gov
100. Risk Factor Modification Reduces Atrial Fibrillation - Professional Heart Dailyprofessional.heart.org
101. 2024 ESC Guidelines for Management of Atrial Fibrillation: Key Points - American College of Cardiologyacc.org
102. Prevention and Treatment of Atrial Fibrillation via Risk Factor Modification - ScienceDirectsciencedirect.com
103. Management of Atrial Fibrillation: Updated Guidance from the AHA, ACC, and HRSaafp.org
104. Optimizing Anticoagulation in Valvular Heart Disease: Navigating NOACs and VKAs - PMCpmc.ncbi.nlm.nih.gov
105. Anticoagulation in device-detected atrial fibrillation with or without vascular disease: a combined analysis of the NOAH-AFNET 6 and ARTESiA trialsacademic.oup.com
106. Direct Oral Anticoagulants for Stroke Prevention in Patients With Device-Detected Atrial Fibrillation: A Study-Level Meta-Analysis of the NOAH-AFNET 6 and ARTESiA Trialsahajournals.org
107. Management of device-detected subclinical atrial fibrillation: a European Heart Rhythm Association surveyacademic.oup.com
108. January 15, 2020 ◆ Volume 101, Number 2 www.aafp.org/afpaafp.org
109. Recurrence of Postoperative Atrial Fibrillation After Cardiac Surgery: Insights from a Tertiary Follow-Up Clinicncbi.nlm.nih.gov
110. Anticoagulation for atrial fibrillation after cardiac surgery: Do guidelines reflect the evidence? - The Journal of Thoracic and Cardiovascular Surgeryjtcvs.org
111. An Update on Guidelines to Prevent and Manage Atrial Fibrillation After Cardiac Surgery and a Survey of Practice in the UK - ScienceDirectsciencedirect.com
112. Postoperative atrial fibrillation following cardiac surgery: a persistent complicationacademic.oup.com
113. Anticoagulation in Atrial Fibrillation and Rheumatic Heart Diseasenejm.org
114. Direct oral anticoagulant use in nonvalvular atrial fibrillation with valvular heart disease: a systematic review - PMCpmc.ncbi.nlm.nih.gov
115. Anticoagulation for Prosthetic Valvesncbi.nlm.nih.gov
116. Postoperative atrial fibrillation (POAF) after cardiac surgery: clinical practice review - Suero - Journal of Thoracic Diseasejtd.amegroups.org
117. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates - PubMedpubmed.ncbi.nlm.nih.gov
118. Temporal trends and patterns in atrial fibrillation incidence: A population-based study of 3·4 million individuals - ScienceDirectsciencedirect.com
119. Prevalence, Age Distribution, and Gender of Patients With Atrial Fibrillation: Analysis and Implicationsjamanetwork.com
120. Worldwide Epidemiology of Atrial Fibrillationahajournals.org
121. Global, regional, and national burdens of atrial fibrillation/flutter from 1990 to 2019: An age-period-cohort analysis using the Global Burden of Disease 2019 studyncbi.nlm.nih.gov
122. Global, regional, and national burden of atrial fibrillation and atrial flutter in the working-age population from 1990 to 2021: A systematic analysis based on 2021 global burden of disease data - PMCpmc.ncbi.nlm.nih.gov
123. Global Epidemiological Transition of Atrial Fibrillation/Flutter (1990–2021): Multidimensional Burden Dynamics and Socioeconomic Health Gradients Across 204 Countries and Territoriesncbi.nlm.nih.gov
124. Global epidemiology of atrial fibrillation - PubMedpubmed.ncbi.nlm.nih.gov
125. One hundred years of atrial fibrillation - PMC - NIHpmc.ncbi.nlm.nih.gov
126. A history of atrial fibrillation - PubMedpubmed.ncbi.nlm.nih.gov
127. From rebellious palpitations to the discovery of auricular fibrillation: Contributions of Mackenzie, Lewis and Einthoven - ScienceDirectsciencedirect.com
128. A brief review: history to understand fundamentals of electrocardiographyncbi.nlm.nih.gov
129. Déjà vu in the theories of atrial fibrillation dynamics - PMCpmc.ncbi.nlm.nih.gov
130. [The concise history of atrial fibrillation]researchgate.net
131. Age and Sex Specific Prevalence of Clinical and Screen-Detected Atrial Fibrillation in Hospitalized Patientsncbi.nlm.nih.gov
132. Global, regional, and national prevalence, incidence, mortality, and risk factors for atrial fibrillation, 1990–2017: results from the Global Burden of Disease Study 2017academic.oup.com
133. Prevalence and Incidence of Atrial Fibrillation Among Older Primary Care Patientsjamanetwork.com
134. Global burden of early-onset atrial fibrillation and flutter: Population based study - ScienceDirectsciencedirect.com
135. A troubled beginning: evolving concepts of an old arrhythmia - ScienceDirectsciencedirect.com
136. Atrial Fibrillation: A Timeline of This Veritable Quandary - American College of Cardiologyacc.org
137. Apixaban Cuts Stroke but Ups Bleeding in Subclinical AFmedscape.com
138. ARTESiA Trial results were presented at the American Heart Association Scientific Sessions 2023stopafib.org
139. Anticoagulation in Patients With Device‐Detected Atrial Fibrillation With and Without a Prior Stroke or Transient Ischemic Attack: The NOAH‐AFNET 6 Trialahajournals.org
140. Subclinical AF Burden Not Enough for Anticoagulation Decision-Making, Trial Analysis Showsmedscape.com
141. Anticoagulation in device-detected atrial fibrillation with or without vascular disease: a combined analysis of the NOAH-AFNET 6 and ARTESiA trialsescardio.org
142. Catheter Ablation Versus Antiarrhythmic Drug Therapy for Atrial Fibrillation Trial (CABANA)biolincc.nhlbi.nih.gov
143. The CABANA Trial - HRShrsonline.org
144. Cost-Effectiveness of Catheter Ablation Versus Antiarrhythmic Drug Therapy in Atrial Fibrillation: The CABANA Randomized Clinical Trialahajournals.org
145. Clinical validation of a novel smartwatch for automated detection of atrial fibrillation - ScienceDirectsciencedirect.com
146. Accuracy and interpretability of smartwatch electrocardiogram for early detection of atrial fibrillation: A systematic review and meta‐analysis - Iqhrammullah - 2025 - Journal of Arrhythmia - Wiley Online Libraryonlinelibrary.wiley.com
147. Artificial intelligence–based electrocardiogram analysis improves atrial arrhythmia detection from a smartwatch electrocardiogram - PMCpmc.ncbi.nlm.nih.gov
148. Interethnic Validation of Artificial Intelligence for prediction of Atrial Fibrillation Using Sinus Rhythm Electrocardiogrammedrxiv.org
149. When to prescribe oral anticoagulation for patients with device-detected subclinical atrial fibrillation? Insights from the NOAH-AFNET 6 and ARTESiA trials - European Journal of Internal Medicineejinme.com
150. Recurrence of Atrial Fibrillation After Catheter Ablation or Antiarrhythmic Drug Therapy in the CABANA Trialjacc.org
151. Catheter Ablation vs Anti-arrhythmic Drug Therapy for Atrial Fibrillation Trialclinicaltrials.gov