Atrial fibrillation in adults: Selection of candidates for anticoagulation

INTRODUCTION — Atrial fibrillation (AF) is a major cause of morbidity and mortality in adults. While ischemic stroke due to embolization of left atrial thrombi is the most frequent clinical manifestation of embolization, embolization to other locations in the systemic circulation (and in the pulmonary circulation from right atrial thrombi) also occurs but is less commonly recognized. Stroke associated with AF tends to be more extensive/larger than stroke related to carotid artery disease. Chronic oral anticoagulation (OAC) is recommended to reduce the risk of thromboembolism for most patients with AF. However, such therapy is associated with an increased risk of bleeding, and recommendations for its use must take both benefit and risk into account through shared decision-making with the patient. (See "Stroke in patients with atrial fibrillation".)

This topic will focus on identifying which patients with AF require long-term/chronic OAC with either vitamin K antagonist (VKA; eg, warfarin) or direct oral anticoagulants (DOAC; also referred to as non-vitamin K oral anticoagulants [NOAC]). The discussion here excludes patients with rheumatic mitral stenosis that is severe or clinically significant (mitral valve area ≤1.5 cm²), a bioprosthetic valve (surgical or bioprosthetic) within the first three to six months after implantation, or a mechanical heart valve. Management for patients with these valve conditions is briefly discussed in a section below that provides links to related topics. (See 'Patients with valvular heart disease' below.)

Other potentially relevant topics to the reader include:

●Choice of OAC for AF (see "Atrial fibrillation in adults: Use of oral anticoagulants")

●(See "Prevention of embolization prior to and after restoration of sinus rhythm in atrial fibrillation".)

●(See "Stroke in patients with atrial fibrillation".)

●(See "Atrial fibrillation: Left atrial appendage occlusion".)

●(Related Pathway(s): Atrial fibrillation: Anticoagulation for adults with atrial fibrillation.)

APPROACH TO DECIDING WHETHER TO ANTICOAGULATE

Decision-making based upon risk assessment — A first step in deciding which patients with AF should receive long-term oral anticoagulation (OAC) is to assess the individual patient’s risks of thromboembolism and bleeding along with patient preferences. Long-term anticoagulation lowers the risk of clinical embolization in patients with AF, but its use is associated with an increased risk of bleeding.

The benefits and risks of OAC with respect to reduction in risk of stroke and increment in risk of bleeding must be carefully considered and discussed with each patient. The greater the estimated reduction in absolute stroke risk compared with the increase in absolute risk of life-threatening or severely debilitating bleeding (such as intracranial hemorrhage), the more likely a patient is to benefit from long-term OAC. The benefit generally outweighs the risk for all but those with the lowest risk of thromboembolism. In cases of more balanced stroke reduction and bleeding risks, OAC is less likely to provide a net benefit. Risk scores are commonly used to assess thromboembolic and bleeding risks, although these tools are subject to a number of limitations. (See 'CHA2DS2-VASc score' below and 'Bleeding risk' below.) (Related Pathway(s): Atrial fibrillation: Anticoagulation for adults with atrial fibrillation.)

●Our approach to deciding whether to prescribe anticoagulant therapy for patients with AF (without severe or clinically significant rheumatic mitral stenosis [mitral valve area ≤1.5 cm²], a bioprosthetic valve [surgical or bioprosthetic] within the first three to six months after implantation, or a mechanical valve) is as follows:

•For a CHA₂DS₂-VASc score ≥2 in males or ≥3 in females (calculator 1) (table 1), we recommend chronic OAC.

•For a CHA₂DS₂-VASc score of 1 in males and 2 in females (calculator 1) (table 1), the specific nonsex risk factor present and the documented burden of AF influences decision making:

-For patients with CHA₂DS₂-VASc score of 1 in males and 2 in females based on age 65 to 74 years, we recommend chronic OAC. Age 65 to 74 years is a stronger risk factor than the other factors conferring one CHA₂DS₂-VASc score point [1].

-For patients with other risk factors, the decision to anticoagulate is based upon the specific nonsex risk factor and the burden of AF. For patients with very low burden of AF (eg, AF that is well documented as limited to an isolated episode that may have been due to a reversible cause such as recent surgery, heavy alcohol ingestion, or sleep deprivation), it may be reasonable to forgo chronic OAC and institute close surveillance for recurrent AF, although it may not be possible to reliably estimate AF burden from surveying symptoms or infrequent monitoring. The frequency and duration of AF episodes vary widely over time and episodes are often asymptomatic. (See "Atrial fibrillation in patients undergoing noncardiac surgery", section on 'Anticoagulation after surgery' and "Atrial fibrillation and flutter after cardiac surgery", section on 'Anticoagulation'.)

•For patients with a CHA₂DS₂-VASc of 0 in males or 1 in females (calculator 1) (table 1), we suggest against anticoagulant therapy. Patient values and preferences may impact the decision. For example, a patient who is particularly stroke averse and is not at increased risk for bleeding (see 'Bleeding risk' below) may reasonably choose anticoagulation, particularly if the patient is a candidate for treatment with a direct oral anticoagulant (DOAC).

●For all potential candidates for OAC, bleeding risk and related possible contraindications to OAC should be reviewed (table 2 and table 3). The appropriate use of bleeding risk assessment is to draw attention to modifiable bleeding risk factors that can be mitigated, and to flag patients with high bleeding risk for early review and follow-up and to identify potential candidates for left atrial appendage occlusion [2-6]. (See 'Bleeding risk' below and "Management of warfarin-associated bleeding or supratherapeutic INR", section on 'Mitigating bleeding risk' and "Risks and prevention of bleeding with oral anticoagulants", section on 'Risk factors for bleeding'.)

Effects of anticoagulation — In identifying which patients with AF are likely to benefit from OAC, the relative risk reduction in thromboembolism with OAC identified in randomized trials (see 'General efficacy' below) is combined with estimates of baseline risk using the CHA₂DS₂-VASc score to estimate the expected absolute risk reduction from OAC (see 'CHA2DS2-VASc score' below). The estimated absolute risk reduction for thromboembolic events is weighed against the estimated increase in absolute risk of intracranial hemorrhage (ICH) and other major bleeding complications. (See 'Bleeding risk' below.)

General efficacy — For patients with AF, randomized trials have shown that therapeutic OAC (vitamin K antagonist [VKA] or DOAC) reduces the risk of ischemic stroke and other embolic events by approximately two-thirds compared with placebo irrespective of baseline risk (figure 1) [7-17].

A meta-analysis included six randomized trials comparing VKA (warfarin) with placebo or no treatment in a total of 2900 participants with AF (mean age at entry 69 years, 20 percent with prior stroke or transient ischemic attack [TIA]) [14]. The overall rate of stroke was 2.2 percent/patient year in the warfarin group and 6.0 percent/patient year in the control group (relative risk reduction 0.64; 95% CI 0.49-0.74). The absolute risk reduction was 2.7 percent/year for primary prevention and 8.4 percent/year for secondary prevention. With warfarin therapy, all-cause mortality was reduced by 1.6 percent/year (relative risk reduction 0.26; 95% CI 0.03-0.43).

While most of the evidence comparing OAC with placebo in patients involved treatment with VKA, a trial comparing edoxaban 15 mg daily with placebo in patients with AF ≥80 years old with low body weight found a similar relative reduction in risk of stroke or systemic embolism (2.3 versus 6.7 percent/year; hazard ratio 0.34, 95% CI 0.19-0.61) [18]. The possible implications of this study for edoxaban dose are discussed separately. (See "Atrial fibrillation in adults: Use of oral anticoagulants" and "Atrial fibrillation in adults: Use of oral anticoagulants", section on 'DOACs'.)

CHA2DS2-VASc score

Use — We use the CHA₂DS₂-VASc score (calculator 1) to estimate thromboembolic risk in patients with AF, while recognizing its limitations (see 'Potential alternatives' below and 'Limitations' below). This estimation of baseline thromboembolic risk is combined with the above information on relative risk reduction (see 'General efficacy' above) to estimate the expected absolute risk reduction.

●The annual risk of ischemic stroke in untreated patients is estimated to be 0.2, 0.6, and 2.2 percent for those with CHA₂DS₂-VASc scores of 0, 1, and 2 (table 1) [19]. However, stroke rates have varied substantially among studies, which may be due to differences in study populations and methodologies [2,20-24]. As an example, studies examining ischemic stroke rates in patients with a single risk factor have identified risks of <1 to 2.7 percent/year [25-27].

Among patients with AF, ischemic stroke is the dominant type of thromboembolic event. As an example, in a study including data on 39,973 participants in four randomized trials of anticoagulation, the incidence of nonstroke systemic embolic events (SEEs) was 0.23/100 person-years, and the incidence of cerebral embolism was 1.92/100 person-years [28]. Among those with SEEs, 58 percent occurred in the lower extremities, 31 percent in the visceral-mesenteric system, and 11 percent in the upper extremities.

●Among patients with AF treated with OAC, annual stroke risk is lowered by approximately two-thirds to <0.1, 0.2, and 0.6 percent, respectively. In addition to the lowering of stroke risk, there is evidence that warfarin, compared with no anticoagulant therapy, leads to less severe stroke episodes and lower 30-day stroke mortality [14,29].

●The annual risk of intracranial bleeding with warfarin is 0.2, 0.3, and 0.5 percent, respectively. The risk of ICH with DOAC is approximately half of that with VKA (table 4). (See "Atrial fibrillation in adults: Use of oral anticoagulants", section on 'Choice of anticoagulant'.)

Among adults with AF, females have a modestly higher risk of thromboembolism than males, but female sex is associated with increased risk primarily in patients with at least two CHA₂DS₂-VASc score non-sex risk score points [1,30]. Thus, we focus on non-sex risk factors when deciding whether OAC is indicated.

●For CHA₂DS₂-VASc score ≥2 in males or ≥3 in females (when the risk score points are from two or more non-sex risk factors), the benefit from OAC generally exceeds the risks of severe bleeding [19,31-33].

●For CHA₂DS₂-VASc score of 1 in males or 2 in females (one non-sex risk factor with a value of 1), the risk of thromboembolism varies depending upon the non-sex risk factor [1]. Among the risk factors with a one-point value, age 65 to 74 years and the presence of heart failure have the greatest effect on thromboembolic risk [1], and OAC is recommended in patients with any of these risk factors.

●For CHA₂DS₂- VASc score of 0 in males or 1 in females (zero nonsex risk factors), the thromboembolic risk is low [27], so no OAC is suggested. (See 'Approach to deciding whether to anticoagulate' above.)

The warfarin versus placebo or aspirin trials were reported in the early 1990s, raising concerns that the findings may not be applicable to contemporary clinical practice [31,34,35]. Studies evaluating more contemporary data have found that the absolute risk of stroke in untreated patients has fallen from approximately 8 percent/year to 4 or 5 percent/year (table 1), but the relative risk reduction attributable to anticoagulant therapy is in the same range as earlier studies [36,37]. A two-thirds risk reduction in thromboembolism using the more contemporary lower absolute risks is clinically important for patients with two or more nonsex risk factors and for selected patients with one nonsex risk factor.

Comparisons of the effects of VKA and DOAC are presented separately. (See "Atrial fibrillation in adults: Use of oral anticoagulants".)

Potential alternatives — A variety of risk scores, imaging methods, and biomarkers have been studied for their potential predictive value for thromboembolic risk in patients with AF [38].

The CHA₂DS₂-VASc score has been compared with potential alternatives including the CHADS₂ and ATRIA risk scores (table 1 and table 5). The clinical utility of a risk score for individuals with AF hinges primarily on its accuracy in identifying those at lowest risk for thromboembolism, as anticoagulation is generally recommended for individuals with all but the lowest level of risk. Systematic reviews suggest that the CHA₂DS₂-VASc score generally performs better than the CHADS₂ and ATRIA scores in identifying low-risk patients, although there have been some discrepant results for comparisons of CHA₂DS₂-VASc and ATRIA [38]. However, all these risk scores are subject to the limitations discussed below. (See 'Limitations' below.)

A potential alternative to the risk score approach is to calculate the risk for each patient based upon risk factors including age as a continuous variable using the Calculator of Absolute Stroke Risk (CARS) [1].

For patients with AF, there is no established role for routine cardiac imaging to assess thromboembolic risk. Transesophageal echocardiography (TEE) is used in patients with AF primarily to evaluate left and right atrial appendage anatomy/function to identify individuals who are free of atrial thrombi and are therefore candidates for early cardioversion (see "Prevention of embolization prior to and after restoration of sinus rhythm in atrial fibrillation"). Thromboembolic risk is associated with cardiac imaging findings, including evidence of left atrial thrombus (generally assessed by TEE; less commonly assessed by cardiovascular magnetic resonance [CMR] or cardiac computed tomography [CCT]) and depressed left ventricular ejection fraction (which can be assessed by transthoracic echocardiography, TEE, CMR, CCT, or nuclear methods) [39]. However, imaging findings have not been shown to improve risk stratification in patients with AF [2].

Limitations — Risk scores to estimate thromboembolic risk in patients with AF have limited predictive value when applied to individual patients.

One limitation is that risk scores utilize point systems that do not reflect differences in risk among included risk factors. Risk factors assigned equal point values are associated with substantially different risks, as illustrated by the following examples for the CHA₂DS₂-VASc score (table 1) [1]:

●Age 65 to 74 years is associated with substantially greater stroke risk than other risk factors assigned one point.

●A history of prior stroke, TIA, or thromboembolic event is assigned two points, but the risk associated with this risk factor is more than five times the risk associated with risk factors assigned one point.

●The continuous risk of age is lumped into categories, so that ages 65 years and 74 years each confer one point, despite the much higher actual risk associated with the older age.

Another limitation is that the event rates observed in populations used to generate risk score may differ from those occurring in different clinical settings (eg, community versus hospitalized) and patient populations with differing risk profiles.

Also, some clinical features or conditions may impact the risk of thromboembolism but are not included in risk models; these include the duration or frequency of episodes of paroxysmal AF and the presence of conditions such as chronic kidney disease and elevated troponin level. Prediabetes has also been implicated as a possible risk factor for stroke in patients with AF [40]. The potential role of troponin measurement in the assessment of the risk of embolization in patients with AF is discussed separately. (See 'Chronic kidney disease' below and "Elevated cardiac troponin concentration in the absence of an acute coronary syndrome", section on 'Atrial fibrillation'.)

Bleeding risk — When OAC is considered, the major safety concern is the increased risk of bleeding, especially major bleeding, which includes events that require hospitalization, transfusion, or surgery, or that involve particularly sensitive anatomic locations. Thus, bleeding risk and related contraindications to OAC should be reviewed (table 2).

A systematic review comparing various bleeding risk scores in patients with AF found that the HAS-BLED risk score (table 3) was the best predictor of bleeding risk [2], although all bleeding risk scores provide imprecise estimates for individual patients, do not provide estimates for specific types of major bleeds, and are based upon bleeding risk with warfarin. Two more recent studies confirmed the efficacy of the HAS-BLED score was comparable to or better than ORBIT score in patients treated with DOACs [41,42]. Among patients with AF, the three most important predictors of major bleeding (including ICH) are overanticoagulation with warfarin (defined as an international normalized ratio greater than 3.0), prior stroke, and older patient age [31,43-45]. (See "Risks and prevention of bleeding with oral anticoagulants".)

The risk of bleeding was evaluated in a cohort of over 16,000 patients diagnosed with AF between 2005 and 2010 [37]. The incidence of major bleeding with current, recent, past, or no VKA (warfarin) exposure was 3.8, 4.5, 2.7, and 2.9/100 patient-years, respectively. However, major bleeding includes ICH and extracranial bleeding, particularly gastrointestinal bleeding. ICH is the most serious bleeding complication, since the likelihood of mortality or subsequent major disability is substantially higher than with bleeding at other sites [46]. In this study and others, the annual risk of ICH in patients with AF who are not anticoagulated is estimated to be 0.2 percent/year; that risk approximately doubles with OAC with VKA [34,37]. Randomized trials have shown that the risk of ICH with DOAC (both direct thrombin and factor Xa inhibitors) is approximately half of that with VKA (table 4). (See "Atrial fibrillation in adults: Use of oral anticoagulants", section on 'Choice of anticoagulant'.)

Given differences in morbidity associated with different types of bleeding, we think most patients would weigh the reduction in risk of ischemic stroke primarily against the increase in risk of ICH, with less weight given to the risk of gastrointestinal bleed or other less serious bleeding. While the incremental absolute risk of ICH with VKA (approximately 0.2 percent/year) is not trivial, it is substantially less than the expected absolute reduction in risk of ischemic stroke from OAC for most patients with AF and two or more nonsex CHA₂DS₂-VASc risk factors.

One problem with the bleeding risk scores is that they were developed from studies that included bleeds of differing severity. While any bleed can lead to death or severe disability, most do not; the major exception is ICH. The morbidity associated with ICH is similar to that for ischemic stroke, while the morbidity associated with gastrointestinal bleeding is generally not as severe. (See "Risks and prevention of bleeding with oral anticoagulants", section on 'Intracranial'.)

For patients in the following clinical settings, the bleeding risk is significantly higher:

●Thrombocytopenia or known coagulation defect associated with bleeding

●Active bleeding or recent surgery with a concern for ongoing bleeding

●Prior severe bleeding (including ICH) while on an oral anticoagulant

●Aortic dissection

●Malignant hypertension

●Combined use of anticoagulant and antiplatelet (including regular use of nonsteroidal antiinflammatory) agents

SPECIFIC PATIENT GROUPS

Patients with valvular heart disease — For patients with valvular heart disease (excluding those with rheumatic mitral stenosis that is severe or clinically significant [mitral valve area ≤1.5 cm²], a bioprosthetic valve [surgical or transcatheter] within the first three to six months after implantation, or a mechanical heart valve), the above general approach to deciding on oral anticoagulation (OAC) applies, although the evidence in patients with severe native valve disease is more limited than for the general population of patients with AF [47]. (See 'Approach to deciding whether to anticoagulate' above.)

Approaches to antithrombotic therapy (including anticoagulation) in patients with AF with specific valve conditions are discussed separately:

●Rheumatic mitral stenosis that is severe or clinically significant (mitral valve area ≤1.5 cm²). (See "Rheumatic mitral stenosis: Overview of management", section on 'Prevention of thromboembolism'.)

●Mechanical heart valve. (See "Antithrombotic therapy for mechanical heart valves".)

●Surgically implanted bioprosthetic valve. (See "Antithrombotic therapy for mechanical heart valves".)

●Transcatheter bioprosthetic valve. (See "Transcatheter aortic valve implantation: Antithrombotic therapy", section on 'General approach'.)

AF type and management

Paroxysmal AF — Our approach to deciding whether to anticoagulate is generally similar for patients with paroxysmal AF (PAF; with or without symptoms) as for persistent, or permanent, AF, as described above (see 'Decision-making based upon risk assessment' above). However, the burden of AF (duration and frequency of episodes) is a factor for decision-making for selected patients in whom the balance of benefit versus risk of anticoagulation is uncertain, recognizing that it may not be possible to accurately estimate AF burden except in patients with cardiac implantable electronic devices that can measure AF burden. We consider the burden of AF in decision-making for patients aged <65 years and who have one nonsex CHA₂DS₂-VASc risk factor. On the other hand, patients with AF with past history of embolic stroke are at high risk for a recurrent thromboembolic event, so the burden of AF would generally not impact the decision to anticoagulate. (See 'Decision-making based upon risk assessment' above.)

As discussed separately, the risk of thromboembolism in patients with PAF appears to be lower than in patients with persistent AF, and thromboembolic risk is higher in those with greater AF burden (percentage of time in AF). (See "Paroxysmal atrial fibrillation", section on 'Risk of embolization'.)

There are no definitive data to establish a threshold duration of AF episodes for the initiation of anticoagulant therapy. Some of our experts recommend a single threshold for duration of AF of 30 seconds, and others use a threshold as long as 24 hours [48]. Those experts who do not routinely anticoagulate patients with shorter-duration AF believe that the benefit is small and potentially outweighed by the bleeding risk. However, the AF burden is likely to vary over time, so a patient with 30 seconds of AF in one monitoring period may well have 30 hours of AF in the next monitoring period. While a large proportion of patients with short episodes of AF will go on to experience longer episodes, it is also true that the reverse occurs in a sizable percentage of patients experiencing long episodes of AF [49]. Also, the extent to which thromboembolic risk may continue during periods of sinus rhythm is uncertain. (See "Paroxysmal atrial fibrillation", section on 'Risk of embolization'.)

Device-detected AF — Similar considerations apply to patients with device-detected (subclinical) AF as to patients with PAF, with the understanding that many patients with subclinical AF have very low AF burden.

Device-detected AF (also known as subclinical AF or atrial high-rate episodes [AHRE]) is commonly identified in patients with a pacemaker, implantable cardioverter-defibrillator, or implanted cardiac monitor. The effects of anticoagulation for device-detected AF were assessed by a meta-analysis of two randomized controlled trials [50]: a trial enrolling 2536 individuals (mean age 77.5 years; median CHA₂DS₂-VASc score 4; prior stroke, systemic embolism, or transient ischemic attack [TIA] 10 percent) with AHREs (median duration 2.8 hours; interquartile range [IQR] 0.8-9.2 hours) that compared edoxaban with placebo or aspirin 100 mg [51] and a trial enrolling 4012 individuals (mean age 76.8 years; mean CHA₂DS₂-VASc score 3.9; prior stroke, systemic embolism, or TIA 9 percent) with subclinical AF (median duration 1.5 hours; IQR 0.2-5 hours) that compared apixaban with aspirin 81 mg [52].

The stroke rates were low in both arms of both trials [50]. In the trial comparing edoxaban with placebo/aspirin, ischemic stroke rates were 0.9 and 1.1 percent per patient-year [51]. In the trial comparing apixaban with aspirin, ischemic stroke rates were 0.64 and 1.02 percent per patient-year [52].

Meta-analysis found that oral anticoagulation reduced ischemic stroke (relative risk [RR] 0.68, 95% CI 0.50-0.92) but did not reduce all-cause mortality (RR 1.08, 95% CI 0.96-1.21). Oral anticoagulation increased major bleeding (RR 1.62, 95% CI 1.05-2.5) [50].

Rhythm versus rate control — For patients with AF, the process of deciding whether to anticoagulate is generally the same regardless of the choice between rhythm control or rate control strategies. As discussed separately, the risk of thromboembolism is not reduced by clinical maintenance of sinus rhythm. (See "Management of atrial fibrillation: Rhythm control versus rate control", section on 'Thromboembolic risk'.)

AF after surgery — Approaches to OAC in patients with AF after cardiac surgery and after noncardiac surgery are discussed separately. (See "Atrial fibrillation and flutter after cardiac surgery", section on 'Our approach to postoperative anticoagulation' and "Atrial fibrillation in patients undergoing noncardiac surgery", section on 'Anticoagulation after surgery'.)

Older adults — For older adults, we follow the general approach described above, including careful assessment the relative benefits and risks of OAC (see 'Decision-making based upon risk assessment' above). The approach to chronic kidney disease is discussed below. (See 'Chronic kidney disease' below.)

In patients with documented frequent falls but without prior trauma (eg, fracture, subdural), we weight the risks and benefits of OAC versus left atrial appendage occlusion. In this clinical setting, we often recommend OAC and work to reduce the risk of falls. The risk of falls leading to subdural hematomas is increased in older adult patients taking oral anticoagulants independent of the agent chosen. (See "Risks and prevention of bleeding with oral anticoagulants", section on 'Age, race, and sex' and "Atrial fibrillation: Left atrial appendage occlusion".)

A Taiwanese database study compared 15,756 older (≥90 years of age) adults with AF (11,064 receiving no antithrombotic therapy, 4075 receiving antiplatelet therapy, and 617 on warfarin) with 14,658 older adult patients without AF and without antithrombotic therapy [53]:

●Patients with AF had a greater risk of ischemic stroke (5.75 versus 3.00 percent/year; hazard ratio [HR] 1.93, 95% CI 1.74-2.14) and a similar risk of intracranial hemorrhage (ICH; 0.97 versus 0.54 percent/year; HR 0.85, 95% CI 0.66-1.09) compared with those without AF.

●Among patients with AF, warfarin use was associated with a lower stroke risk (3.83 versus 5.75 percent/year; HR 0.69, 95% CI 0.49-0.96) compared with no antithrombotic therapy. There was a nominal but nonsignificant increase in risk of ICH (HR 1.26, 95% CI 0.70-2.25).

In a second, later cohort of patients ≥90 years of age with AF, 768 patients treated with warfarin were compared with 978 patients treated with a direct oral anticoagulant (DOAC) [53]. DOACs were associated with a lower risk of ICH compared with warfarin (0.42 versus 1.63 percent/year; HR 0.32, 95% CI 0.10-0.97) and similar rate of ischemic stroke (4.07 versus 4.59 percent/year; HR 1.16; 95% CI 0.61–2.22). (See "Atrial fibrillation in adults: Use of oral anticoagulants", section on 'Older adults'.)

Potential use of reduced-dose DOAC (edoxaban) in selected older adults with AF with low body weight is discussed separately. (See "Atrial fibrillation in adults: Use of oral anticoagulants", section on 'DOACs'.)

Chronic kidney disease — For most patients with AF and chronic kidney disease (CKD), we follow the general approach to selection of candidates for OAC described above (see 'Approach to deciding whether to anticoagulate' above). However, some of our authors consider anticoagulation for the very uncommon CKD patient with a CHA₂DS₂-VASc score of 0 in males or 1 in females.

For patients with CKD and AF, the following is our approach for deciding whether to anticoagulate (figure 2):

●Stages 2, 3, and 4 and 5 (not on dialysis) – For patients with estimated glomerular filtration rate (eGFR) of 15 to 89 mL/min/1.73 m², our approach is similar to the general approach described above (see 'Decision-making based upon risk assessment' above), although there are very limited data for patients with end-stage kidney disease. Individualized risk assessment is performed to carefully weigh the benefits and risks of anticoagulation, with special attention to the bleeding risk associated with CKD. (See "Overview of the management of chronic kidney disease in adults", section on 'Uremic bleeding'.)

●Stage 5 on dialysis – Among patients with end-stage kidney disease on dialysis, we anticoagulate some higher-risk individuals (based on the CHA₂DS_2-VASc score) after shared decision-making and discussion of risks and benefits between the clinician and the patient.

However, it is reasonable to not anticoagulate the following groups of individuals with AF and eGFR <30 mL/min (stages 4 and 5) given our uncertainty of the benefit-to-risk ratio for antithrombotic therapy in these patients:

●Patients with high frailty

●Patients with prior life-threatening bleeding or recurrent bleeding

●Patients with poorly controlled hypertension

AF is common in patients with CKD [54-59], with prevalence between 8 and 35 percent in patients on hemodialysis and approximately 7 percent in patients undergoing peritoneal dialysis [60-62]. This rate is significantly higher than in the general population [63-66]. Rates are even higher in studies which used prolonged/continuous monitoring for identifying AF (figure 3) [54,55,67]. CKD significantly increases thromboembolic risk above baseline and is also associated with increased risk of bleeding [68-71]. Studies assessing the independent predictive value of presence of CKD for thromboembolic risk beyond the CHA₂DS₂-VASc score have yielded mixed results [68,72,73]. The thromboembolic risk associated with CKD may be due to alterations in the normal hemostatic mechanisms. The increased bleeding risk, particularly from the gastrointestinal tract, is due to pathophysiologic mechanisms including impairment of normal platelet function secondary to factors such as uremic toxins, abnormal platelet arachidonic acid metabolism, altered von Willebrand factor, and reduction in intracellular adenosine diphosphate and serotonin, as well as an increase in the frequency of the need for invasive procedures [63]. (See "Uremic platelet dysfunction".)

The evidence to support OAC (vitamin K antagonist [VKA] or DOAC) is less robust in individuals with creatinine clearance <30 mL/min, as many such patients were excluded from the important randomized trials [74]. However, we believe that the benefit outweighs the risk in most cases.

The efficacy and safety of warfarin in patients with AF and CKD have been evaluated in observational studies which have come to differing conclusions [69,75-80]. A 2020 meta-analysis of 15 studies (with a total of 47,480 patients with AF and end-stage renal disease) found no difference in the risk of ischemic stroke (HR 0.96, 95% CI 0.82-1.13), a higher risk of hemorrhage stroke (HR 1.46, 95% CI 1.05-2.04), and no significant difference in mortality (HR 0.95, 95% CI 0.83-1.09) or major bleeding (HR 1.20, 95% CI 0.99-1.47) in comparing warfarin users with those not taking warfarin [81]. Many of the observational cohorts did not evaluate the quality of the OAC with warfarin, such as the time in the therapeutic range (TTR). This may be relevant since evidence suggests that higher TTR is associated with better outcomes. (See "Warfarin and other VKAs: Dosing and adverse effects", section on 'Monitoring (PT/INR)'.)

Hyperthyroidism — The role of anticoagulant therapy is less well defined in patients in whom the underlying disease associated with AF can be corrected, as in hyperthyroidism. (See "Epidemiology, risk factors, and prevention of atrial fibrillation" and "Cardiovascular effects of hyperthyroidism", section on 'Atrial fibrillation'.)

For patients with AF attributable to hyperthyroidism, we follow the general approach described above for identifying candidates for OAC. (See 'Approach to deciding whether to anticoagulate' above.)

After successful treatment of the disorder, and after documentation that AF has not been present for at least three months, most of our experts suggest discontinuing anticoagulant treatment with periodic reassessment of the patient for recurrence of AF. We consider the absence of symptoms or signs of AF and two-week continuous monitoring showing no AF as adequate documentation. Some experts prefer additional documentation. However, some of our experts make a decision about continuing anticoagulant therapy based on the CHA₂DS₂-VASc score independent of monitored rhythm in these patients.

Hypertrophic cardiomyopathy — The role of OAC in patients with hypertrophic cardiomyopathy and AF is discussed separately. (See "Hypertrophic cardiomyopathy in adults: Supraventricular tachycardias including atrial fibrillation", section on 'Long-term management'.)

Patients with cancer on chemotherapy — Several chemotherapy drugs have been associated with AF and atrial flutter. Depending on severity, dose reduction or discontinuation of the offending chemotherapy agent may be indicated. (See "Cardiotoxicity of cancer chemotherapy agents other than anthracyclines, HER2-targeted agents, and fluoropyrimidines".)

For most patients with AF and cancer who are on chemotherapy, we follow the general approach to selection of candidates for OACs described above. (See 'Approach to deciding whether to anticoagulate' above.)

For patients who have AF in the setting of chemotherapy-related thrombocytopenia, OACs may require a dose reduction in order to prevent bleeding. (See "Anticoagulation in individuals with thrombocytopenia", section on 'Atrial fibrillation'.)

ALTERNATIVES TO ANTICOAGULATION

Left atrial appendage occlusion — As discussed separately, left atrial appendage occlusion is the primary alternative for patients with AF (excluding those with severe or clinically significant rheumatic stenosis, a bioprosthetic valve [surgical or bioprosthetic] within the first three to six months after implantation, or a mechanical valve) who have an indication for anticoagulation but have a contraindication for long-term anticoagulation. (See 'Decision-making based upon risk assessment' above and "Atrial fibrillation: Left atrial appendage occlusion".)

Pharmacologic agents — For patients with AF, no other antithrombotic regimen is an effective and safe alternative to standard therapeutic oral anticoagulation (OAC). In this setting, other antithrombotic regimens are less effective in lowering thromboembolic risk than standard therapeutic OAC and some antithrombotic regimens entail a bleeding risk similar to or greater than standard therapeutic OAC:

●Aspirin plus clopidogrel – Dual antiplatelet therapy is preferred to aspirin alone in the occasional high-risk patient with AF who cannot be treated with any OAC for a reason other than risk of bleeding. Given the availability of the direct oral anticoagulant (DOAC) agents as alternatives to vitamin K antagonists (VKAs), this situation should be extremely uncommon. One possible example is a patient with contraindications to DOAC agents who cannot receive effective international normalized ratio (INR) monitoring for VKA. Of note, dual antiplatelet therapy and OAC have similar bleeding risks. Thus, a patient who would not be a candidate for OAC because of bleeding risk is also not a candidate for dual antiplatelet therapy.

In patients with AF, dual antiplatelet therapy (with aspirin plus clopidogrel) reduces the risk of thromboembolism compared with aspirin monotherapy but offers less protection against thromboembolism than OAC (with VKA or DOAC).

The safety and efficacy of dual antiplatelet therapy in patients with AF were investigated in the ACTIVE W and ACTIVE A trials. All patients in the two trials had AF and one or more risk factors for stroke. The primary endpoint in both trials was a composite outcome (the first occurrence of stroke, systemic [non-central nervous system] embolization, myocardial infarction, or vascular death). The ACTIVE W trial included 6706 patients who were randomly assigned to combined therapy with clopidogrel (75 mg/day) and aspirin (75 to 100 mg/day) or to OAC with a VKA (target INR 2.0 to 3.0) [82]. The trial was stopped at an interim analysis after a median follow-up of 1.3 years because VKA lowered the annual rate of the primary endpoint compared with combined antiplatelet therapy (3.9 versus 5.6 percent; relative risk [RR] 0.69, 95% CI 0.57-0.85).

The ACTIVE A trial included 7554 patients with AF who were not candidates for warfarin OAC and were randomly assigned to combined therapy with clopidogrel (75 mg/day) and aspirin (75 to 100 mg/day) or to aspirin alone at the same dose [83]. After a median follow-up period of 3.6 years, patients treated with clopidogrel plus aspirin had a significantly lower annual rate of the primary combined endpoint (6.8 versus 7.8 percent; RR 0.89, 95% CI 0.81-0.98), which was primarily driven by a reduction in stroke (2.4 versus 3.3 percent; RR 0.72, 95% CI 0.62-0.83). On the other hand, dual antiplatelet therapy resulted in a higher rate of major bleeding (2.0 versus 1.3 percent/year; RR 1.57, 95% CI 1.29-1.92).

●Aspirin monotherapy – Aspirin (or other antiplatelet agent) is not an effective therapy for preventing thromboembolic events in patients with AF. In patients with AF, some but not all meta-analyses of clinical trials comparing aspirin with placebo have found that aspirin reduced the risk of stroke and systemic embolism (table 6) [14,38,84]. In contrast, clinical trials have demonstrated that OAC (with VKA or DOAC) lowers the risk of thromboembolism compared with aspirin (table 6) [9,14-17,85-87].

●Aspirin plus low-dose warfarin – In contrast to therapeutic adjusted-dose warfarin (target INR 2.0 to 3.0), low-dose warfarin (1.25 mg/day or target INR between 1.2 and 1.5) in combination with aspirin (300 to 325 mg/day) should not be used to reduce stroke risk in patients with nonvalvular AF [17,88,89]. In the SPAF-III trial of 1044 patients with AF who were at high risk for embolism, low-dose warfarin plus aspirin had a much higher rate of ischemic stroke and systemic embolism than therapeutic adjusted-dose warfarin (figure 4A-B) [88].

●Aspirin plus full-dose warfarin – Limited available data suggest that there is no benefit from adding aspirin to therapeutic OAC in patients with AF. In a post-hoc analysis of the SPORTIF trials in patients with AF, among patients taking aspirin plus warfarin (or aspirin plus the factor Xa inhibitor ximelagatran) experienced similar rates of stroke and systemic embolism as those taking warfarin alone (or ximelagatran alone) [90]. The risk of major bleeding was higher with aspirin plus warfarin compared with warfarin alone (3.9 versus 2.3 percent/year).

The management of antithrombotic therapy for patients with AF treated with OAC who have a concurrent indication for antiplatelet therapy is discussed separately. (See "Atrial fibrillation in adults: Use of oral anticoagulants", section on 'Concomitant antiplatelet therapy'.)

RECOMMENDATIONS OF OTHERS — Recommendations for choosing which patients with atrial fibrillation should be anticoagulated are available from the American Heart Association/American College of Cardiology/Heart Rhythm Society, the European Society of Cardiology, and the American College of Chest Physicians [38,91-93]. In general, we agree with relevant recommendations made in these guidelines.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Atrial fibrillation" and "Society guideline links: Arrhythmias in adults" and "Society guideline links: Anticoagulation".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, “The Basics” and “Beyond the Basics.” The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on “patient info” and the keyword(s) of interest.)

●Basics topics (see "Patient education: Atrial fibrillation (The Basics)" and "Patient education: Medicines for atrial fibrillation (The Basics)" and "Patient education: Choosing an oral medicine for blood clots (The Basics)" and "Patient education: Taking oral medicines for blood clots (The Basics)")

●Beyond the Basics topics (see "Patient education: Atrial fibrillation (Beyond the Basics)" and "Patient education: Warfarin (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Role of oral anticoagulation (OAC) in atrial fibrillation (AF) – In patients with AF, OAC reduces the risk of thromboembolism by approximately two-thirds across clinical risk factor profiles but also entails an increased risk of major bleeding.

●Deciding whether to anticoagulate – For each patient, their estimated absolute risk reduction for thromboembolic events is weighed against their estimated increase in absolute risk of intracranial hemorrhage and other life-threatening or severely debilitating bleeding complications. (See 'Approach to deciding whether to anticoagulate' above.)

●CHA₂DS₂-VASc risk score – Our approach to deciding whether to prescribe anticoagulant therapy for patients with AF (excluding those with rheumatic mitral stenosis that is severe or clinically significant [mitral valve area ≤1.5 cm²], a bioprosthetic valve [surgical or transcatheter] within the first three to six months after implantation, or a mechanical heart valve) is as follows (see 'Approach to deciding whether to anticoagulate' above):

•For a CHA₂DS₂-VASc score ≥2 in males or ≥3 in females (calculator 1) (table 1), we recommend chronic OAC (Grade 1A).

•For a CHA₂DS₂-VASc score of 1 in males and 2 in females (calculator 1) (table 1):

-For patients with CHA₂DS₂-VASc score of 1 in males and 2 in females based on age 65 to 74 years, we recommend chronic OAC (Grade 1A). Age 65 to 74 years is a stronger risk factor than the other factors conferring one CHA₂DS₂-VASc score point.

-For patients with other risk factors, the decision to anticoagulate is based upon the specific nonsex risk factor and the burden of AF. For patients with very low burden of AF (eg, AF that is well documented as limited to an isolated episode that may have been due to a reversible cause such as recent surgery, heavy alcohol ingestion, or sleep deprivation), it may be reasonable to forgo chronic OAC and institute close surveillance for recurrent AF, although it may not be possible to reliably estimate AF burden from surveying symptoms or infrequent monitoring. (See "Atrial fibrillation in patients undergoing noncardiac surgery", section on 'Anticoagulation after surgery' and "Atrial fibrillation and flutter after cardiac surgery", section on 'Anticoagulation'.)

•For patients with a CHA₂DS₂-VASc of 0 in males or 1 in females (calculator 1) (table 1), we suggest against OAC (Grade 2C). Patient values and preferences may impact the decision. For example, a patient who is particularly stroke averse and is not at increased risk for bleeding (see 'Bleeding risk' above) may reasonably choose anticoagulation, particularly if the patient is a candidate for treatment with a direct oral anticoagulant (DOAC).

●Bleeding risk – For all potential candidates for OAC, bleeding risk and related possible contraindications to OAC should be reviewed (table 2 and table 3). (See 'Bleeding risk' above.)

The appropriate use of bleeding risk assessment is to draw attention to modifiable bleeding risk factors that can be mitigated to flag high-bleeding-risk patients for early review and follow-up. (See "Risks and prevention of bleeding with oral anticoagulants", section on 'Bleeding risk scores' and "Management of warfarin-associated bleeding or supratherapeutic INR", section on 'Mitigating bleeding risk'.)

●Specific patient groups

•Paroxysmal AF – Our approach to deciding whether to anticoagulate is generally similar for patients with paroxysmal AF (PAF; with or without symptoms) as for persistent, or permanent, AF, as described above (see 'Decision-making based upon risk assessment' above). However, the burden of AF (duration and frequency of episodes) is a factor for decision-making for selected patients in whom the balance of benefit versus risk of anticoagulation is uncertain. (See 'Paroxysmal AF' above.)

Similar considerations apply to patients with device-detected (subclinical) AF, with the understanding that many patients with subclinical AF have very low AF burden. (See 'Device-detected AF' above.)

•Concurrent conditions – Our approach to OAC for patients with AF who are older, have chronic kidney disease, hyperthyroidism, and hypertrophic cardiomyopathy can sometimes differ for patients who are younger or do not have these conditions. (See 'Specific patient groups' above.)

●Contraindication to OAC – For patients with AF (excluding those with severe or clinically significant rheumatic stenosis, a surgical bioprosthetic valve within the first three to six months after implantation, or a mechanical valve) with an indication for OAC but who have a contraindication for long-term OAC, the primary alternative is left atrial appendage occlusion. For such patients, no other antithrombotic regimen is an effective and safe alternative to standard therapeutic OAC. (See 'Alternatives to anticoagulation' above and "Atrial fibrillation: Left atrial appendage occlusion".)