Definition and pathophysiology
Atrial fibrillation (AF) is a major independent risk factor for thromboembolism, including ischaemic stroke. This risk applies across the clinical patterns of AF—paroxysmal, persistent, long-standing persistent and permanent—and is not determined solely by the apparent frequency or duration of documented episodes.
The central therapeutic objective is prevention of thromboembolic complications, particularly stroke. In untreated AF, stroke risk is increased approximately five-fold, and AF is associated with roughly one in five strokes. Thrombus formation is most commonly related to the left atrial appendage, although competing mechanisms such as large-artery atherosclerosis and small-vessel disease may also contribute to stroke in patients with AF.
Stroke risk is heterogeneous and is determined principally by the patient’s underlying clinical risk factors rather than by the AF pattern. Important predictors include previous stroke or transient ischaemic attack (TIA), mitral stenosis, hypertrophic cardiomyopathy, diabetes mellitus, hypertension, heart failure, older age, vascular disease and renal dysfunction.
The commonly used CHA2DS2-VASc score incorporates the following variables:
| Risk factor | Points |
|---|---|
| Congestive heart failure | 1 |
| Hypertension | 1 |
| Age ≥75 years | 2 |
| Diabetes mellitus | 1 |
| Prior stroke or TIA | 2 |
| Vascular disease | 1 |
| Age 65–74 years | 1 |
| Female sex | 1 |
Bleeding risk must be considered alongside thromboembolic risk. The HAS-BLED score includes hypertension, abnormal renal or liver function, prior stroke, bleeding tendency or previous bleeding, labile international normalized ratio (INR), older age, drugs predisposing to bleeding and alcohol use. A high bleeding-risk score should prompt correction of modifiable risk factors and closer surveillance; it does not, by itself, establish that anticoagulation should be withheld when the thromboembolic risk is substantial.
Clinical presentation and symptoms
The clinical manifestations of AF are variable. AF may be clinically apparent, first detected during management of another condition, or identified only by implanted or wearable devices as subclinical AF. Some patients have frequent symptomatic episodes, whereas others have brief or asymptomatic episodes that are not detected during routine clinical assessment.
The clinical burden of AF does not reliably predict stroke risk. Infrequent episodes observed during office-based assessments may coexist with asymptomatic episodes. Consequently, the absence of AF during intermittent monitoring is not sufficient evidence that thromboembolic risk is low.
In the setting of acute coronary syndrome, AF is the most frequent supraventricular arrhythmia. It may predate the coronary event, be detected for the first time, or develop during acute management. AF is often haemodynamically tolerated and may require no specific acute therapy beyond appropriate anticoagulation and management of ventricular rate. However, AF causing acute haemodynamic instability requires prompt treatment, with electrical cardioversion preferred.
Evaluation and risk assessment
The evaluation of a patient with AF who may require anticoagulation should establish:
Whether AF is clinical or device-detected subclinical AF.
The AF pattern, while recognising that pattern does not determine the anticoagulation indication.
Previous stroke or TIA.
The presence of mitral stenosis, a mechanical heart valve or hypertrophic cardiomyopathy.
Heart failure, hypertension, diabetes and vascular disease.
Age and sex category.
Renal and hepatic function.
Previous bleeding, bleeding predisposition, alcohol use and concomitant drugs that increase bleeding risk.
Adherence and the suitability of the proposed anticoagulant dose.
Whether concomitant antiplatelet therapy is genuinely indicated.
The strongest established predictors of ischaemic stroke and systemic embolism are previous stroke or TIA and mitral stenosis. In patients with AF and previous ischaemic stroke treated with aspirin, the risk of recurrent stroke has been reported in the range of 10–12% per year. Conversely, younger patients without comorbidities have a much lower long-term risk.
Renal dysfunction is an independent thromboembolic risk factor. The risk is increased in chronic kidney disease and is higher still in patients requiring haemodialysis or renal transplantation. Renal function is also important for drug selection and dosing because dabigatran, rivaroxaban and apixaban undergo renal excretion.
Diagnostics
Electrocardiography and rhythm documentation
The source material does not provide detailed electrocardiographic criteria for AF. It distinguishes between clinical AF and subclinical, device-detected AF. Subclinical AF is generally asymptomatic and may be detected only by prolonged monitoring with implantable loop recorders, pacemakers or defibrillators.
The absence of AF on periodic monitoring does not establish absence of clinically relevant thromboembolic risk. The role of continuous monitoring with implanted devices to guide anticoagulation in patients with borderline risk remains uncertain.
Cardiac imaging
Cardiac imaging can assist in evaluating thromboembolic risk and in planning rhythm-control procedures. Before AF catheter ablation, imaging should be considered in patients at high risk of ischaemic stroke or thromboembolism despite anticoagulation, to exclude intracardiac thrombus.
In patients with acute ischaemic stroke and AF, echocardiography contributes to decisions about the timing of anticoagulation. The presence of an intracardiac thrombus may favour earlier anticoagulation when the cerebral bleeding risk is acceptable. Conversely, absence of intracardiac thrombus may permit waiting several days when clinically appropriate.
The source material does not provide detailed echocardiographic criteria for diagnosis of AF or for quantifying atrial structure and function.
Biomarkers and laboratory findings
Laboratory assessment is relevant primarily to safe and effective anticoagulant use rather than to the diagnosis of AF itself.
Renal and hepatic assessment
Renal function should be assessed before and during treatment because renal impairment affects both thromboembolic and bleeding risk and may require dose adjustment with renally excreted direct oral anticoagulants (DOACs). The source material also identifies hepatic dysfunction as a bleeding-risk factor, although it does not specify testing schedules or dose thresholds.
INR monitoring
Warfarin requires monitoring of the prothrombin time and INR to adjust dosing and maintain therapeutic anticoagulation. It takes several days to achieve a therapeutic effect, described in the source material as an INR above 2. Labile INR is both a practical limitation and a component of bleeding-risk assessment.
When ischaemic stroke occurs despite anticoagulation, INR measurement or assessment of DOAC levels may help identify non-adherence, underdosing or another correctable explanation. The source material does not specify validated therapeutic ranges for DOAC concentrations or routine monitoring protocols.
Anticoagulant therapy
General principles
The default strategy is oral anticoagulation for eligible patients with AF, except those at low risk of incident stroke or thromboembolism. For patients with an estimated annual stroke or thromboembolic risk above 2%, treatment selection should be determined by thromboembolic risk rather than by whether AF is paroxysmal, persistent or permanent.
Anticoagulation should be reconsidered periodically. Reassessment should address:
Stroke and thromboembolic risk.
Bleeding risk.
Net clinical benefit.
Adherence.
Drug interactions.
Renal function.
Correctness of the prescribed dose.
Direct oral anticoagulants
The direct-acting anticoagulants discussed in the source material are:
Dabigatran.
Rivaroxaban.
Apixaban.
Edoxaban.
In non-valvular AF, these agents were non-inferior to warfarin in the individual trials described. Pooled analyses suggested small absolute advantages over warfarin for outcomes including mortality, stroke, major bleeding and intracranial haemorrhage. Direct-acting agents are simpler to administer, achieve anticoagulation promptly and generally do not require dose adjustment based on serial coagulation blood tests.
Renal excretion is clinically relevant for dabigatran, rivaroxaban and apixaban. Dose adjustment may be necessary with modest renal impairment, particularly in older patients who are also at increased risk of bleeding. P-glycoprotein inducers and inhibitors can influence drug exposure.
The source material provides no standard AF dosing schedules for these agents. Dosing must therefore be individualised according to the specific drug, renal function, age, interacting medicines and the applicable prescribing guidance.
Vitamin K antagonist therapy
Warfarin remains necessary for patients with rheumatic mitral stenosis or mechanical heart valves. In patients with rheumatic heart disease-associated AF, warfarin has been more effective than rivaroxaban for the composite of cardiovascular events or death without a higher bleeding rate. Apixaban and dabigatran have not demonstrated non-inferiority to warfarin in patients with mechanical heart valves and are less effective for preventing valve thrombosis or thromboembolism in that setting.
Practical limitations of warfarin include:
Delayed onset of therapeutic anticoagulation.
Requirement for INR monitoring.
Numerous drug and food interactions.
Difficulty maintaining a stable therapeutic effect.
Warfarin can be reversed with fresh frozen plasma, prothrombin complex concentrate and vitamin K.
Reversal of DOAC effect
Idarucizumab is available for reversal of dabigatran, while andexanet alfa is available for factor Xa inhibitors. Both are administered intravenously. Because reversal agents may be prothrombotic, their use requires careful clinical judgement.
Antiplatelet therapy and combined antithrombotic treatment
Aspirin or clopidogrel alone should not be substituted for oral anticoagulation when anticoagulation is indicated for AF-related stroke prevention. Antiplatelet therapy is less effective than warfarin and does not provide a compensatory reduction in bleeding risk. Dual antiplatelet therapy with aspirin and clopidogrel is also inferior to warfarin and causes more bleeding than aspirin alone.
Combining an anticoagulant with an antiplatelet agent increases bleeding risk and has not shown a clear general benefit for stroke or mortality prevention in AF. Such combinations should generally be reserved for selected patients with a separate indication, such as acute coronary syndrome or recent coronary or peripheral arterial stenting.
For AF patients undergoing percutaneous coronary intervention, oral platelet inhibition with a P2Y inhibitor—preferably clopidogrel—is recommended in the source material. Triple antithrombotic therapy, preferably incorporating a DOAC, may be considered in patients with high ischaemic risk, such as acute coronary syndrome, for up to 30 days.
The low-dose rivaroxaban regimen of 2.5 mg combined with aspirin, although associated with reduced stroke risk in chronic vascular disease, cannot be generalised to AF because patients requiring full-dose anticoagulation were excluded.
Anticoagulation after acute ischaemic stroke
The timing of anticoagulation after an acute cerebral infarction requires balancing recurrent embolism against haemorrhagic transformation. The approach is influenced by infarct size, MRI findings and echocardiographic evidence of intracardiac thrombus.
| Clinical finding | Suggested approach in the source material |
|---|---|
| Small infarct with low bleeding risk | Consider early anticoagulation |
| Medium-sized infarct | Consider waiting 2–4 days |
| Large infarct with high bleeding risk | Wait at least 7 days |
| No gradient-echo MRI haemorrhagic signal | Early anticoagulation is more likely to be safe |
| A few scattered gradient-echo haemorrhagic signals | Consider waiting several days |
| Numerous gradient-echo haemorrhagic signals | High risk of further bleeding; consider waiting at least 10 days or longer |
| No intracardiac thrombus | Waiting several days may be appropriate |
| Intracardiac thrombus | Consider early anticoagulation if cerebral bleeding risk permits |
Large infarcts, particularly those caused by cardioembolism or infection, and uncontrolled hypertension carry a high risk of spontaneous haemorrhagic transformation. Systemic anticoagulation is therefore generally avoided during the first 5–7 days in these circumstances.
Acute anticoagulation may also be considered in selected situations such as an intraluminal clot in a major vessel, acute major cerebral-vessel dissection with superimposed thrombus, cerebral venous sinus thrombosis or an intracardiac thrombus. Recurrent strokes of unclear cause despite aggressive antiplatelet therapy may represent another circumstance in which acute systemic anticoagulation is considered, although robust clinical evidence is limited.
Device-detected subclinical AF
The benefit of anticoagulation is firmly established for clinical AF, whereas the treatment threshold for device-detected subclinical AF remains less certain.
Subclinical AF is associated with an approximately 2.5-fold increase in stroke risk. In patients with device-detected episodes lasting 6 minutes to 24 hours, apixaban reduced stroke or systemic embolism compared with aspirin but increased major bleeding. In another trial of patients with device-detected atrial high-rate episodes, edoxaban did not significantly reduce the composite of cardiovascular death, stroke or embolism compared with placebo and increased the composite of death or major bleeding. That trial was stopped early because of safety concerns and futility for efficacy.
These findings indicate that anticoagulation decisions for subclinical AF require attention to both AF burden and the individual thromboembolic and bleeding profiles. The source material does not establish a universal duration threshold at which anticoagulation should be initiated.
AF associated with acute coronary syndrome
AF during acute coronary syndrome is associated with more comorbidity and a higher risk of complications. In most cases it is haemodynamically tolerated and requires no specific acute intervention other than appropriate anticoagulation and management of the ventricular rate.
For acute rate control:
Beta-blockers may be used depending on the presence of heart failure and reduced left ventricular ejection fraction.
In patients with depressed left ventricular ejection fraction, amiodarone or digoxin may be used, with amiodarone preferred.
In hypotension, digoxin is preferred over amiodarone or beta-blockers.
Electrical cardioversion is preferred when AF produces acute haemodynamic instability.
Patients with AF and thromboembolic risk factors should receive chronic oral anticoagulation. Even transient, self-terminating AF during ST-elevation myocardial infarction may predict increased long-term stroke risk.
Anticoagulation around AF ablation
Periprocedural anticoagulation is required because catheter ablation carries a risk of ischaemic stroke and thromboembolism.
| Recommendation | Class | Level |
|---|---|---|
| Begin oral anticoagulation at least 3 weeks before catheter ablation in AF patients at elevated thromboembolic risk | I | C |
| Continue oral anticoagulation without interruption during AF catheter ablation | I | A |
| Continue oral anticoagulation for at least 2 months after ablation in all patients, regardless of rhythm outcome or CHA2DS2-VA score | I | C |
| After the first 2 months, determine anticoagulation according to CHA2DS2-VA score rather than perceived procedural success | I | C |
| Consider cardiac imaging before ablation in high-risk patients despite anticoagulation, to exclude thrombus | IIa | B |
The source material uses the CHA2DS2-VA score in the post-ablation recommendation; this excludes the sex-category component used in CHA2DS2-VASc.
Left atrial appendage occlusion and surgery
Because most atrial thrombi are thought to arise in the left atrial appendage, appendage exclusion may provide an alternative or adjunctive strategy in selected patients.
Surgical closure of the left atrial appendage is recommended as an adjunct to oral anticoagulation in patients with AF undergoing cardiac surgery. It should be considered as an adjunct during endoscopic or hybrid AF ablation. Stand-alone endoscopic closure may be considered when long-term anticoagulation is contraindicated.
Percutaneous left atrial appendage occlusion devices may be considered in patients with high thromboembolic risk and serious bleeding risk from chronic oral anticoagulation. The source material describes these devices as comparable to warfarin for stroke prevention, with additional reductions in major bleeding—particularly haemorrhagic stroke—and all-cause mortality. They also appear non-inferior to DOACs for major AF-related cardiovascular, neurological and bleeding events.
In patients undergoing mitral valve surgery who are suitable for rhythm control, concomitant surgical ablation is recommended. Concomitant surgical ablation should also be considered during non-mitral valve cardiac surgery in appropriate patients. Intraprocedural imaging for left atrial thrombus is recommended during surgical ablation regardless of oral anticoagulant use.
Residual stroke risk despite anticoagulation
Oral anticoagulation substantially reduces, but does not eliminate, ischaemic stroke risk. Stroke occurring during anticoagulant therapy may reflect:
A non-AF-related mechanism, such as large-artery or small-vessel disease.
Non-adherence.
An inappropriately low dose.
Drug interactions.
Subtherapeutic warfarin anticoagulation.
Thromboembolism despite adequate anticoagulation.
INR measurement or assessment of DOAC levels may help identify an actionable cause. Management should include systematic evaluation of adherence, dosing, interactions and vascular risk factors.
Routine switching between DOACs, or from a DOAC to a vitamin K antagonist, is not recommended because proven efficacy is lacking. Switching may be justified for an individual reason, including a clinically important drug interaction, but observational evidence suggests only limited reduction in recurrent ischaemic stroke. Adding an antiplatelet agent to anticoagulation may increase bleeding risk and should not be used routinely.
Guideline-based treatment framework
The principal recommendations can be summarised as follows:
Patients with AF and an estimated annual stroke or thromboembolic risk above 2% should receive stroke-prevention therapy selected according to thromboembolic risk, irrespective of AF pattern.
Oral anticoagulation is the standard preventive treatment for eligible patients.
Antiplatelet drugs alone are not an alternative to anticoagulation for AF-related stroke prevention.
Thromboembolic and bleeding risks, renal function, adherence and dosing should be reassessed periodically.
Warfarin is required in rheumatic mitral stenosis and mechanical heart valves.
DOACs are generally easier to administer in non-valvular AF and have lower intracranial bleeding risk than warfarin in the source material.
Anticoagulant–antiplatelet combinations should be restricted to selected patients with an independent vascular indication.
Anticoagulation around AF ablation should be uninterrupted during the procedure, continued for at least 2 months afterwards, and subsequently guided by stroke risk rather than apparent rhythm success.
Left atrial appendage closure may be considered when long-term anticoagulation is unsuitable because of serious bleeding risk.
Following stroke despite anticoagulation, investigate adherence, dose, interactions, anticoagulant effect and competing stroke mechanisms before altering treatment.
Prognosis and follow-up
AF confers an enduring risk of stroke and systemic embolism, including in patients with paroxysmal AF or apparently infrequent episodes. Patients with AF during acute coronary syndrome have worse short- and long-term prognoses than those who remain in sinus rhythm. Even transient, self-terminating AF during ST-elevation myocardial infarction may be associated with increased long-term stroke risk.
Follow-up should be periodic and should include reassessment of:
Thromboembolic risk.
Bleeding risk and modifiable bleeding factors.
Renal and hepatic function where relevant.
Adherence and drug interactions.
Correct anticoagulant dosing.
The continued need for concomitant antiplatelet treatment.
New vascular comorbidities or intervening stroke or TIA.
Anticoagulation decisions should not be based on the perceived success of rhythm control or ablation. After the initial post-ablation period, treatment should continue or be discontinued according to the patient’s stroke-risk profile. Continuous monitoring may detect otherwise unrecognised AF, but its role in guiding anticoagulation for patients with borderline risk remains unresolved.
