A trial fibrillation (AF), one of the most common dysrhythmias, has numerous causes (Table 1). When an underlying cause is not identified, the condition is called “lone” AF. Approximately 10 percent of patients with chronic AF have this type.1 Management of AF requires assessment and treatment of the underlying cause, control of the rate, consideration of pharmacologic or electric cardioconversion and consideration of anticoagulation. This article discusses the issues involved in the risk of a thromboembolic event in association with AF and the indications for anticoagulation therapy.
|Coronary artery disease|
|Myocardial infarction with congestive heart failure|
|Following coronary bypass surgery|
|Valvular heart disease (especially mitral stenosis and regurgitation)|
|Cardiomyopathy (dilated, hypertrophic, restrictive)|
|Metabolic disorders (hypokalemia, hypomagnesemia, hypercalcemia)|
|Vagal stimulation (carotid sinus pressure)|
|Nausea and vomiting|
|Nocturnal atrial fibrillation (preceded by bradycardia)|
|Sympathetic excess (pheochromocytoma, hypoglycemia, psychologic stress)|
|Physical causes (lightning, electric shock, hypothermia, chest contusion)|
|Cold fluids injected into right atrium|
|Long-term ventricular pacing|
|Magnet applied to a dual-chamber pacemaker|
|Automatic cardioverter-defibrillator (AICD) firing, presence of AICD|
|Cardiotoxic drugs (e.g., anthracyclines)|
|Drugs (ophthalmic atropine, digoxin, theophylline, sympathomimetics, adenosine [Adenocard], antidepressants, nicotine gum)|
|Congenital heart disease (especially atrial septal defect)|
|Intrathoracic infections and inflammation (pneumonia, pericarditis, myocarditis, endocarditis)|
|Malignancy (with atrial involvement)|
|Sinus node dysfunction|
|Lone atrial fibrillation|
Incidence of Atrial Fibrillation
The incidence of AF increases with age and with the presence of structural heart disease.2 The Framingham Heart Study,3 which included a cohort of 2,325 men and 2,866 women who were evaluated biennially for 22 years, showed that chronic AF developed in 49 men and 49 women. The incidence of new-onset AF increased with age in both sexes. For men and women aged 22 to 34, the rate was 2.6 and 2.2 per 1,000, respectively. In the group 55 to 64 years of age, however, the rate of AF increased to 37.9 and 29.9 per 1,000 men and women, respectively. Patients with diabetes, hypertension, rheumatic heart disease, coronary artery disease and congestive heart failure have a higher incidence of AF than patients without these disorders.4
In another large study,5 which included 2,254 patients, AF was noted in 2 percent of the patients 65 to 75 years of age and in 5 percent of those older than 75 years of age. Again, the occurrence of AF was found to increase with age, with 14 percent of those older than 84 years of age having this dysrhythmia. In the more recently initiated Cardiovascular Health Study of Americans More Than 65 Years, the prevalence of AF on 24-hour Holter recordings was approximately 5 percent.4
Impact on Morbidity
A significant increase in the incidence of stroke was noted among participants with AF in the Framingham study.3 Patients with rheumatic heart disease and AF had a 17-fold increase in the incidence of stroke, whereas patients with nonrheumatic AF had a fivefold increase in stroke. A fourfold increase in stroke was found in patients with lone AF.11 Most of these patients were older than 60 years of age. In the Stroke Prevention in Atrial Fibrillation Investigators (SPAFI) study,7 patients with lone AF who were younger than 65 years of age had a stroke rate of 1 percent per year, compared with a stroke rate of 4.3 percent per year in patients over 65 years of age. Data from this study highlight the portentous prognosis associated with lone AF in older patients and underscore the need for treatment. Approximately 25 percent of the strokes in patients with lone AF occurred at the onset of AF. Stroke recurrence in the first six months was twice as high in the AF group.
In the Olmsted county population in Minnesota,12 where patients were followed for 30 years, lone AF was identified in 97 (2.7 percent) of 3,623 patients with AF. This study population differed from that of the Framingham study; in the Olmsted county study, patients older than 60 years of age were excluded at entry into the study. The incidence of stroke was 0.35 per 100 person-years, with stroke occurring in 1.3 percent at 15 years. It is important to recognize that patients older than 60 years at entry and those with hypertension were excluded from the study.
Lone AF in younger patients without other risk factors for stroke carries the same risk of stroke as that in the general population. The importance of AF in the general population is that its appearance heralds a mortality rate double that of control subjects. Much of the morbidity and some of the mortality associated with AF is due to stroke. While the risk of stroke is not due solely to AF, AF substantially increases the risk of stroke in the presence of other cardiovascular disorders (Table 3). The attributable risk of stroke from AF is estimated to be 1.5 percent in the 50- to 59-year-old age group and to approach 30 percent in persons 80 to 89 years of age.13,14
Clinical Trials of Anticoagulation in Nonvalvular AF
General consensus has been reached regarding the recommendations for anticoagulation in patients with rheumatic AF. However, until recently, clear-cut recommendations have not been available for management of patients with nonvalvular AF. Since 1989, many large, prospective, randomized trials have been conducted to evaluate the risks and benefits of warfarin (Coumadin) or aspirin therapy in the prevention of stroke. The results of these trials provide a strong foundation for recommendations regarding the use of anticoagulation therapy in patients with nonvalvular AF.
A reduction in the risk of thromboembolism with anticoagulation or antiplatelet agents as compared with placebo was noted in the following trials: the Copenhagen Atrial Fibrillation Aspirin and Anticoagulation Study (AFASAK),8 the Boston Area Anticoagulation Trial in Atrial Fibrillation (BATAF),15 the Canadian Atrial Fibrillation Anticoagulation Study (CAFA)16 and the Stroke Prevention in Atrial Fibrillation (SPAF) trial.6,7,17–19
Stroke Prevention in Atrial Fibrillation Trial
The SPAF III trial18 included 1,044 patients (mean age: 71 years) with nonrheumatic AF and underlying risk factors for stroke, including systemic embolism. Patients in the study were randomized to receive full-dose warfarin (International Normalized Ratio [INR]: 2.0 to 3.0) or 325 mg of aspirin plus low-dose warfarin (INR: 1.2 to 1.5). The risk factors for stroke in these patients included a history of hypertension (67 percent), congestive heart failure (44 percent), diabetes (19 percent), myocardial infarction (19 percent) and a mean left atrial size of 4.4 cm on two-dimensional echocardiography.
After approximately one year of follow-up, the trial was prematurely stopped because the stroke rate in the aspirin plus low-dose warfarin group was substantially higher than that in the warfarin group (7.8 percent as opposed to 2.6 percent). The advantage of full-dose warfarin (INR: 2.0 to 3.0) in reducing stroke was evident in the subgroup analysis of patients with a history of stroke. The rate of stroke in this subgroup was 3.4 percent in patients receiving full-dose warfarin, compared with a rate of 11 percent in those receiving aspirin and low-dose warfarin. In patients without a previous stroke, the rate of stroke was 1.1 percent in those receiving full-dose warfarin and 5.1 percent in those receiving low-dose warfarin and aspirin. The data clearly favor anticoagulation with warfarin to an intended INR of 2.0 to 3.0 in high-risk groups with AF.
Anticoagulation Before Cardioversion
In hemodynamically compromised patients, AF must be treated emergently with synchronized electric cardioversion. There are no controlled studies available to suggest recommendations regarding anticoagulation in these patients. In a nonrandomized trial of 437 patients (532 instances of cardioversion),20 the incidence of an embolic event was 0.8 percent in the group receiving anticoagulant therapy, compared with 5.3 percent in the group not receiving anticoagulant therapy. These results are impressive because the number of patients with congestive heart failure, hypertension and rheumatic heart disease was greater in the anticoagulation group than in the group not receiving anticoagulation.
The prevalence of intracardiac thrombi in patients with AF has been investigated in several studies.21–24 One study included 233 patients with AF of more than 48 hours' duration who had not previously received anticoagulants.22 A left atrial thrombus was found in 15 percent of the patients. In all but one patient, the thrombi were located in the atrial appendage. In another study,24 left atrial thrombi were detected by transesophageal echocardiography in 13 percent of 113 patients with no acute embolic event, even though AF had been present for fewer than three days. In 122 patients with AF of three days' duration or longer and no embolic event, the prevalence of left atrial thrombi was 29 percent.24
Based on these observations, it is generally recommended that anticoagulation be instituted for three weeks before cardioversion is attempted in patients with AF of more than two days' duration. To minimize thromboembolic complications, anticoagulants should be continued for four weeks after cardioversion. This period is required for recovery of atrial mechanical contractility after conversion to a sinus rhythm. Such a recommendation is based more on tradition and theory than on findings from controlled trials.
The time required for a clot to form in the atrium during AF is unknown. Once a clot is formed, the assumption is that a poorly adherent clot is more likely to dislodge at the time of cardioversion. It has been estimated that the time needed for organization of the newly formed clot is approximately two weeks. Therefore, anticoagulation for four weeks following cardioversion would cover the time required for organization of a clot, adherence of any existing clot and prevention of a new clot. Continued anticoagulation after cardioversion would theoretically protect the patient from embolic events until atrial mechanical activity is restored, which can take up to two weeks.
An alternate approach using transesophageal echocardiography has been suggested for use in hospitalized patients with AF of more than 48 hours' duration.22 In one prospective study,22 intravenous heparin was administered for at least 12 hours and transesophageal echocardiography was used to determine whether atrial thrombi were present. Patients without atrial thrombi underwent electrical or pharmacologic cardioversion, followed by warfarin therapy for four weeks after cardioversion. No clinical embolic events occurred in this group. Patients with atrial thrombi identified on transesophageal echocardiography received warfarin therapy for three weeks before cardioversion was attempted and for four weeks after cardioversion.
The need for anticoagulation before cardioversion in patients with AF of short duration (less than 48 hours) is less clear. These patients can harbor left atrial thrombi and systemic emboli. Intravenous administration of heparin before cardioversion, followed by warfarin therapy for four weeks after cardioversion, may be beneficial but supportive data are not available.
Risk of Bleeding with Antithrombotic Therapy
In pooled data from five randomized trials (mean age of patients was 69 years), the rate of intracranial hemorrhage in warfarin-treated patients (therapeutic target range varied) was 0.3 percent per year.25 In the subgroup of elderly patients (mean age: 80 years) in SPAF II,17 the rate of intracranial hemorrhage was 1.8 percent in those who received warfarin (INR: 2.0 to 4.5) and 0.8 percent in those who received aspirin. One explanation for the higher rate of intracranial hemorrhage in this elderly subgroup is that the selection of patients for enrollment was less restrictive than that used in the five trials. Elderly patients overall would be expected to have a rate of warfarin-associated intracranial hemorrhage higher than 0.3 percent per year, which was reported in collaborative AF trials analysis.25
In addition to increasing age of the patient, the strongest risk factors for warfarin-associated intracranial hemorrhage are poorly controlled hypertension, excessive anticoagulation and a history of cerebrovascular disease.
Anticoagulation with warfarin (INR: 2.0 to 3.0) is clearly beneficial in patients with non-rheumatic AF who are at moderate risk for stroke and have a low risk of bleeding as a result of anticoagulation (Table 4).26,27 Anticoagulation with warfarin should also be instituted in patients with a high risk of emboli unless the risk of bleeding is very high.
|Age of patient||Other risk factors for stroke||No other risk factors for stroke|
|< 65 years||Warfarin (Coumadin)||Aspirin or no anticoagulant therapy|
|65 to 75 years||Warfarin||Aspirin or warfarin|
|> 75 years||Warfarin||Warfarin|
Patients with lone AF who are under 60 years of age with no other risk factors for stroke are at low risk for systemic embolism and stroke. The risks of anticoagulation in these patients outweigh the potential benefits. Patients over 60 years of age are at intermediate risk of stroke and will benefit from anticoagulation. Younger patients without heart failure or cardiac chamber dilatation who are at high risk for atherosclerosis may take aspirin in a dosage of 325 mg per day.
A decision to initiate anticoagulation must be individualized, with the physician carefully weighing the risk of bleeding against the potential reduction in embolic risks. Current recommendations for antithrombotic therapy for AF necessitate individualization of therapy after an integrated clinical assessment that evaluates thromboembolic risk due to AF alone, other potential indications for anticoagulation, hemorrhagic risk and nonmedical factors related to compliance, the patient's ability to follow through with INR monitoring, the patient's gait stability, the risk of other trauma and the patient's preferences.26