Home SVCC                                                  Area: English - Español - Português

Efficacy of Anticoagulation for
Secondary Stroke Prevention

Andrew Evans, MRCP

Guy's, King's & St Thomas' School of Medicine,
London, United Kingdom

   I am conscious that to Cardiologists, stroke is often seen as a feared complication for patients under their care. Much of your practice will revolve around primary prevention whether it relates to patients with mitral stenosis, non-rheumatic atrial fibrillation or the use of statins to prevent stroke in patients with ischaemic heart disease. It may be routine in your own country to leave decisions regarding secondary prevention to your colleagues in Neurology or Internal Medicine. Nevertheless, although the pathological processes in stroke have many differences from those of myocardial ischaemia and infarction, there is still considerable overlap between these vascular diseases. Furthermore, decisions made in the context of a patient's heart condition may well impact on the management of their stroke.

   (Figure 1) In this lecture, I will discuss the importance of secondary prevention of stroke. I will then review the evidence behind the use of anticoagulants to prevent further stroke in a variety of pathological conditions. Despite strong evidence to support the use of anticoagulants in certain circumstances, this does not always happen in actual clinical practice; I will explore the reasons why this might be so, and present some of our own group's work to examine if these reasons are justified. I will then discuss the areas in which there is still doubt and in which there is scope for further work.

Figure 1

   (Figure 2) Stroke is a devastating disease. It is repeatedly quoted that it is the third greatest cause of death (after ischaemic heart disease and cancer) and the number one cause of severe disability in the Western world. While community-based stroke registers suggest that 19-25% of patients will die within the first month of their stroke, this still leaves a large number (with varying degrees of disability) alive. After this time, the risk of death is approximately 10% per year, giving a cumulative risk of death of 60% over 5 years.

Figure 2

   Of the stroke survivors, approximately two-thirds will be functionally independent. It is this group of patients that have the most to gain from effective secondary prevention, particularly as the risk of recurrent stroke is high. There is a 9% risk of recurrent stroke over the first 6 months rising to a 22% cumulative risk over 5 years. A quarter of these recurrent strokes will be fatal and the majority will leave permanent disability. Nevertheless this high level of risk means that interventions that reduce that risk will have a greater individual impact and be more clinically and cost-effective.

   The basic pathological aetiology of ischaemic stroke is arterial occlusion whether from embolization or in-situ thrombosis. It is simply postulated that treatment with anticoagulation may act to prevent both of these processes and so reduce stroke. However, anticoagulant treatment carries both risk and inconvenience. Therefore, we have a number of questions to consider:
(Figure 3)

Figura 3

    Is it effective in reducing stroke?
    Are the benefits outweighed by the risks of treatment?
    Is the level of overall benefit worth the inconvenience and expense of monitoring?

   The effectiveness of anticoagulation may well be determined by the precise mechanism. The underlying pathologies of stroke can be grouped as follows:
[Figure 4]

Figure 4

   1. Intracranial bleeding. This may be primary (usually associated with hypertension) or secondary to underlying structural defects such as aneurysms, arteriovenous malformations or tumours. In the context of secondary prevention with anticoagulants, bleeding may occur into areas of recent cerebral infarction, so-called haemorrhagic transformation of an infarct.

   2. Cardioembolism. Thrombus within the heart may embolise to the cranial vessels leading to arterial occlusion.

   3. Artery to artery embolisation. The commonest cause is from stenosis of the extracranial carotid arteries

   4. Large vessel occlusion. In situ thrombosis of the extracranial or intracranial carotid, vertebral or basilar arteries.

   5. Small vessel occlusion. Occlusion of the small deep penetrating arteries of the brain. These produce characteristic small areas of infarction (known as lacunes) in the basal ganglia or brainstem, which are often visible on computerised tomographic (CT) scanning and usually seen with magnetic resonance imaging (MRI). Diffuse ischaemic changes around the ventricles, known as leukoaraiosis, is also thought to be secondary to widespread small vessel ischaemia.

   Strokes due to cardioembolism, artery-to-artery embolization and large vessel occlusion tend to be more severe than those due to small vessel disease. This is partially because of the potentially larger area of brain supplied by these arteries, but also that they will often produce the particularly disabling conditions of dysphasia, field deficits, apraxias and neglect (so-called cortical signs).

Secondary Prevention in the presence of Atrial Fibrillation
   Data from the Framingham project, established AF as a strong independent risk factor for stroke and it is by far the commonest cause of cardioembolic stroke. It has long been known that patients with mitral stenosis and AF had a high risk of stroke. Although never subjected to the rigours of a randomised-controlled trial, long-term anticoagulation with warfarin for rheumatic AF has been accepted standard practice.

   (Figure 5) In this context a number of randomised controlled trials were performed in the 1980s to study the effect of anticoagulation and antiplatelet therapy (AFASAK, BAATAF, SPAF, CAFA and SPINAF) on non-rheumatic atrial fibrillation. Although these studies are often quoted as primary prevention, they did include a number of patients with previous stroke. (Figure 6) Meta-analysis of these studies enabled stratification of stroke risk such that those with the highest stroke risk would benefit from long-term anticoagulation with warfarin, whereas aspirin was the preferred agent in those with lower risk. Previous stroke or transient ischaemic attack (TIA) put patients into the high-risk group.

Figure 5

Figure 6

   (Figure 7) This finding was directly tested in the European Atrial Fibrillation Trial (EAFT), the definitive study in secondary prevention in AF. This was a multicentre [108 centres] RCT of patients with minor stroke or TIA with 2 schedules- (1) open label anticoagulation (Target International Normal Ratio 2.5-4.0) vs double blind aspirin or placebo [669 patients] (2) double blind aspirin v placebo [338 patients]. Follow up was for an average 2.3 years and outcome was blinded. Overall results showed a reduction in stroke from 12% per year for placebo to 4% with long-term anticoagulation. Aspirin showed a lesser reduction from 12% to 10% per year. The risk of major bleeding on warfarin was 2.8% per year. 21% of patients discontinued warfarin during the course of the study. (Figure 8) It is on this Grade I evidence that it is recommended that in the absence of contra-indications to anticoagulants, all patients with TIA or minor stroke who are in AF should be anticoagulated.

Figure 7

Figure 8

   When to start anticoagulation is more controversial. Studies which looked at acute anticoagulation with heparin within 48 hours of ischaemic stroke do show a reduction in ischaemic stroke recurrence (and venous thromboembolism). However, this is entirely offset by an increase in symptomatic haemorrhagic transformation of infarct and extracranial bleeding. There is thus no randomised-controlled trial evidence to favour commencing anticoagulation with stroke and AF before 2 weeks after onset. Nevertheless, many centres do commence anticoagulation earlier in the case of small infarcts, where the risk of intracranial bleeding is likely to be lower.

Secondary Prevention in Sinus Rhythm
    (Figure 9) Until, recently much less was known of the effect of anticoagulation on patients in sinus rhythm. Systematic review of secondary prevention trials of oral anticoagulants versus placebo showed no clear evidence of benefit with a significantly increased risk of intracranial haemorrhage. However, most of these studies were small and some presumed ischaemic stroke patients had not had CT scanning.

Figure 9

    When aspirin is compared with warfarin, the first major trial (SPIRIT) had to be stopped early because of an excess of cerebral bleeds in the anticoagulation limb, particularly in the elderly and in those with leukoaraiosis. This study was criticised for the relatively high level of anticoagulation used (target INR 3.0-4.5), but does emphasise the potential for harm with anticoagulants.

    (Figure 10) The preliminary results of the Warfarin-Aspirin Recurrent Stroke Study (WARSS) have only recently been presented at international conferences. However, it does shed light on a number of previously unanswered questions. The trial itself was a model of scientific rigour, in that it is one of the few studies in anticoagulation that was truly double-blind. 2206 patients were randomly allocated to warfarin (target INR 1.4-2.8) or aspirin 325mg. The triallists achieved blinding of treatment allocation by providing a centralised INR monitoring service, which despatched dummy INR values for patients on aspirin. The major end-point of recurrent ischaemic stroke showed a non-significant benefit in favour of aspirin, with a good safety record for both aspirin and warfarin. Subgroup analysis showed non-significant tends in favour of aspirin for small-vessel stroke and in favour of warfarin for large vessel stroke with no carotid disease. Given the relative inconvenience and expense of treatment with anticoagulants, it is unlikely that any small benefit in favour of warfarin would be clinically significant.

Figure 10

   Further, information may become available when the results of the ESPRIT trial (open-label aspirin vs aspirin/dipyridamole vs warfarin (INR 2.0-3.0)).

    In the acute phase, there is also no randomised-controlled data to suggest any benefit from immediate anticoagulation with heparin or heparinoids in sinus rhythm. Again, any reduction in recurrent ischaemic stroke is offset by increases in intracranial bleeding, predominantly due to haemorrhagic transformation of infarcts.

Secondary Prevention in Specific Stroke Pathologies
(Figure 11)
   As you may have noticed, the spectrum of "stroke in sinus rhythm" is broad, covering some cardioembolic stroke, artery-to-artery embolisation and both large and small vessel occlusions. Anticoagulation is advocated by many neurologists and stroke physicians in a number of specific circumstances. However, none has been shown to benefit in the context of a randomised-controlled trial.

Figure 11

   1) Structural heart disease. Aside from mitral stenosis and metallic prosthetic valves, where consensus has been reached on the value of anticoagulation, there is no proven benefit in patients with ventricular dyskinesia, other valvular pathologies or cardiomyopathy. A particular management problem is patent foramen ovale. I'm sure we have all seen dramatic post mortem specimens of thrombus traversing a septal defect. Potential options in secondary prevention are anticoagulation, aspirin, no treatment or surgery. There is no definite advantage to any strategy, although a number of studies are underway and patients with PFO were included as a substudy of WARSS.

   2) Intracranial stenosis. Retrospective studies suggest that the risk of recurrent stroke is lower on anticoagulation but there is no prospective randomised data.

   3) Arterial dissection. This not uncommon cause of stroke, particularly in younger adults, is commonly treated with heparin and warfarin. However, evidence is non-randomised and based on case series and anecdote.

   4) Antiphospholipid syndrome. Recurrent stroke associated with anticardiolipin antibodies and often migraine, venous thrombosis and livedo reticularis, Patients are often treated with high-dose warfarin but without grade I evidence of its efficacy

   (Figure 12)
   It is recognised that despite strong evidence and guidelines, the uptake of anticoagulation for atrial fibrillation (for both primary and secondary prevention) is much lower in clinical practice than one would anticipate. There may be a number of reasons for this. They include:

Figure 12

1) Trial patients not representative of patients in practice. Trial patients tend to be younger and fitter than those met in ones own practice. Physicians may have concerns that trials cannot be extrapolated to older patients with comorbidity.

2) Monitoring in practice may be less effective than within a clinical trial. This may be particularly important when we consider anticoagulation, as poorer INR monitoring may result in reduced benefit and increased risk

3) Patient choice. Patients in real-life may be unwilling to attend coagulation clinics or suffer minor bleeding, when they are not part of a trial.

4) Resources. Potentially costly interventions may not be able to be met outside the funding of a trial.

5) Inadequate knowledge of evidence.

   (Figure 13) With the first three of these questions in mind we investigated how trial evidence translated into clinical practice. We prospectively looked at 288 stroke patients in atrial fibrillation who were eligible for anticoagulants. Warfarin control (target INR 2.0-3.0) was managed by local anticoagulant services. We compared patient characteristics, anticoagulation practice and outcome events with the data from the 225 patients in the anticoagulation limb of the European Atrial Fibrillation Trial.

Figure 13

   (Figure 14) Patients in our cohort were significantly older (76 years v 71 years) and more likely to have ischaemic heart disease or heart failure. (Figure 15) Comparing anticoagulation practice, we found that patients in clinical practice spent a longer time above the target range than in the EAFT (Figure 16). Although, there was no significant difference between the proportion withdrawing from treatment, real-world patients were more likely to withdraw due to choice, compliance or logistics (Figure 17). Despite these differences, the rates of ischaemic stroke (5.7% v 3.9%), intracranial bleeding (0.5% v 0.4%) and extracranial bleeding {major (2.5% v 2.6%) and minor (7% v 9.2%)} were not significantly different.

Figure 14


Figura 15


Figure 16


Figure 17

    This suggests that the results of the EAFT can be reproduced in clinical practice and that fears over the issues of efficacy and safety are probably unfounded. It is important to recognise however, that in an, often elderly, population with stroke, contraindications to warfarin should include recurrent falls and cognitive difficulties preventing compliance.

    (Figure 18) When we looked at the 32 recurrent strokes in this study, 14 (44%) were definitely cardioembolic; 11 of these occurred when the INR was subtherapeutic. However, 11 (34%) were lacunar strokes; of these 8 occurred during adequate anticoagulation. This raises the question of whether atrial fibrillation was the cause of these lacunar strokes or just acting as a bystander. If these lacunar strokes are not cardioembolic in nature, anticoagulation may not be effective in preventing them. There is some evidence that anticoagulation may not be effective in atrial fibrillation with small vessel disease.

Figure 18


   (Figure 19) There remain a number of aspects of anticoagulation in secondary prevention that remain uncertain. These include:

Is anticoagulation worthwhile for patients in atrial fibrillation with severe stroke? Some small studies have suggested that dependent patients may be at greater risk of complications.
Is there a role for anticoagulants in specific causes of stroke, such as patent foramen or antiphospholipid syndrome.
Is there a role for anticoagulation in sinus rhythm at higher INRs?
Is the subtype of stroke important in atrial fibrillation?
Are there alternatives to warfarin, which may be safer or more convenient?

Figure 19

   The answers to some of these questions will be discovered from ongoing trials. Others may remain more elusive.

   (Figure 20) A randomised controlled trial has shown anticoagulation to be the most effective secondary preventative strategy in atrial fibrillation
   The results of this trial can be reproduced in actual clinical practice.
   There is no evidence for anticoagulation for secondary prevention in acute stroke.
   Anticoagulation is unlikely to be effective in sinus rhythm
   There is no strong evidence for anticoagulation in any other specific stroke pathologies, although many trials are ongoing.

Figure 20



Your questions, contributions and commentaries will be answered
by the lecturer or experts on the subject in the Stroke list.
Please fill in the form (in Spanish, Portuguese or English) and press the "Send" button.

or commentary
Name and Surname:
E-Mail address:


2nd Virtual Congress of Cardiology

Dr. Florencio Garófalo
Steering Committee
Dr. Raúl Bretal
Scientific Committee
Dr. Armando Pacher
Technical Committee - CETIFAC

Copyright© 1999-2001 Argentine Federation of Cardiology
All rights reserved


This company contributed to the Congress: