Editorial. Atrial Fibrillation and Stroke

Transcription

1 Editorial 1337 Atrial Fibrillation and Stroke Revisiting the Dilemmas Robert G. Hart, MD; Jonathan L. Halperin, MD A trial fibrillation is a common cardiac dysrhythmia / \ of the elderly, and stroke is its most devastating JL ^ complication. Six years ago in Stroke, we discussed controversies surrounding the prevention of stroke in patients with nonvalvular atrial fibrillation (AF). 1 Since then, six recent randomized trials have changed the approach to stroke prevention for AF patients. 2-3 Above all else, the randomized trials have confirmed the magnitude of the problem: AF raises the risk of stroke in more than 1.5 million Americans to a degree approaching that which follows transient ischemic attack (TLA). The rate of ischemic stroke among people with AF averages 5%/y, approximately six times that of people without AF Additionally, considering TTAs (often causing radiographic evidence of brain infarction) and clinically occult "silent" strokes detected radiographicalry, the rate of brain ischemia accompanying nonvalvular AF exceeds 7%/y The absolute rate of stroke in AF patients varies importantly with coexistent cardiovascular disease, and therefore stratification of AF patients into those at high and low risk of thromboembolism has become a crucial determinant of optimal antithrombotic prophylaxis. AF increases in prevalence and risk with age, and more than half of AF-associated strokes occur in patients aged older than 75 years Special consideration of these older patients is therefore a critical aspect of an effective stroke prevention strategy. Most ischemic strokes associated with AF are probably due to embolism of stasis-induced thrombi forming in the left atrium and particularly its appendage. Transesophageal echocardiography shows left atrial thrombi more frequently in AF patients with ischemic stroke than in AF patients without stroke. 16 Nevertheless, perhaps 25% of AF-associated strokes are due to intrinsic cerebrovascular diseases, other cardiac sources of embolism, or atheromatous pathology in the proximal aorta Approximately half of elderly AF patients have chronic hypertension (a major risk factor for cerebrovascular disease), 2 and approximately 12% of elderly AF patients harbor cervical carotid artery stenosis. 19 But the frequency of carotid stenosis is not substantially greater in AF patients with stroke, and carotid stenosis is a minor Received March 29, 1994; accepted April 20, From the Department of Medicine (Neurology), University of Texas Health Science Center, San Antonio (R.G.H.), and the Division of Cardiology, Mount Sinai Medical Center, New York, NY (J.L.H.). Correspondence to Robert G. Hart, MD, Department of Medicine (Neurology), University of Texas Health Science Center, 7703 Floyd Curl Dr, San Antonio, TX 782S American Heart Association, Inc. contributing factor to the increased stroke risk associated with AF. 19 Although atherosclerotic carotid lesions may contribute to stroke risk through thromboembolic mechanisms unrelated to the severity of luminal stenosis, the available data do not support routine screening of AF patients without symptoms of brain ischemia for cervical carotid stenosis. 19 Subgroups of AF Patients With High and Low Rates of Stroke Identification of subpopulations with relatively high or low absolute rates of stroke determines which AF patients gain the greatest benefit from long-term anticoagulation-with warfarin, offsetting its risk, expense, and inconvenience. Because clinical classification of ischemic strokes according to etiologic subtypes is imperfect and not adequately validated, 17 risk stratification is presently based on combined analysis of all ischemic strokes. Schemes for stratification of risk in AF patients have been extensively pursued, using clinical and echocardiographic parameters Two prospective studies encompass sufficient numbers of patients and stroke events for meaningful multivariate analysis, and these provide the most reliable basis for risk stratification available (Table I). 2-1 "- 11 The differences in the findings (involving age and congestive heart failure) are not substantially conflicting, as clinical variables frequently overlap (heart failure, hypertension, and diabetes are all more frequent in older patients). Intermittent (or paroxysmal) AF does not appear to be an independent predictor of thromboembolic risk Other studies support the notion that stroke in AF patients is directly related to coexistent heart disease, hypertension, age, and perhaps female sex Still missing from these stratification models are firm links to stroke mechanism and cohesive pathophysiological correlations. Echocardiographic correlates of thromboembolic risk in AF are enlargement of the left atrium (mitigated by mitral regurgitation) and impairment of left ventricular systolic function, which may promote stasis of blood within the left atrium in the presence of AF Precordial echocardiographic findings can be combined with clinical risk stratifiers to identify AF patients without risk factors, who are at very low inherent risk of thromboembolism (Table I). 11 Transesophageal echocardiography offers better visualization of the left atrium and its appendage than precordial echocardiography and more often identifies atrial thrombi and spontaneous echogenic contrast ("smoke" probably indicative of stasis) in AF patients

2 1338 Stroke Vol 25, No 7 July 1994 TABLE 1. Risk Stratification In Atrial Fibrillation*: Independent Predictors of Thromboembollc Risk No. of patients No. of events High-risk variables Thromboembolic rate (95% Cl) SPAF-I Placebo" Prior stroke/tta Diabetes Recent heart failure AFI Pooled Analysis Prior stroke/tia Diabetes Age >65 years SPAF Echo Analysis Prior stroke/tia Recent heart failure LA >2.5 cm/m 2 LV dysfunction Low-risk variables High-risk variables Percentage at low risk 1.4%/y ( ) >7%/y %/y ( ) >5%/y %/y ( ) >6%/y 26 SPAF indicates Stroke Prevention in Atrial Fibrillation study; AFI, Atrial Fibrillation Investigators; Cl, confidence interval; Hx, history; TIA, transient ischemic attack; LA, left atrial size by M-mode echocardiography; and LV dysfunction, left ventricular dysfunction by two-dimensional echocardiography. *Prospectively acquired data analyzed by multivariable techniques. The SPAF-I placebo data set 10 is included In the pooled analysis of clinical trials by the AFI. 2 with thromboembolism than in those without thromboembolism. 16 The sensitivity and specificity of this technique for detection of small thrombi in the atrial appendage are unclear. The predictive value of these findings for subsequent stroke has yet to be validated by adequate clinical studies, and available data are insufficient to justify routine transesophageal echocardiography to stratify thromboembolic potential in our view. A large multicenter clinical trial (Stroke Prevention in Atrial Fibrillation [SPAF] III) is evaluating the role of transesophageal echocardiography in AF. 28 Antithrombotic Therapy for Stroke Prevention Anticoagulation with the oral vitamin K antagonist warfarin is highly effective for reducing ischemic stroke in AF patients. Five recent randomized clinical trials using prothrombin time international normalized ratio (INR) ranges between approximately 1.8 and 4.2 showed a mean reduction in ischemic stroke of nearly 70% in patients assigned anticoagulation (Fig 1); ontherapy analysis indicated an even greater benefit. 2 The incremental risk of severe bleeding was less than 1%/y among anticoagulated patients, a selected group followed up carefully according to clinical trial protocols. Whether such low bleeding risks can be achieved in general clinical practice is an important, unresolved issue. 29 Safe anticoagulation monitoring requires use of the INR, which corrects for variable thromboplastin sensitivities. Low-intensity anticoagulation (INR range, 2 to 3) clearly confers benefit, 2-30 and there is no doubt that warfarin is highly effective in subgroups of AF patients who carry a high inherent risk of thromboembolism (Table I)."- 31 The safety and tolerability of long-term anticoagulation to conventional levels has not been completely defined among patients older than 75 years, who account for more than half of AF-associated strokes. All but one of the placebo-controlled anticoagulation trials enrolled AF patients who were, on average, substantially younger. 2 The single placebo-controlled trial involving AF patients with a mean age of 75 years reported an overall withdrawal rate from anticoagulation of 38% after 1 year, and the attrition rate among the oldest patients was probably even higher. 32 The risk of major hemorrhage during anticoagulation (INR range, 2.0 to 4.5; mean, 2.7) in another trial was substantially greater among AF patients older than 75 years than among younger patients anticoagulated to similar intensities. 31 Although the elderly have a greater risk of AF-associated stroke, the benefit of anticoagulation is offset somewhat by this greater toxicity, and the AFASAK BAATAF CAFA SPAF SPiNAF Combined 68% (50-79%, p<.001) 100% 50% 0% -50% Warfarin better Warfarin worse Risk Reduction Fra 1. Risk-reduction plot of the results of five randomized clinical trials comparing warfarin with control for prevention of ischemic stroke in atrial fibrillation patients. 2 Horizontal lines indicate the 95% confidence intervals around the point estimates (vertical lines) for each trial. The combined risk reduction was 68% (95% confidence interval, 50% to 79%; P<.001). See Atrial Fibrillation Investigators' pooled analysis 2 for specific data. AFASAK indicates Atrial Fibrillation, Aspirin, and Anticoagulant Therapy study; BAATAF, Boston Area Anticoagulation Trial for Atrial Fibrillation; CAFA, Canadian Atrial Fibrillation Anticoagulation; SPAF, Stroke Prevention in Atrial Fibrillation study; and SPINAF, Stroke Prevention in Nonrheumatic Atrial Fibrillation study.

3 Hart and Halperin Atrial Fibrillation and Stroke 1339 Combined 100% 50% 0% -50% -100% Warfarin better Aspirin better Risk Reduction FIG 2. Risk-reduction plot of the results of three randomized clinical trials comparing warfarin with aspirin for prevention of ischemic stroke in atrial fibrillation patients & Horizontal lines indicate the 95% confidence intervals around the point estimates (vertical lines) for each trial. The combined estimated risk reduction from published results (intent-to-treat) was 47% (95% confidence interval, 28% to 61%) by warfarin relative to aspirin. SPAF indicates Stroke Prevention in Atrial Fibrillation study; EAFT, European Atrial Fibrillation Trial; and AFASAK, Atrial Fibrillation, Aspirin, and Anticoagulant Therapy study. inability of the very elderly to sustain long-term anticoagulation is a noteworthy finding of several trials Intracerebral hemorrhage is the most feared and most often fatal complication of warfarin in elderly patients, occurring in 0.3%/y of AF patients given warfarin in recent clinical trials. 2 Predictors and precipitants of warfarin-related intracerebral bleeding have not been consistently identified in the literature. In one study the rate of intracerebral bleeding was 1.5%/y in AF patients with a mean age of 80 years and with a mean INR of 2.6, and advanced age was predictive of its occurrence. 31 The rate of intracerebral hemorrhage in AF patients older than 75 years was substantially lower in combined analysis of other clinical trials, in which anticoagulation intensity was generally lower Although it is uncertain whether the intensity of anticoagulation is related to intracerebral bleeding, 33 a lower intensity of anticoagulation (INR approximately 2) may be sensible in older AF patients until ongoing trials addressing alternatives are completed. The efficacy of aspirin, a platelet inhibitory agent, for stroke prevention in AF patients is less clear and more controversial. 2 * 34 The effect of aspirin in doses between 75 and 325 mg/d has been assessed in three placebocontrolled trials with a statistically significant pooled risk reduction of approximately 25% (range, 14% to 44%) in aspirin-treated patients Aspirin was significantly less effective than warfarin in two of these trials analyzed along intention-to-treat paradigms 3-35 and according to on-treatment (efficacy) analysis of the third (Fig 2). 31 There is no compelling evidence that a specific dose of aspirin between 75 and 325 mg/d confers more or less benefit. Aspirin thus appears active for preventing stroke in AF patients but substantially less effective than warfarin. Aspirin is inexpensive, easily administered, and produces less bleeding toxicity than anticoagulation in elderly patients. 31 Selecting Optimal Antithrombotic Therapy to Prevent Stroke Platelet inhibitors and anticoagulants may influence cardioembolic and noncardioembolic sources of stroke TABLE 2. Rates of Thromboembollsm During Aspirin Therapy in High- and Low-Risk Patients: Results of the SPAF II Study Patients s75 years old % of cohort Event rate, % (95% Cl) Patients >75 years old % of cohort Event rate, % (95% Cl) Clinical Risk Factors for Thromboflmbollsm* Yes No ( ) 0.5 ( ) ( ) 1.8 ( ) SPAF indicates Stroke Prevention in Atrial Fibrillation; Cl, confidence interval. *One or more of a history of hypertension, prior thromboembolism, or recent (within 3 months) congestive heart failure differently. 17 Aspirin may have a greater prophylactic impact on noncardioembolic mechanisms than on strokes of presumed cardioembolic origin in AF patients. 17 Aspirin is particularly effective in younger AF patients with a history of diastolic hypertension. 3 * The efficacy of platelet inhibitors for reducing venous thrombosis in the lower extremities (a situation in which stasis is a major contributing factor) has been suggested by a recent meta-analysis. 37 Most AF-related stroke, particularly in elderly women, seems attributable to cardiogenic embolism, and this type of stroke is more effectively prevented by anticoagulation. 38 AF patients with high rates of stroke during aspirin therapy are typically older women and patients with impaired left ventricular function, where cardiogenic embolism is the predominant mechanism of cerebral infarction These pathophysiological constructs nevertheless require further study before they can be confidently applied to patient management. The risk of thromboembolism in AF patients aged 75 years or younger given aspirin is relatively low, below 3%/y Among these younger AF patients, those prospectively identified as at lowriskbased on clinical criteria had an extremely low rate of thromboembolism averaging 0.5%/y (95% confidence interval, 0.1%/y to 1.9%/y; Table 2) when treated with aspirin in one study. 31 Pending the results of ongoing confirmatory studies, it is our view that low-risk AF patients can be treated with aspirin 325 mg/d to prevent stroke, but they must be observed carefully for the development of risk factors for thromboembolism (including systolic hypertension greater than 160 mm Hg, episodes of minor cerebral ischemia, and changes in ventricular function) High-risk patients deemed safe candidates for anticoagulation should be treated with warfarin. For high-risk AF patients aged 75 years or younger, an INR range of 2.0 to 3.0 is effective and fairly safe; for those aged older than 75 years, a lower target INR of 2.0 seems more sensible. AF patients who cannot safely take warfarin should be given aspirin; the value of other antiplatelet agents has not been assessed in this situation. 41 Secondary Prevention of Stroke in AF Patients Conventional wisdom holds that most ischemic strokes in AF patients are large and disabling, but it is

4 1340 Stroke Vol 25, No 7 July 1994 now clear that minor stroke and TIA are frequent accompaniments of AF. 2 AF-related stroke carries a particularly high mortality rate, related to advanced patient age and associated heart disease Previously unrecognized AF discovered in a patient with acute stroke might be a consequence of brain infarction mediated by other processes. 43 Recent studies indicate that the risk of early recurrent stroke (within 2 weeks) is lower in AF patients than thought a decade ago. 18 ' 42 Initiating oral anticoagulation within a few days after submassive infarcts is a reasonable practice; delaying warfarin for 1 week or more in AF patients with large infarcts may also be prudent to avoid accentuating secondary brain hemorrhage. Prophylaxis against deep venous thrombosis with low-dose, subcutaneous heparin in those with lower limb paresis is safe. 18 Once ischemic stroke has occurred in a patient with AF, diagnostic evaluation is aimed at identifying those who require chronic anticoagulation for secondary prevention (most patients) or carotid endarterectomy (few patients). Carotid sonography discloses ipsilateral cervical arterial stenosis in approximately 15% of AF patients with stroke. 17 ' 18 Precordial echocardiography is seldom helpful in defining stroke mechanism because left atrial thrombi are not reliably detected. Whether or not carotid artery stenosis or atrial thrombi are identified by ultrasound studies, anticoagulation is warranted for secondary prevention of stroke or TIA in AF patients whenever it can be used safely, because the rate of recurrent stroke is high, exceeding 10%/y without treatment. 3 ' 42 A large randomized trial found warfarin (INR range, 2.5 to 4.0) highly effective, superior to aspirin, and relatively safe. 3 In AF patients with brain ischemia harboring ipsilateral carotid artery stenosis, carotid endarterectomy followed by chronic anticoagulation should be considered. Restoration of Sinus Rhythm to Prevent Stroke There is no solid clinical evidence that cardioversion followed by prolonged maintenance of sinus rhythm reduces thromboembolism in AF patients However, based on the pathophysiological concept that most AF-associated strokes are due to embolism of stasisinduced thrombi from the left atrial appendage, resumption of atrial contractility should logically reduce thromboembolic risk. Atrial enlargement and some perturbation of flow dynamics may persist chronically even after successful cardioversion. Further, AF often recurs after iatrogenic cardioversion, and thromboembolism before recognition of recurrence is a potential problem. Because of these uncertainties, it is unclear whether efforts to restore and maintain sinus rhythm for the purpose of preventing stroke are worthwhile. This choice is influenced by the toxicity of anticoagulation versus antiarrhythmic therapy. Quinidine, the most commonly used antiarrhythmic agent to suppress AF, has been possibly associated with increased mortality, and other antiarrhythmic agents may be preferred. 44 Clinical studies of this crucial issue are needed. 45 At present, the decision to restore sinus rhythm to prevent stroke is complex, fraught with uncertainties, and must consider the likelihood of long-term suppression of AF and the relative toxicity of antithrombotic versus antiarrhythmic drugs. Persisting Management Dilemmas The volume of sound scientific data that have recently emerged regarding prevention of stroke in AF patients represents a considerable research achievement. We know now that the majority of strokes in AF patients can be prevented, more reliably, in fact, than almost any other common type of brain ischemia. AF can today be seen not only as a marker of a patient's risk of stroke but also as a valuable opportunity for effective clinical intervention. Even so, many issues remain unresolved. Lowerintensity regimens of anticoagulation, which should cause less bleeding, and combinations of anticoagulants plus platelet inhibitory agents are under study to establish efficacy. These efforts are particularly important for AF patients aged older than 75 years, for whom conventional anticoagulation may carry substantial toxicity. 31 Better characterization of the role of aspirin and of low-risk patients awaits clarification of stroke mechanisms in AF. The links between risk stratification schemes, pathophysiology, and stroke mechanisms have not been fully elucidated. The value of transesophageal echocardiography for risk stratification and for evaluation of acute stroke patients requires further study. 48 Whether perturbations in intrinsic hemostatic balance cause or contribute to stroke in AF is unknown. Ongoing clinical trials 28 may yet answer these questions during this "decade of the brain." References 1. Halperin JH, Hart RG. Atrial fibrillation and stroke: new ideas, persisting dilemmas. Stroke. 1988;19: Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation: analysis of pooled data from five randomized controlled trials. Arch Intern Med In press. 3. EAFT Study Group. European Atrial Fibrillation Trial: secondary prevention of vascular events in patients with nonrheumatic atrial fibrillation and recent transient ischemic attack or minor ischemic stroke. Lancet 1993;342: Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke. 1991;22: Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation: a major contributor to stroke in the elderly. Arch Intern Med. 1987;147: Stroke Prevention in Atrial Fibrillation Investigators. The Stroke Prevention in Atrial Fibrillation Study: final results. Circulation. 1991;84: Feinberg WM, Seeger IF, Carmody RF, Anderson DC, Hart RG, Pearce LA. Epidemiologic features of asymptomatic cerebral infarction in patients with nonvarvular atrial fibrillation. Arch Intern Med. 1990;15O: Kempster PA, Gerraty RP, Gates PC. Asymptomatic cerebral infarction in patients with chronic atrial fibrillation. Stroke. 1988; 19: Petersen P, Madsen EB, Brun B, Pedersen F, Gyldensted C, Boysen G. Silent cerebral infarction in chronic atrial fibrillation. Stroke. 1987;18: Stroke Prevention in Atrial Fibrillation Investigators. Predictors of thromboembolism in atrial fibrillation, I: clinical features of patients at risk. Ann Intern Med 1992;116:l Stroke Prevention in Atrial Fibrillation Investigators. Predictors of thromboembolism in atrial fibrillation, II: echocardiographic features of patients at risk. Ann Intern Med 1992;] 16: Kopecky SL, Gersh BJ, McGoon MD, Whisnant JP, Holmes DR, Ilstrup DM, Frye RL. The natural history of lone atrial fibrillation: a population-based study over three decades. N EnglJ Med. 1987; 317: Asplund K, Carlberg B, Sundstrom G. Stroke in the elderly. CerebrovascDis. 1992;2:

5 Hart and Halperin Atrial Fibrillation and Stroke Boysen G, Nyboe J, Appleyard M, Sorensen PS, Boas J, Somnier F, Jensen G, Schnohr P. Stroke incidence and risk factors in Copenhagen, Denmark. Stroke. 1988;19: Moulton AW, Singer DE, Haas JS. Risk factors for stroke in patients with nonrheumatic atrial fibrillation: a case-control study. AmJMed. 1991;91: Chimowitz MI, DeGeorgia MA, Poole RM, Hepner A, Armstrong WM. Left atrial spontaneous echo contrast is highly associated with previous stroke in patients with atrial fibrillation. Stroke. 1993;24: Miller VT, Rothrock JF, Pearce LA, Feinberg WM, Hart RG, Anderson DC. Ischemic stroke in patients with atrial fibrillation: effect of aspirin according to stroke mechanism. Neurology. 1993; 43: Bogousslavsky J, Van Melle G, Regli F, Kappenberger L. Pathogenesis of anterior circulation stroke in patients with nonvalvular atrial fibrillation: the Lausanne Stroke Registry. Neurology. 1990; 40: Kanter MC, Tegeler CH, Pearce LA, Weinberger J, Feinberg WM, Anderson DC, Gomez CR, Rothrock JF, Helgason CM, Hart RG, on behalf of the SPAF Investigators. Carotid stenosis in patients with atrial fibrillation: prevalence, risk factors and relationship to stroke. Arch Intern Med. In press. 20. Cabin HS, Clubb KS, Hall C, Perlmutter RA, Feinstein AR. Risk for systemic embolization of atrial fibrillation without mitral stenosis. Am J Cardiol. 1990;65: Flegel KM, Hanley J. Risk factors for stroke and other embolic events in patients with nonrheumatic atrial fibrillation. Stroke. 1989;20: Wiener I. Clinical and echocardiographic correlates of systemic embolization in nonrheumatic atrial fibrillation. Am J Cardiol. 1987;59: Aronow WS, Gutstein H, Hsieh FY. Risk factors for thromboembolic stroke in elderly patients with chronic atrial fibrillation. Am J Cardiol. 1989;63: Petersen P, Kastrup J, Helweg-Larsen S, Boysen G, Godtfredsen J. Risk factors for thromboembolic complications in chronic atrial fibrillation. Arch Intern Med. 1990;150: Blackshear JL, Pearce LA, Asinger RW, Dittrich HC, Goldman ME, Zabalgoitia M, Rothbart RM, Halperin JL. Mitral regurgitation associated with reduced thromboembolic events in high risk patients with atrial fibrillation. J Am Coll Cardiol. 1993;73: Zabalgoitia M, Dipeshkumar KG, McPherson D. Spontaneous echo contrast in severe left ventricular dysfunction: a risk factor for thromboembolism. Circulation. 1990;84(suppl III):III-109. Abstract. 27. Rosenthal MS, Halperin JL. Thromboembolism in nonvalvular atrial fibrillation: the answer may be in the ventricle. Int J Cardiol. 1992;37: Major Ongoing Stroke Trials. Stroke. 1993;24: Lundstrom T, Ryden L. Haemorrhagic and thromboembolic complications in patients with atrial fibrillation on anticoagulant prophylaxis. J Intern Med. 1989;226: Veterans Affairs Stroke Prevention in Nonrheumatic Atrial Fibrillation Investigators. Warfarin in the prevention of stroke associated with nonrheumatic atrial fibrillation, N Engl J Med. 1992; 327: Stroke Prevention in Atrial Fibrillation Investigators. Warfarin versus aspirin for prevention of thromboembolism in atrial fibrillation: Stroke Prevention in Atrial Fibrillation II Study. Lancet. 1994;343: Petersen P, Boysen G, Godtfredsen J, Anderson ED, Andersen B. Placebo-controlled, randomized trial of warfarin and aspirin prevention of thrombo-embolic complications in chronic atrial fibrillation. Lancet. 1989;1: Franke CL, De Jonge J, van Swieten JC, Op de Coul AAW, van Gijn J. Intracerebral hematomas during anticoagulant treatment. Stroke. 1990;21: Singer DE. Aspirin and prevention of stroke. Lancet. 1994;343: Letter. 35. Petersen P, Boysen G. Prevention of stroke in atrial fibrillation. N Engl J Med. 1990;323: Stroke Prevention in Atrial Fibrillation Investigators. A differential effect of aspirin for prevention of stroke in atrial fibrillation. J Stroke Cerebrovasc Dis. 1993;3: Antiplatelet Trialists Collaboration. Collaborative overview of randomized trials of antiplatelet therapy, III: reduction in venous thrombosis and pulmonary embolism by antiplatelet prophylaxis among surgical and medical patients. Br Med J. 1994;308: D'Olhaberriague L, Hernandez-Vidal A, Molina L, Soler-Singla L, Marrugat J, Pons S, Moral A, Pou-Serradel A. A prospective study of atrial fibrillation and stroke. Stroke. 1989;20: McAnulty JH, Hart RG, Pearce LA, for the SPAF Investigators. Atrial fibrillation and thromboembolism: predictors of high risk during aspirin therapy. Circulation. 1993;88(suppl II):II-222. Abstract. 40. Singer DE, Hughes RA, Gress DR, Sheehan MA, Oertel LB, Maraventano SW, Blewett DR, Rosner B, Kistler JP, for the BAATAF Investigators. The effect of aspirin on the risk of stroke in patients with nonrheumatic atrial fibrillation: the BAATAF study. Am Heart J. 1992;124: Morocutti C, Coccheri S. Aspirin and prevention of stroke. Lancet. 1994;343:234. Letter. 42. Sandercock P, Bamford J, Dennis M, Burn J, Slattery J, Jones L, Boonyakarnkul S, Warlow C. Atrial fibrillation and stroke: prevalence in different types of stroke and influence on early and long-term prognosis. BMJ. 1992;305: Vangerhoets F, Bogousslavsky J, Regli F, Van Melle G. Atrial fibrillation after acute stroke. Stroke. 1993;24: Disch DL, Greenberg ML, Holzberger PT, Malenka DJ, Birkmeyer JD. Managing chronic atrial fibrillation: a Markov decision analysis comparing warfarin, quinidine and low-dose amiodarone. Ann Intern Med. 1994;120: National Heart, Lung, and Blood Institute Working Group on Atrial Fibrillation. Atrial fibrillation: current understandings and research imperatives. J Am Coll Cardiol. 1993;22: Coplen SE, Antmann EM, Berlin JA, Hewitt P, Chalmers TC. Efficacy and safety of quinidine therapy for maintenance of sinus rhythm after cardioversion: a meta-analysis of randomized controlled trials. Circulation. 1990;82:l Flaker GC, Blackshear JL, McBride R, Kronmal RA, Halperin JL, Hart RG. Anti-arrhythmic drug therapy and cardiac mortality in atrial fibrillation. J Am Coll Cardiol. 1992;20: Verhorst PMJ, Kamp O, Visser CA, Verheugt FWA. Left atrial appendage flow velocity assessment using transesophageal echocardiography in nonrheumatic atrial fibrillation and systemic embolism. Am J Cardiol. 1993;71: KEY WORDS clinical trials anticoagulants aspirin stroke prevention atrial fibrillation