Long-term Anticoagulation Therapy in Prevention of StrokeDavid Sherman, MD

AddressDivision of Neurology, Department of Medicine, The University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, TX 78229, USA.E-mail: sherman@uthscsa.eduCurrent Treatment Options in Neurology 2002,4:411–416Current Science Inc. ISSN 1092-8480Copyright © 2002 by Current Science Inc.

Introduction

PRIMARY AND SECONDARY PREVENTION OF CARDIOEMBOLIC STROKEThrombi arising in the heart producing emboli to abrain artery account for approximately 20% of ischemicstrokes. The proportion is higher in young patients withstroke due to congenital heart defects, and in olderpatients with ischemic stroke due to the increasingoccurrence of atrial fibrillation with advancing age. In asmany as one third of patients with ischemic stroke,some cardiac abnormality is detected. However, there isoften uncertainty as to whether a particular abnormalitydid in fact produce emboli to the cerebral circulation.With more sensitive and extensive cardiac investiga-tions, such as transesophageal echocardiography, thedetection of unsuspected cardiac abnormalitiesincreases. Cardiac abnormalities can be broadly classi-

fied as major risk or minor risk for embolic stroke(Table 1). The most frequent cardiac abnormality isatrial fibrillation, accounting for almost half of allcardiac causes of stroke (Fig. 1). The diagnosis ofcardiac embolus as the cause of stroke is generally basedon the demonstration of a potential cardiac source ofembolus, and no alternative mechanism for stroke.Although some clinical features are more common witha cardioembolic stroke, none are specific. Patients witha cardioembolic stroke are less likely to have precedingstereotyped transient ischemic attacks, and are some-what more likely to have the abrupt onset of a maximalneurologic deficit at onset, possibly a seizure or loss ofconsciousness at onset and a cortical infarct. Thesefeatures may also be seen in non-cardioembolic strokesand are, thus, not definitive.

Opinion statementLong-term anticoagulation continues to be used and investigated as a means to prevent new or recurrent stroke. The best-established indications for long-term anti-coagulation are cardiac abnormalities capable of producing intracardiac thrombi that may embolize into the brain or systemic circulation. The firmest cardiac indications are mechanical prosthetic heart valves, mitral stenosis with atrial fibrillation, and atrial fibrillation with additional features placing them at increased risk for stroke. These and other “major” cardiac potential sources of emboli should be considered as anticoagulation candidates unless the estimated risk of bleeding is prohibitive. A number of noncardiac potential causes of stroke are generally managed with long-term anticoagulation. These include arterial dissections, cerebral venous sinus thrombosis, the antiphospholipid antibody syndrome, and congenital and acquired coagulopathies. Recent randomized studies do not support the use of long-term anticoagulation for the prevention of recurrent stroke in patients with noncardioembolic stroke that is not due to the previously outlined disorders. Whether long-term anticoagulation is beneficial in the specific population of patients with major documented intracranial atherosclerotic stenosis is currently under investigation.

412 Cerebrovascular Disorders

PRIMARY PREVENTIONAtrial fibrillation Nonvalvular atrial fibrillation is acommon arrhythmia and the major cardiac abnormal-ity that causes embolic stroke [1•]. Nonvalvular atrialfibrillation conveys a five-fold increase in stroke risk.Approximately 15% of patients with an ischemic strokewill have associated atrial fibrillation. This proportion isnearly doubled in the very elderly stroke population.About two thirds of the strokes occurring in patientswith atrial fibrillation are concluded to have beencaused by an embolus arising from a left atrialthrombus, and most often in the atrial appendage. Sev-eral randomized clinical treatment trials have studiedthe use of long-term oral anticoagulation for theprimary or secondary prevention of stroke in individu-als with atrial fibrillation. These studies confirmed anoverall annual risk of ischemic stroke of about 5% with-out anticoagulation. This risk is reduced by two thirdsor greater with the long-term use of warfarin [2,3••,

Class I]. These studies showed clinical features aspredictive of increased stroke risk. High-risk predictorsare prior stroke, transient ischemic attack or systemicembolus, age over 75 years, a history of hypertension, orpoor left ventricular systolic function by echocardio-gram or by virtue of clinical congestive heart failure,rheumatic mitral valve disease, and prosthetic heartvalve. These patients have an annual stroke risk of 8% orgreater without long-term anticoagulation. Patients 65to 75 years old and those with diabetes mellitus orcoronary heart disease have moderate risk. Patients lessthan 65 years old with no high-risk predictors have anannual stroke risk of about 1% to 2%.

Mitral stenosis Rheumatic mitral valve stenosis is thecardiac valvular disease with the greatest risk of stroke orsystemic embolus. A decline in mitral stenosis as a causeof stroke has been noted in the United States with betterprevention and management of rheumatic fever. The riskof emboli increases dramatically with the development ofatrial fibrillation [4]. Other important predictors of strokeor systemic embolus include advancing age and reducedcardiac indexes. The influence of left atrial size is unclear.Long-term anticoagulation is recommended in rheumaticmitral stenosis patients with a history of stroke or sys-temic embolus, coexistent atrial fibrillation, and in thosewithout atrial fibrillation who have a left atrial size ofgreater than 5.5 centimeters.

Prosthetic cardiac valves Mechanical prosthetic valvesrequire life-long anticoagulation to prevent stroke andsystemic embolism risk. The greatest risk of emboliza-tion is with the older caged ball or disk valves and withmechanical valves in the mitral rather than aortic posi-tion. Coexistence of atrial fibrillation increases strokerisk. As the estimated stroke and embolic risk increases,the recommended target International NormalizedRatio (INR) also raises and the use of concomitant anti-platelet, usually aspirin, therapy is recommended. Atarget INR of 2.5 (range 2.0 to 3.0) is recommended for

Table 1. Common sources of cardiogenic embolism

Major risk sources* Minor risk sources†

Atrial fibrillation Mitral valve prolapse±myxomatous changesMitral stenosis Severe mitral annular calcificationMechanical cardiac valvesRecent myocardial infarction Patent foramen ovaleLeft ventricular thrombus (especially if mobile or protruding) Atrial septal aneurysm; calcified aortic stenosisAtrial myxoma Left ventricular regional wall motion abnormalities

(prior myocardial infarction)Infective endocarditisDilated cardiomyopathiesMarantric endocarditis Aortic arch atheromatous plaques

*These are associated with substantial absolute risk of stroke, firmly linked to an embolic mechanism.†These are associated with low or uncertain absolute risk of initial or recurrent stroke.

Figure 1. Sources of cardiogenic emboli.

Long-term Anticoagulation Therapy in Prevention of Stroke Sherman 413

patients with a bileaflet valve or those with a tilting diskvalve when these valves are in the aortic position andthe patient is in sinus rhythm [5••]. Patients withbileaflet mechanical valves in the mitral position andthose with associated atrial fibrillation require higherlevels of anticoagulation with a target INR or 3.0 (range2.5 to 3.5). An alternative recommendation for thesehigher-risk patients is to target a lower INR of 2.5 andadd aspirin 80 to 100 mg per day. Patients with cagedball or disk valves are at the greatest risk for stroke, and,thus, a target INR of 3.0 (range 2.5 to 3-5) is recom-mended in combination with aspirin 80 to 100 mg perday. Any patient with a mechanical valve suffering astroke or systemic embolus with anticoagulation in therecommended target range should have the targetincreased to 3.0 (range 2.5 to 3.5) and aspirin 80 to 100mg per day added to their anticoagulation regimen.

Bioprosthetic heart valves have a lower risk of strokeor systemic embolism than do mechanical valves. Anti-coagulation is recommended for the first 3 monthsfollowing placement of a bioprosthetic valve in themitral or aortic position although the evidence of use isnot firm. A target INR of 2.5 (range 2.0 to 3.5) is recom-mended. Anticoagulation should be continued indefi-nitely in those patients with atrial fibrillation and for 3to 12 months in patients demonstrated to have a leftatrial thrombus at surgery or those with a stroke orsystemic embolus.

Patent foramen ovale and atrial septal aneurysm Emboliarising in the systemic or pulmonary arterial circulationmay pass through an intracardiac right to left shunt toproduce “paradoxical” embolism to the brain orsystemic arterial circulation. The most common poten-tial intracardiac shunt is a patent foramen ovale (PFO).Patent foramen ovale can be demonstrated in 25% to35% of autopsied hearts and, by contrast, echocardio-graphy in 10% to 15% of normal adults free of stroke.In one study from France, 581 patients less than 55years old with cryptogenic stroke were evaluated withechocardiography as a part of their diagnostic evalua-tion. Patent foramen ovale alone was discovered in37%, atrial septal aneurysm alone in 1.7%, and bothPFO and atrial septal aneurysm in 8.8% of the patients[6•, Class II]. After 4 years of follow-up on aspirintherapy, the risk of recurrent stroke was highest, 15.2%,in the group with both PFO and atrial septal aneurysm.The risk of stroke among patients with PFO alone oratrial septal aneurysm alone was not increased abovethat of the group without these abnormalities. Transe-sophageal echocardiography was performed in over 600patients in the Warfarin Aspirin Recurrent Stroke Study(WARSS) with PFO diagnosed in 34%. These patientshad been randomized to aspirin or warfarin, andpreliminary results suggest that stroke recurrence wasnot increased in the patients with PFO or atrial septal

aneurysm. There was no difference in recurrencebetween the two assigned therapies [7••,8].

Aortic arch atheromas Thrombi overlying atheromatousplaques of the aortic arch may be a source of emboli intothe systemic circulation. These atheromatous plaques aremost often visualized during transesophageal echo-cardiography. In one case control study, plaques 4 milli-meters and larger in thickness were significantly morecommon in patients with ischemic stroke than incontrols [9, Class II]. The predictors of increased risk forstroke appear to be size, plaques 4 millimeters or largerin thickness, and the lack of calcification in the plaque[10]. Observational studies suggest that patients withhigh-risk echocardiographic features have fewer ischemicevents with warfarin therapy than with antiplatelet treat-ment. These assignments were, however, not random-ized. These observations form the basis for the currentrecommendation that warfarin therapy be initiated inpatients with mobile aortic atheromas and aortic plaquesgreater than 4 millimeters in diameter by trans-esophageal echocardiography [4, Class I].

Ischemic heart disease Patients with coronary arterydisease have a risk of embolization from thrombi aris-ing in the left ventricle following myocardial infarctionor left atrium if atrial fibrillation is present. The risk ofstroke is greatest in patients with anterior myocardialinfarctions where the risk of stroke is 2% to 6%. Mostemboli occur within the first few weeks of an acute ante-rior myocardial infarction. The clinical features predica-tive of an increased risk of embolus and, thus,warranting anticoagulation include anterior Q-wavemyocardial infarction, severe left ventricular dysfunc-tion or congestive heart failure, a history of systemic orvenous embolus, and echocardiography demonstrationof a mural thrombus or atrial fibrillation. Anticoagula-tion is recommended for 3 months except for thepatients with atrial fibrillation who should be anti-coagulated indefinitely. A target INR of 2.5 (range 2.0 to3.0) is recommended [11••].

SECONDARY PREVENTIONThe WARSS compared aspirin and warfarin for preven-tion of recurrent stroke in 2206 patients with noncardi-oembolic ischemic stroke [7••]. Patients wererandomized to aspirin 325 mg daily or warfarin with atarget INR range of 1.4 to 2.8. After 2 years of follow-up,a primary event, death or recurrent ischemic stroke, hadoccurred in 17.8% of the warfarin-assigned group, andin 16% of the aspirin-assigned group. The rates of majorhemorrhage were 2.22 per 100 patient-years in thewarfarin group, and 1.49 per 100 patient-years in theaspirin group. Thus, warfarin anticoagulation was notsuperior to aspirin for prevention of recurrent ischemicstroke in this population. Potential high-risk subgroups

414 Cerebrovascular Disorders

of patients were examined in the WARSS study includ-ing patients with PFO demonstrated by transesophagealechocardiography. Thirty-four percent of the patientsundergoing transesophageal echocardiography hadPFO demonstrated. The presence of PFO did not predictan increased risk of recurrent stroke or differential treat-ment effect even when size and coexistence of atrialseptal aneurysm were considered [8].

Extracranial arterial dissection No randomized trialshave compared anticoagulation with antiplatelet therapyfor carotid or vertebral dissection. Case series suggestinga high risk of recurrent stroke presumably due to embo-lization of thrombotic material overlying the dissectionhave advocated the use of unfractinated or low molecularweight heparin initially followed by oral anticoagulationfor 3 to 6 months. At that time, an imaging study of thearea of dissection is repeated, and is used in the decisionto continue or discontinue anticoagulation. Alternativetherapies include antiplatelet therapy, stenting of theincurred vessel or surgical vascular repair.

Cerebral venous sinus thrombosis Cerebral venous sinusthrombosis presents clinically with a spectrum ofclinical signs and severity. Patients may have headache,seizures, focal neurologic deficits, and altered alertness.The diagnosis may be suggested by CT, but is now typi-cally demonstrated on magnetic resonance angiographywith attention to the cerebral veins and sinuses. Anti-coagulation with unfractinated or low molecular weightheparin is recommended initially and warfarin anti-coagulation continued for 3 to 6 months with a targetINR of 2.5 (range 2.0 to 3.0) [12••, Class III].

ENDOCARDITISInfective endocarditis Anticoagulation is generallyavoided in patients with native valve infectiveendocarditis, because of the risk of intracranial hemor-rhage. Prompt effective antibiotic therapy has reducedthe occurrence of emboli in these patients. If a patientwith a mechanical prosthetic valve develops infectiveendocarditis they should generally be continued ontheir anticoagulation, because of the high risk of throm-boemboli with the mechanical valve in the unanti-coagulated state.

Nonbacterial thrombotic endocarditis Nonbac te r ia lthrombotic endocarditis is associated with emboli inabout 42% of cases. Malignancy, chronic debilitatingillnesses, sepsis, and disseminated intravascular coagu-lation predispose to the deposition of small thinthrombi on the mitral or aortic valves that may embo-lize into the systemic circulation. The disorder may be

difficult to diagnose, because cardiac murmurs are lesscommon and vegetations less prominent on echocar-diography than with infective endocarditis. The diagno-sis or strong suspicion of nonbacterial thromboticendocarditis warrants anticoagulation with heparinbecause of the substantial risk of emboli. Studies of war-farin anticoagulation are insufficient to support its userather than heparin. Efforts should be directed toresolve the underlying predisposing disorder.

MITRAL VALVE PROLAPSEMitral valve prolapse (MVP) is the most common valvu-lar abnormality in adults. Mitral valve prolapse ispresent in about 6% of women and 4% of men.Although it certainly seems possible, based on casestudies, that MVP can cause transient ischemic attacks,stroke, or systemic embolus, the risk seems extremelysmall considering the frequency of MVP. Some studieshave suggested that thickened or redundant mitralvalves are associated with an increased embolic risk, butthere seems to be no consensus as to clinical or echocar-diograph characteristics that will identify patients withMVP at increased risk for embolus. Antiplatelet therapyis recommended for initial treatment of patients withtransient ischemic attack or stroke that is possiblyrelated to MVP. Long-term anticoagulation is recom-mended if there is associated atrial fibrillation or ifembolic events continue despite antiplatelet therapy.

HYPERCOAGULABLE STATESAntiphospholipid antibody syndrome Some pa t i en t swith venous and arterial thrombosis have circulatingantiphospholipid antibodies as the presumed basis fora hypercoagulable state. There is a population ofyounger patients, generally in their 20s or 30s, withrecurrent deep venous thrombosis, stroke, and sponta-neous abortions with circulating antiphospholipid anti-bodies. These patients have a high risk of recurrentthromboembolism, and cohort studies suggest theyhave reduced recurrence rates on long-term warfarinanticoagulation [13, Class III]. The high risk in this pop-ulation contrasts to the findings of the WARSSantiphospholipid substudy, which found no differencein the recurrence risk or response to treatment withaspirin or warfarin in the patients with circulatingantiphospholipid antibodies [14]. Thus, it would seemappropriate to consider long-term warfarin therapy inpatients with the antiphospholipid antibody syndromewith multiple venous and arterial thromboemboli, andno alternative explanation, however, the older patientwith a noncardioembolic ischemic stroke does notseem to have their risk altered by the presence ofantiphospholipid antibodies.

Long-term Anticoagulation Therapy in Prevention of Stroke Sherman 415

Treatment

Warfarin

Standard dosage The dose of warfarin ranges typically between 2 and 10 mg per day, and must be individualized and adjusted to achieve the desired INR. Therapy is initiated with an estimated daily maintenance dose, for example, at 5 mg per day. Lower doses of 3 to 4 mg per day may be initiated in elderly and small individuals. The prothrom-bin time and INR are monitored frequently until in the target range and stable. Thereafter, stable patients are typically monitored at least monthly. The target INR for most indications is 2.5 (range 2.0 to 3.0). The target is increased to 3.0 (range 2.5 to 3.5) in patients with high-risk mechanical prosthetic cardiac valves and in patients with the antiphospholipid antibody syndrome [15••,16•].

Contraindications Warfarin therapy is contraindicated in patients at risk for major hemorrhage including those with active bleeding, recent surgery, pregnancy, esophageal varices, thrombocytopenia, concurrent use of thrombolytics, recent lumbar puncture, and congenital clotting defects. Patients at increased risk for falls or other trauma may not be candidates for long-term anticoagulation therapy. The importance of compliance dictates that unreliable or demented patients should only receive warfarin therapy if adequate supervision of their warfarin therapy and INR monitoring is possible.

Main drug interactions There are numerous potential drug interactions with warfarin. Some potentate and others inhibit the anticoagulant effects of warfarin. In addition, certain foods are high in vitamin K, and can thereby inhibit the effects of warfarin. Other substances such as alcohol may increase bleeding risk. The best course is to monitor the INR following any change in medications in a patient on warfarin and adjust the dose as needed. In addition patient education about the potential effects of over the counter preparations and foods is an important element of warfarin management.

Main side effects The main potential side effect of warfarin is bleeding. The only other serious side effect is skin necrosis. This, fortunately, uncommon complication is seen on the third to eighth day of therapy, and is caused by thrombosis of venules and capillar-ies within the subcutaneous fat. A number of factors influence bleeding risk with warfarin therapy. The intensity of anticoagulation is directly related to bleeding risk with major bleeds increased dramatically with INR levels above 4.0. Patients over age 75 have increased risk of hemorrhage particularly intracerebral hemor-rhage. Patients with cerebrovascular disease are at greater risk for intracerebral hemorrhage than are patients without an ischemic cerebrovascular history. Other comorbid conditions increasing bleeding risk include hypertension, heart disease, renal insufficiency, and malignancy. Alcoholism and liver disease are also consid-ered by many to increase the bleeding risk associated with warfarin therapy.

Special points Warfarin management by anticoagulation clinics or anticoagulation management services appears to reduce thromboembolic events and bleeding complications in patients of warfarin.

Cost/cost effectiveness The annual cost of warfarin is estimated to be $232.43 (generic) for the 5 mg per day dose and $353.72 for the 10 mg per day dose. In addition the cost of regular, typically at least monthly, monitoring of the INR, outpatient visits to review and adjust dosing must be considered. Several analyses have concluded that there is a cost saving when an anticoagulation monitoring service or clinic manages anticoagulation. Fewer bleeding complications and reduced thromboembolic complications account for these cost savings [17•].

Pharmacologic treatment

416 Cerebrovascular Disorders

• Alternative agents for anticoagulation are under investigation. The motivation for development of alternatives is to discover effective means of anticoagulation and prevent thromboemboli without excessive bleeding complications, a broader therapeutic window and less need for intensive monitoring of coagulation parameters. One group of agents under investi-gation acts as direct thrombin inhibitors. These agents act against free and clot bound thrombin, a theoretic advantage. The first agents studied, hiru-din, bivalirudin, and argatraban had the disadvantage of only parenteral administration. Melagatran has an oral prodrug, ximelagatran, that is currently under investigation and my prove to be an alternative to warfarin anticoagulation not requiring monitoring of anticoagulation level (INR).

References and Recommended ReadingPapers of particular interest, published recently, have been highlighted as:• Of importance•• Of major importance

1.• Falk RH: Atrial fibrillation. N Engl J Med 2001, 344:1067–1078.

An excellent review of atrial fibrillation management.2. Albers GW: Choice of antithrombotic therapy for

stroke prevention in atrial fibrillation: warfarin, aspi-rin, or both? Arch Intern Med 1998, 158:1487–1491.

3.•• Albers GW, Dalen JE, Laupacis A, et al.: Antithrombotic therapy in atrial fibrillation. Chest 2001, 119(suppl):194S–206S.

Review of clinical trials and recommendations for manage-ment of atrial fibrillation.4. Salem DN, Daudelin HD, Levine HJ, et al.: Antithrom-

botic therapy in valvular heart disease. Chest 2001, 119(suppl):207S–219S.

5.•• Stein PD, Alpert JS, Bussey HI, et al.: Antithrombotic therapy in patients with mechanical and biological pros-thetic heart valves. Chest 2001, 119(suppl):220S–227S.

An excellent review and consensus statement.6.• Mas JL, Arquizan C, Lamy C, et al.: Recurrent cere-

brovascular events associated with patent foramen ovale, atrial septal aneurysm, or both. N Engl J Med 2001, 345:1740–1746.

7.•• Mohr JP, Thompson JLP, Lazar RM, et al.: A comparison of warfarin and aspirin for the prevention of recurrent ischemic stroke. N Engl J Med 2001, 345:1444–1451.

Report of this large randomized trial.8. Homma S, Sacco RL, DiTullio MR, et al.: PICSS-PFO

in cryptogenic stroke study. Circulation 2002, 105:2625–2631.

9. Amarenco P, Cohen A, Tzourio C, et al.: Atherosclerotic disease of the aortic arch and the risk of ischemic stroke. N Engl J Med 1994, 331:1474–1479.

10. Cohen A, Tzourio C, Bertrand B, et al.: Aortic plaque morphology and vascular events: a follow-up study in patients with ischemic stroke. FAPS investigators: French study of aortic plaques in stroke. Circulation 1997, 96:3838–3841.

11.•• Cairns JA, Theroux P, Lewis HD, et al.: Antithrombotic agents in coronary artery disease. Chest 2001, 119(suppl):228S–252S.

An excellent review and consensus statement.12.••Albers G, Amarenco P, Easton JD, et al.: Antithrombotic

and thrombolytic therapy for ischemic stroke. Chest 2001, 119(suppl):300S–320S.

An excellent review and consensus statement.13. Khamashta MA, Cuadrado MJ, Mujic F, et al.: The

management of thrombosis in the antiphospholipid-antibody syndrome. N Engl J Med 1995, 332:993–997.

14. Brey R: Preliminary concordance between antiphos-pholipid (aPL) assays in a subset of ischemic stroke patients enrolled in the WARSS/APASS collaboration. Stroke 2002, In press.

15.••Levine MN, Raskob G, Landefeld S, Kearon C: Hemorrhagic complications of anticoagulant treatment. Chest 2001, 119(suppl):108S–121S.

An excellent review and consensus statement.16.• Hirsh J, Dalen JE, Anderson DR, et al.: Oral anticoagu-

lants: mechanism of action, clinical effectiveness, and optimal therapeutic range. Chest 2001, 119:8–21.

An excellent review and consensus statement.17.• Ansell J, Hirsh J, Dalen J, et al.: Managing oral anti-

coagulant therapy. Chest 2001, 119(suppl):22S–38S.An excellent review and consensus statement.

Emerging therapies