Atrial Fibrillation: How Best to Use Rate Control and Anticoagulation

April 1, 2006

Atrial fibrillation (AF) is the most commonsustained cardiac arrhythmia; itaffects about 2.2 million Americans.The prevalence of AF, which increaseswith age,1 is approximately 5.9% in personsolder than 65 years2 and greaterthan 10% in those older than 75 years.3

Atrial fibrillation (AF) is the most commonsustained cardiac arrhythmia; itaffects about 2.2 million Americans.The prevalence of AF, which increaseswith age,1 is approximately 5.9% in personsolder than 65 years2 and greaterthan 10% in those older than 75 years.3

Framingham data revealed a morethan 2-fold increase in the incidence ofAF for each decade of life in personsolder than 55 years.4 Thus, more thantwo thirds of patients with AF are between65 and 85 years of age.4

Men are approximately 1.5 timesmore likely to have chronic AF thanwomen of the same age. However, becauseelderly women outnumber elderlymen, the absolute number ofwomen with AF is greater.2

Here, we discuss the most commonpresentations of AF and describetreatment strategies for rate controland prevention of thromboembolism.In a second article, beginning on page539, we will address the benefits andrisks of conversion to sinus rhythm.



AF may be classifiedas:

  • Acute (less than a few days).
  • Paroxysmal (episodic with spontaneousconversion to normal sinusrhythm or requiring intervention forconversion).
  • Chronic and/or permanent (refractoryto conversion).5

Paroxysmal AF ultimately becomeschronic in 25% of cases.6

Lone AF has no known precipitatingcause, and patients with this conditionhave no clinical or echocardiographicevidence of cardiopulmonarydisease. Patients with lone AF who areyounger than 60 to 65 years and do nothave hypertension are at low risk forthromboembolism.

Comorbid conditions. Althoughisolated AF can occur, the arrhythmiais usually associated with other conditions(Table 1). The FraminghamStudy revealed several independentrisk factors for AF, including age, congestiveheart failure, systemic hypertension,coronary artery disease, anddiabetes mellitus.4 A history of valvularheart disease and increased left atrialdimensions are among the strongestpredictors of AF.7 A study of echocardiographicpredictors of AF suggestedthat increased left atrial dimensions,left ventricular wall thickness, and reducedleft ventricular fractional shorteningsignificantly heighten the risk ofnonrheumatic AF.8

 Table 1 –Causes of atrial fibrillation



Pathophysiology. AF is thoughtto result from the continuous electricalactivity of multiple wavelets ofreentry that arc around the atrium.9For propagation of AF to occur, a criticalmass of excitable tissue must bepresent, and reentrant wavelets mustnot encounter refractory tissue remainingfrom a previous wavelet.Therefore, a short refractory periodfor the atrial myocardium or slow conductiontime (which allows atrial myocardialtissue to recover) enhancesthe propagation of AF.10

The atrial enlargement that occurswith valvular heart disease, ischemicheart disease, cardiomyopathy,and AF itself promotes persistence ofAF by accommodating more wavelengths.11 Atrial refractory periods mayshorten in the case of increased vagalor sympathetic tone, thyrotoxicosis, orelectrical remodeling that is further enhancedby persistence of this arrhythmia.Myocardial ischemia, fibrosis orinflammation, electrolyte abnormalities,and increased vagal tone may alsoenhance AF propagation by decreasingconduction velocity in atrial myocardialtissue.10

Associated risks. AF is associatedwith significant morbidity12 andmortality13-even in patients who donot have organic heart disease--principallybecause of the increased incidenceof thromboembolic events.These events are thought to result primarilyfrom stasis of blood within theatria (caused by loss of atrial contractility)and subsequent thrombus formation.Left heart emboli may result instroke, myocardial infarction (MI), orother systemic embolic phenomena.Moreover, AF may itself confer a hypercoagulablestate, as indicated by elevatedplasma concentrations of fibrin,D-dimer, and β-thromboglobulin.14

Framingham data show that nonrheumaticAF increases the overall riskof stroke from 1% to approximately 5%per year. Additional risk factors forstroke include transient ischemic attack(TIA) or previous stroke, hypertension,diabetes, and advancing age.15Patients with mitral stenosis and AFhave a 17-fold risk of stroke comparedwith the general population.16 In the Framingham cohort, not only didstroke occur more commonly in patientswith AF, but AF-associated strokewas twice as likely to be fatal and survivorsexperienced more recurrencesand greater functional deficits.17

Morbidity may also result fromthe hemodynamic compromise associatedwith AF. Loss of "atrial kick" mayreduce cardiac output by up to 30%. Inaddition, the rapid heart rates that areoften associated with new-onset AFshorten the period of diastolic fillingand further reduce cardiac output.18

Thus, patients with a history ofcardiac disease, such as MI or left ventriculardysfunction (systolic or diastolic),are at particular risk for hemodynamicdecompensation if subsequentAF develops.19 Morbidity mayalso result from a persistently rapidheart rate, which can lead to tachycardia-induced cardiomyopathy20 and atrialenlargement.21,22 These conditionsmay further predispose to the persistenceof AF through electrical remodelingand fibrosis. Finally, morbiditymay result from pharmacologic treatmentof AF.

The initial presentation of AFvaries considerably according to theventricular rate and severity of anyconcomitant cardiovascular or pulmonarydisease. Patients may haveminimal or no symptoms, or nonspecificmanifestations such as palpitations,decreased exercise tolerance, orfatigue of recent onset.

A study of paroxysmal AFshowed that patients were generallyasymptomatic during these episodes.23Patients with underlying coronaryartery disease may experience new oraccelerated anginal symptoms, andthose with heart failure may havesymptoms of fluid overload or frankpulmonary edema. Stroke may sometimesbe the initial presentation in anotherwise asymptomatic patient. Onestudy of hospitalized patients with AFfound that among those with symptoms,52% had dyspnea, 34% had chestpain, 26% had palpitations, and 19% haddizziness or syncope.24 In the unusualscenario of Wolff-Parkinson-White syndromewith AF, the accessory pathwayis likely to transmit the rapid atrial signalsto the ventricle, which may resultin exceedingly rapid ventricular ratesand risk of deterioration to ventricularfibrillation.

ACHIEVINGRATE CONTROLImmediate. The initial approachto AF is similar to that of any otheracute cardiac arrhythmia. Patientswho are hemodynamically unstable(eg, who have low blood pressure or adecreased level of consciousness) as aresult of a rapid ventricular responseto AF should undergo immediate cardioversion,initially with a 200-J synchronizedshock. Most patients whopresent with AF are hemodynamicallystable; rate control is therefore theprimary short-term treatment objective.25 Once hemodynamic stability isassured, the goals of management areto determine the underlying or precipitatingcauses (see Table 1), relievesymptoms, and prevent thromboemboliccomplications.


Two strategies are appropriatefor long-term treatment ofAF:

  • Ventricular rate control plus anticoagulationtherapy.
  • Restoration and maintenance of normalsinus rhythm.

Although the second optionseems preferable, the drugs that maybe necessary to accomplish this goalall have the potential for significantmorbidity and even mortality.

Three multicenter randomized trialsnow under way will compare ratecontrol with rhythm control: Atrial FibrillationFollow-up Investigation ofRhythm Management (AFFIRM),Rate Control Versus Electrical Cardioversion(RATE), and PharmacologicIntervention in Atrial Fibrillation(PIAF).26,27 Results of these studiesshould provide guidance in choosingbetween the 2 strategies. Until then,the choice must be made on an individualbasis.

Pharmacologic therapy. The rapid,irregular ventricular rate usually associatedwith new-onset AF is responsiblefor many of its symptoms. Thiscondition may also aggravate ventriculardysfunction or cause a tachycardiainducedcardiomyopathy.15,28

Digoxin, calcium channel antagonists,and β-blockers--which all workby slowing atrioventricular (AV) nodalconduction--can be used for immediateand long-term rate control. Until recently,digoxin had been consideredstandard first-line therapy for the treatmentof new-onset rapid AF. This drugworks primarily through a vagotonicmechanism to slow the ventricular response.Digoxin is likely to be most effectivefor patients with AF and heartfailure.

Digoxin has a relatively slow onsetof action (30 minutes to 2 hours),and it is often clinically ineffective forrate control in settings in which vagaltone is low and/or endogenous catecholaminelevels are high, such as inexercise-induced AF, thyrotoxicosis, oracute illness.29 Contrary to a previouslyheld belief, digoxin does not convertAF to sinus rhythm more effectivelythan placebo.30 Given these drawbacks,many clinicians now select alternativeagents for initial rate control.

The calcium channel antagonistsdiltiazem and verapamil are both veryeffective for overall rate control at restand during exercise. In patients whodo not have underlying sick sinus syndrome,these drugs usually do notcause significant bradycardia at rest.They have a rapid onset of action (usually2 to 3 minutes) when given intravenously.Diltiazem is less likely tocause hypotension than verapamil andhas a less negative inotropic effect. Inaddition, it can be carefully titratedwith a continuous intravenous infusion.

Despite its negative inotropic effect,diltiazem is effective and has agood safety profile even when administeredto patients with mild to moderateheart failure.31 Verapamil, and to alesser extent diltiazem, raises serumdigoxin levels, so levels should bemonitored closely when these agentsare used in combination with digoxin.

Some clinicians favor β-blockersas first-line therapy for initial rate control.Use of these agents makes physiologicsense, because they not onlyblock AV nodal conduction but alsocounteract increased sympathetic activity.Like calcium channel antagonists,they are highly effective and featurerapid onset of action (usually within5 minutes) and easy titratabilitywhen given as continuous intravenousinfusions. They may be particularlyuseful for patients whose heart rate increasessubstantially with activity orstress.32

Use β-blockers cautiously in patientswith reactive airways disease, diabetes,hypotension, or systolic heartfailure, because these agents may aggravateor mask these conditions. Allβ-blockers are equally effective for ratecontrol, but certain ones may offer advantagesin specific clinical settings.For example, β-blockers with intrinsicsympathomimetic activity, such as pindololor acebutolol, may be desirablefor patients in whom adequate ratecontrol cannot otherwise be achievedwithout excessive bradycardia.33

β-Blockers and calcium channelantagonists may also be used in combinationwith digoxin if either agent isineffective alone. Although the combi-nation of β-blockers and calcium channelantagonists is usually safe in patientswho do not have sick sinus syndrome,caution is warranted becauseof the potential risk of excessive bradycardia.Avoid concomitant use of intravenousβ-blockers and calcium channelantagonists because of the risk ofsevere bradycardia, heart block, oreven asystole.

If rate control is not achievedafter usual therapy, several otheragents may be considered. Magnesiumsulfate (1 to 4 g infused over severalhours) may enhance the effect ofother drugs.34,35 Candidates for magnesiuminfusion include patients at riskfor magnesium depletion, such asthose with other electrolyte abnormalities(hypokalemia, hyponatremia,hypocalcemia, or hypophosphatemia),those undergoing cancer chemotherapywith cisplatin or similar agents,those undergoing diuretic therapy, andthose with a history of alcoholism oracute MI. Intravenous amiodaronemay also be considered, because itprolongs AV nodal refractory time.

Radiofrequency ablation. If ratecontrol is not achieved despitemaximum medical therapy, considerreferral to a cardiac electrophysiologistfor more invasive alternative treatments.Both AV node modificationand AV node ablation with permanentpacemaker implantation can achievelong-term rate control.36-38 These procedures,which use radiofrequencyenergy, are performed with the patientunder conscious sedation inthe electrophysiology laboratory.Major complication rates of lessthan 3% have been reported.36 Asmall risk of torsades de pointes existswith these procedures; hospital monitoringfor 48 to 72 hours is thereforerecommended.36,39

Many patients who undergo AVnode ablation with pacemaker implantationfeel much better after the procedure.Although long-term anticoagulationis still needed (because theatria are still fibrillating), left ventricularfunction often improves significantlybecause heart rate can now becontrolled.

 Table 2 – Recommendations for ischemic stroke prevention in patients with atrial fibrillation
For patients younger than 60 years*: 

For patients aged 60 years or older†: 

AF, atrial fibrillation; MI, myocardial infarction; CHF, congestive heart failure; TIA, transient ischemic attack; CVA, cerebrovascular accident. *Younger than 65 years, according to some experts. †Aged 65 years or older, according to some experts. Adapted from Albers GW et al. Chest. 200152; Fuster V et al. Circulation. 2001.

Rate control through pharmacologicor nonpharmacologic treatmentusually improves symptoms and mayprevent or at least partially reversetachycardia-induced cardiomyopathy.Unfortunately, it does not prevent ischemicstroke; hence the necessity forlong-term anticoagulation with warfarinin patients for whom the benefitsof this therapy outweigh the risks.

Warfarin. In 5 randomized, controlledprimary prevention trials, warfarinreduced the risk of stroke by 37%to 86%.40-44 Pooled data from these trialsyielded an average stroke risk reductionof 68%; the reduction was greaterin women than in men (84% vs 60%).12Risk reduction was noted across allage groups studied and was independentof such risk factors as MI, con-gestive heart failure, hypertension, diabetes,and previous TIA or stroke.Risk reduction was not seen in patientswith lone AF; however, very fewstrokes occurred in this group. Warfarinalso reduced overall mortality by31%.41,43,44 Warfarin may be even moreeffective for secondary prevention ofstroke in patients with AF and a previouscerebral ischemic event.45

The target INR in the randomizedtrials ranged from 1.4 to 4.5; most embolicevents occurred with INRs lowerthan 2, and most hemorrhagic eventsoccurred with INRs higher than 4.46 Intracranialhemorrhage occurred at anaverage rate of 1.3% per year.

Aspirin. The role of aspirin in preventingstroke related to nonrheumaticAF is less clear. Aspirin was comparedwith both placebo and warfarinin several of the large anticoagulationtrials.40,43,46,47 Pooled data from 3 studiesshow a 21% risk reduction with aspirintherapy--much lower than thatassociated with warfarin. However, inthe Stroke Prevention in Atrial FibrillationII study, the incidence of intracerebralhemorrhage was significantlyhigher in patients older than 75 yearswho received warfarin than in thosewho received aspirin.48

The combination of low-dose warfarin(INR, 1.2 to 1.5) and aspirin (325mg) has not proved effective for reducingischemic stroke or intracerebralhemorrhage.49

Recommendations. Warfarin isclearly more effective than aspirin forstroke risk reduction and should bestrongly considered for all patients (ifno contraindication exists) at high riskfor embolic stroke from AF. Recentmeta-analyses confirm that warfarinprovides greatest benefit for those athighest risk for stroke and that thisbenefit is not offset by the occurrenceof major hemorrhage.50,51

Those at low risk for stroke mayderive sufficient protection from aspirin.50 (The slightly greater preventionbenefit provided by warfarin is offsetby the greater risk of hemorrhage inthis low-risk population.) Practicallyspeaking, patients at low risk forstroke are those who have true loneAF (as defined by the absence of anyevidence of heart disease--includinghypertension--in patients youngerthan 60 to 65 years).

The American College of ChestPhysicians52 and the American HeartAssociation53 have developed guidelinesfor the use of warfarin and aspirinto prevent stroke in patients with nonrheumaticAF. These guidelines arebased on anticoagulation trials and expertopinion; the major recommendationsare listed in Table 2.


REFERENCES:1. Wolf PA, Benjamin EJ, Belanger AJ, et al. Seculartrends in the prevalence of atrial fibrillation: TheFramingham Study. Am Heart J. 1996;131:790-795.
2. Feinberg WM, Blackshear JL, Laupacis A, et al.Prevalence, age distribution, and gender of patientswith atrial fibrillation: analysis and implications.Arch Intern Med. 1995;155:469-473.
3. Phillips SJ, Whisnant JP, O’Fallon WM, Frye RL.Prevalence of cardiovascular disease and diabetesmellitus in residents of Rochester, Minnesota. MayoClin Proc. 1990;65:344-359.
4. Benjamin EJ, Levy D, Vaziri SM, et al, for theFramingham Heart Study. Independent risk factorsfor atrial fibrillation in a population-based cohort.JAMA. 1994;271:840-844.
5. Narayan SM, Cain ME, Smith JM. Atrial fibrillation.Lancet. 1997;350:943-950.
6. Takahashi N, Seki A, Imataka K, Fujii J. Clinicalfeatures of paroxysmal atrial fibrillation. An observationof 94 patients. Jpn Heart J. 1981;22:143-149.
7. Psaty BM, Manolio TA, Kuller LH, et al. Incidenceof and risk factors for atrial fibrillation inolder adults. Circulation. 1997;96:2455-2461.
8. Vaziri SM, Larson MG, Benjamin EJ, Levy D, forthe Framingham Heart Study. Echocardiographicpredictors of nonrheumatic atrial fibrillation. Circulation.1994;89:724-730.
9. Allessie MA, Konings K, Kirchhof CJ, Wijffels M.Electrophysiologic mechanism of perpetuation ofatrial fibrillation. Am J Cardiol. 1996;77:10A-23A.
10. Dell'Orfano JT, Luck JC, Wolbrette DL, et al.Drugs for conversion of atrial fibrillation. Am FamPhysician. 1998;58:471-480.
11. Flaker GC, Fletcher KA, Rothbart RM, et al.Clinical and echocardiographic features of intermittentatrial fibrillation that predict recurrent atrialfibrillation. Am J Cardiol. 1995;76:355-358.
12. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillationas an independent risk factor for stroke: theFramingham Study. Stroke. 1991;22:983-988.
13. Benjamin EJ, Wolf PA, D’Agostino RB, et al,for the Framingham Heart Study. Impact of atrialfibrillation on the risk of death. Circulation. 1998;98:946-952.
14. Lip GY, Lip PL, Zarifis J, et al. Fibrin D-dimerand beta-thromboglobulin as markers of thrombogenesisand platelet activation in atrial fibrillation:effects of introduction of ultra–low-dose warfarinand aspirin. Circulation. 1996;94:425-431.
15. Atrial Fibrillation Investigators. Risk factorsfor stroke and efficacy of antithrombotic therapy inatrial fibrillation. Analysis of pooled data from 5 randomizedcontrolled trials. Arch Intern Med. 1994;154:1449-1457.
16. Wolf PA, Kannel WB, McGee DL, et al. Durationof atrial fibrillation and imminence of stroke:the Framingham study. Stroke. 1983;14:664-667.
17. Lin HJ, Wolf PA, Kelly-Hayes M, et al. Strokeseverity in atrial fibrillation: the Framingham Study.Stroke. 1996;27:1760-1764.
18. Daoud EG, Weiss R, Bahu M, et al. Effect of anirregular ventricular rhythm on cardiac output.Am J Cardiol. 1996;78:1433-1436.
19. Tischler MD, Lee TH, McAndrew KA, et al.Clinical, echocardiographic and Doppler correlatesof clinical instability with onset of atrial fibrillation.Am J Cardiol. 1990;66:721-724.
20. Catherwood E, Fitzpatrick WD, Greenberg ML,et al. Cost-effectiveness of cardioversion antiarrhythmictherapy in nonvalvular atrial fibrillation.Ann Intern Med. 1999;130:625-636.
21. Sanfilippo AJ, Abascal VM, Sheehan M, et al.Atrial enlargement as a consequence of atrial fibrillation:a prospective echocardiographic study. Circulation.1990;82:792-797.
22. Suarez GS, Lampert S, Ravid S, Lown B.Changes in left atrial size in patients with lone atrialfibrillation. Clin Cardiol. 1991;14:652-656.
23. Page RL, Wilkinson WE, Clair WK, et al. Asymptomatic arrhythmias in patients with symptomaticparoxysmal atrial fibrillation and paroxysmalsupraventricular tachycardia. Circulation. 1994;89:224-227.
24. Lip GY, Tean KN, Dunn FG. Treatment of atrialfibrillation in a district general hospital. Br Heart J.1994;71:92-95.
25. Nattel S. Newer developments in the managementof atrial fibrillation. Am Heart J. 1995;130:1094-1106.
26. Atrial Fibrillation Follow-up Investigation ofRhythm Management: the AFFIRM study design.Am J Cardiol. 1997;79:1198-1202.
27. Waldo AL. Management of AF: the need forAFFIRMative action. AFFIRM investigators: AtrialFibrillation Follow-up Investigation of Rhythm Management.Am J Cardiol. 1999;84:698-700.
28. Peters KG, Kienzle MG. Severe cardiomyopathydue to chronic rapidly conducted atrial fibrillation:complete recovery after restoration of sinusrhythm. Am J Med. 1988;85:242-244.
29. Falk RH, Leavitt JI. Digoxin for atrial fibrillation:a drug whose time has gone? Ann Intern Med.1991;114:573-575.
30. Falk RH, Knowlton AA, Bernard SA, et al.Digoxin for converting recent-onset atrial fibrillationto sinus rhythm: a randomized, double-blinded trial.Ann Intern Med. 1987;106:503-506.
31. Goldenberg IF, Lewis WR, Dias VC, et al. Intravenousdiltiazem for the treatment of patients withatrial fibrillation or flutter and moderate to severecongestive heart failure. Am J Cardiol. 1994;74:884-889.
32. Falk RH. Pharmacologic control of heart rate inatrial fibrillation. Cardiol Clin. 1996;14:521-536.33. Channer KS, James MA, MacConnell T, ReesJR. Beta-adrenoceptor blockers in atrial fibrillation:the importance of partial agonist activity. Br J ClinPharmacol. 1994;37:53-57.
34. Hays JV, Gilman JK, Rubal BJ. Effect of magnesiumsulfate on ventricular rate control in atrialfibrillation. Ann Emerg Med. 1994;24:61-64.
35. Brodsky MA, Orlov MV, Capparelli EV, et al.Magnesium therapy in new-onset atrial fibrillation.Am J Cardiol. 1994;73:1227-1229.
36. Morady F, Hasse C, Strickberger SA, et al.Long-term follow-up after radiofrequency modificationof the atrioventricular node in patients with atrialfibrillation. J Am Coll Cardiol. 1997;29:113-121.
37. Feld GK. Radiofrequency catheter ablation versusmodification of the AV node for control of rapidventricular response in atrial fibrillation. J CardiovascElectrophysiol. 1995;6:217-228.
38. Wood MA, Brown-Mahoney C, Kay GN, EllenbogenKA. Clinical outcomes after ablation and pacing therapy for atrial fibrillation: a meta-analysis.Circulation. 2000;101:1138-1144.
39. Langberg JJ, Chin MC, Rosenqvist M, et al.Catheter ablation of the atrioventricular junctionwith radiofrequency energy. Circulation. 1989;80:1527-1535.
40. Petersen P, Boysen G, Godtfredsen J, et al, forthe Copenhagen AFASAK study. Placebo controlled,randomised trial of warfarin and aspirin for preventionof thromboembolic complications in chronicatrial fibrillation. Lancet. 1989;1:175-179.
41. Boston Area Anticoagulation Trial for AtrialFibrillation Investigators. The effect of low-dosewarfarin on the risk of stroke in patients with nonrheumaticatrial fibrillation. N Engl J Med. 1990;323:1505-1511.
42. Connolly SJ, Laupacis A, Gent M, et al. CanadianAtrial Fibrillation Anticoagulation (CAFA) Study.J Am Coll Cardiol. 1991;18:349-355.
43. Stroke Prevention in Atrial Fibrillation Investigators.Stroke prevention in atrial fibrillation study:final results. Circulation. 1991;84:527-539.
44. Ezekowitz MD, Bridgers SL, James KE, et al.Warfarin in the prevention of stroke associated withnonrheumatic atrial fibrillation. N Engl J Med. 1992;327:1406-1412.
45. Hylek EM, Skates SJ, Sheehan MA, Singer DE.An analysis of the lowest effective intensity of prophylacticanticoagulation for patients with nonrheumaticatrial fibrillation. N Engl J Med. 1996;335:540-546.
46. European Atrial Fibrillation Trial Study Group.Optimal oral anticoagulant therapy in patients withnonrheumatic atrial fibrillation and recent cerebralischemia. N Engl J Med. 1995;333:5-10.
47. The Atrial Fibrillation Investigators. The efficacyof aspirin in patients with atrial fibrillation:analysis of pooled data from 3 randomized trials.Arch Intern Med. 1997;157:1237-1240.
48. Stroke Prevention in Atrial Fibrillation Investigators.Warfarin versus aspirin for prevention ofthromboembolism in atrial fibrillation: Stroke Preventionin Atrial Fibrillation II Study. Lancet. 1994;343:687-691.
49. The Atrial Fibrillation Investigators. Adjusteddosewarfarin versus low intensity fixed dose warfarinplus aspirin for high risk patients with atrialfibrillation: Stroke Prevention in Atrial Fibrillation IIIrandomised clinical trial. Lancet. 1996;348:633-638.
50. Ezekowitz MD, Levine JA. Preventing strokein patients with atrial fibrillation. JAMA. 1999;281:1830-1835.
51. Hart RG, Benavente O, McBride R, Pearce LA.Antithrombotic therapy to prevent stroke in patientswith atrial fibrillation: a meta-analysis. Ann InternMed. 1999;131:492-501.
52. Albers GW, Dalen JE, Laupacis A, et al. Antithrombotictherapy in atrial fibrillation. Chest. 2001;119(suppl):194S-206S.
53. Fuster V, Ryden LE, Asinger RW, et al.ACC/AHA/ESC guidelines for the management ofpatients with atrial fibrillation: executive summary.Circulation. 2001;104:2118-2150.