Primary and Secondary Prevention of Cerebral Ischaemia.*

Joint Guidelines issued by the German Neurological Society (Deutsche Gesellschaft für Neurologie; DGN) and the German Stroke Society (Deutsche Schlaganfallgesellschaft; DSG)

* published in German September 2005

What's new?

	When performed by an experienced vascular surgeon, endarterectomy is capable of reducing stroke risk (A) in patients suffering from asymptomatic high-grade carotid artery stenosis (> 60% degree of stenosis based on Doppler or duplex sonography). The benefit is markedly lower in women than in men. The number needed to treat (NNT) is 40 for the prevention of severe strokes including perioperative mortality (over a 5-year period).
	In patients with symptomatic carotid artery stenoses, the time point of surgery is critical to achieve a preventative effect with carotid endarterectomy as secondary stroke prevention. If the operation is performed later than 14 days after the initial event, it no longer has a preventative effect (B).
	Prophylaxis with ASA should be continued before, during and after carotid surgery (C).
	Acetylsalicylic acid is not effective in the primary prevention of stroke in men (A). In women with vascular risk factors aged > 45 years, strokes but not myocardial infarctions are prevented (B).
	The combination of clopidogrel and acetylsalicylic acid is not more effective than clopidogrel monotherapy as secondary stroke prevention, and leads to a significant increase in life-threatening and severe haemorrhagic complications ( ).
	GP-llb/llla antagonists are not more effective than acetylsalicylic acid as secondary stroke prevention, and lead to a significant increase in haemorrhagic complications ().

The most important recommendations at a glance

Primary prevention – risk factors

	Primary stroke prevention includes a "healthy lifestyle" with at least 30 minutes of exercise three times a week and a diet rich in fruits and vegetables and/or a Mediterranean diet (A). Cardiovascular risk factors (blood pressure, blood sugar, disorders of lipid metabolism) should be checked regularly and then treated appropriately (B).
	Patients suffering from arterial hypertension (BP systolic > 140 mmHg, diastolic > 90 mmHg; diabetics: BP systolic > 130 mmHg, diastolic > 85 mmHg) have to be treated by means of diet (DASH diet, low-sodium diet), endurance sport and/or antihypertensives (A). In this context, the preventative effect of antihypertensives is greater, the stronger the blood-pressure-lowering effect (A). Individual antihypertensives differ only slightly in their effectiveness in preventing stroke (A). Alpha blockers are less effective than other antihypertensives (B).
	Smokers should stop smoking. Pharmacological (nicotine plaster, nicotine chewing gum, selective serotonin reuptake inhibitors or buproprion) or non-pharmacological aids (behavioural therapy, group therapy; B) have proved effective.
	Patients suffering from coronary heart disease or with myocardial infarction in their history and an LDL > 100 mg% should be treated with a statin (A). In persons without CHD and one or no vascular risk factors, a statin should be given when their LDL levels > 160 mg%, with moderate risk and LDL > 130 g % and > 100 mg% and several vascular risk factors. The best evidence exists for simvastatin, pravastatin and atorvastatin.
	Diabetics should be treated with diet, regular exercise, antidiabetics and insulin, as needed (B). Normoglycaemic values should be striven for. In diabetics, the importance of antihypertensive treatment with ACE inhibitors or sartans and the administration of statins is particularly high with regard to stroke prevention (B).

Primary prevention – atrial fibrillation

	Patients suffering from persistent or paroxysmal atrial fibrillation and concomitant vascular risk factors (hypertension, coronary heart disease, heart failure, age over 75 years) should receive oral anticoagulation with a target INR of 2.0–3.0 (A). In patients aged over 75 years, an INR of about 2.0 should be targeted. No anticoagulation or platelet aggregation inhibition is necessary in patients aged under 65 years with so-called lone atrial fibrillation, a rare form of atrial fibrillation, and lacking vascular risk factors. Acetylsalicylic acid (100–300 mg) is recommended for patients aged over 65 years without any vascular risk factors and atrial fibrillation. ASA is likewise used in patients with contraindications to oral anticoagulants such as cerebral microangiopathy, incipient dementia and an elevated risk of falls.
	Oral anticoagulation for stroke prevention for longer than 3 months is not necessary in patients who have received heart valve replacement with a bioprosthesis.
	An asymptomatic patent foramen ovale with or without atrial septal aneurysm (ASA) does not require treatment (A).

Primary prevention – platelet aggregation inhibitors

Acetylsalicylic acid is not effective for primary stroke prevention in men (A). It prevents strokes, but not myocardial infarctions in women aged > 45 years with vascular risk factors (B).

Primary prevention – high-grade stenosis with origin in the internal carotid artery

Surgery for asymptomatic carotid artery stenosis with a degree of stenosis > 60% based on Doppler or duplex sonographic criteria significantly reduces the stroke risk. However, this is only true when the combined 30-day mortality and morbidity of the intervention is less than 3% (A) (  ). The patient's life expectancy should be > 5 years.

Secondary stroke prevention – risk factors

	Rigorous treatment of arterial hypertension reduces the risk of stroke (  ) (A). The combination of perindopril plus indapamide is significantly more effective than placebo ( ), and eprosartan is significantly more effective than the calcium antagonist nitrendipine ( ). In post-stroke patients, ramipril reduces vascular endpoints, but not the stroke risk.
	Nevertheless, all antihypertensives are likely to provide effective secondary stroke prevention (B).
	Treatment of diabetes mellitus reduces the risk of stroke (C), although this has not been well investigated in prospective studies to date.
	Statins should be given to patients suffering from focal cerebral ischaemia and CHD irrespective of their baseline levels of LDL cholesterol (  ) (A). The target values for LDL should range between 70 and 100 mg%.
	Simvastatin (40 mg) can be given to patients suffering from focal cerebral ischaemia without CHD. The principal effect of this is to lower the vascular risk in general ( ) (B). The other statins are probably also effective (C).
	The treatment of hyperhomocysteinaemia with vitamin B6, B12 and folic acid is not effective in secondary stroke prevention () (B).
	Postmenopausal hormone replacement is not effective in secondary stroke prevention () (B).

Secondary prevention – platelet aggregation inhibitors

	Platelet aggregation inhibitors are effective for secondary prevention in patients suffering from focal ischaemia (  ) (A). This is true for ASA (50–150 mg), ASA (25 mg bid) plus dipyridamole (200 mg bid) and clopidogrel (75 mg) (A).
	After TIA and ischaemic stroke and with a low risk of recurrence (< 4%/year), the daily administration of 50–150 mg acetylsalicylic acid is recommended (B).
	In patients with a high risk of recurrence (· 4% per year), the twice-daily administration of the fixed combination of 25 mg acetylsalicylic acid plus 200 mg sustained-release dipyridamole is recommended (B).
	In patients suffering from high risk of recurrence (· 4%/year) and in conjunction with peripheral arterial disease (PAD), clopidogrel 75 mg is recommended (C).
	In patients with a contraindication or intolerance to ASA, clopidogrel is recommended (A).
	In patients, who develop gastric or duodenal ulcers under ASA prophylaxis, continuation of ASA in combination with a proton pump inhibitor after a period off drug is recommended (B).
	ASA in doses > 150 mg leads to an elevated risk of haemorrhagic complications ( ).
	The combination of clopidogrel plus ASA does not show any better efficacy than clopidogrel monotherapy, but does lead to a higher rate of haemorrhagic complications (  )
	The duration of treatment with platelet aggregation inhibitors has not yet been investigated beyond the 4th year after the initial event. Theoretically, prophylaxis should be given for the patient's lifetime, if tolerated (C).
	If a recurrent ischaemic event occurs under ASA, the pathophysiology and risk of recurrence should be re-evaluated. If a cardiac embolic source is revealed, oral anticoagulation should follow. If the risk of recurrence has not changed (< 4%/year), prophylaxis with ASA is continued (C). Otherwise, see above.
	Patients suffering from a TIA or a stroke and acute coronary syndrome should be treated with a combination of 75 mg clopidogrel and 75 mg ASA over a period of 3 months ( ) (C).
	GP-llb/llla antagonists should not be used for secondary stroke prevention (A) (). They are not more effective than acetylsalicylic acid, but their use is associated with a significantly elevated bleeding risk.

Secondary prevention – atrial fibrillation

	In patients with a cardiac embolic source, and particularly those with atrial fibrillation, oral anticoagulation with INR levels of 3.0 is recommended ( ) (A).
	After TIA and mild ischaemic stroke and atrial fibrillation, oral anticoagulation can be started within 3–5 days (C)
	In patients with mechanical heart valves, anticoagulation should be continued with INR values between 2.0 and 3.5 () (C).
	In patients with biological valves and cerebral ischaemia, temporary anticoagulation is recommended for 3 months () (C).

Secondary prevention – patent foramen ovale (PFO)

	In patients with patent foramen ovale (PFO) alone, irrespective of the size, and first cerebral ischaemic event, prophylaxis with ASA (100–300 mg) is given (B).
	In case of recurrence while using acetylsalicylic acid (ASA) or a PFO exists concomitant with ASA (atrial septal aneurysm), give oral anticoagulation with an INR of 2.0–3.0 is recommended (for at least 2 years) (C).
	If there is a further recurrence or contraindications to oral anticoagulation, interventional PFO occlusion should be performed (umbrella occlusion) (C).

Secondary prevention – high-grade carotid artery stenosis

	Patients with high-grade symptomatic carotid artery stenoses should undergo endarterectomy (  ) (A). The benefit of surgery increases with the degree of stenosis of 70–95%. The benefit of surgery is lower when the degree of stenosis is between 50% and 70%, in subtotal stenoses, in women, and when the operation is performed more than 2 weeks after the index event ( ) (B).
	The benefit of surgery is lost when the complication rate is > 6% (  ).
	The period between event and operation should be bridged with platelet aggregation inhibitors. ASA should continue to be given before, during and after the operation (B). Clopidogrel should be replaced by ASA at the latest 5 days before surgery (C).
	Neurosonographic procedures, MRI or CT angiography are sufficient for verifying the diagnosis of carotid artery stenosis (A). A DSA is usually not necessary (B).
	At present, carotid angioplasty with stenting has not yet become a routine procedure. Stenting can be considered for patients suffering from recurrent stenosis after TEA, high-grade stenoses after radiotherapy or high-carotid stenoses poorly accessible to surgical intervention (C). Isolated angioplasty without stenting should no longer be conducted given its high restenosis rate (C).
	Prophylaxis with clopidogrel (75 mg) plus ASA (100 mg) should be given before, during and after stenting for 1–3 months.

Intracranial stenoses

	In patients suffering from high-grade intracranial stenoses or occlusions, anticoagulation with an INR of 3.0 is no more effective than the administration of 1300 mg ASA (), but leads to a higher rate of haemorrhagic complications and can therefore not be recommended (B). In view of the poor tolerability of 1300 mg ASA, we recommend prophylaxis with 100–300 mg ASA (C).
	Stent implantation can be considered for patients with recurrent events (C). Afterwards, 75 mg clopidogrel and 100 mg ASA are administrated over a period of 1–3 months (C).

Anticoagulation – early secondary prophylaxis

	PTT-effective treatment with heparin is not indicated for the treatment of acute stroke () (A).
	This is also true for patients suffering from non-rheumatic atrial fibrillation (ß) (A).
	In special indications, the risks and benefits must be weighed.
	Heparin in low-dose and low-molecular-weight heparins reduce the risk of deep vein leg thrombosis in stroke patients with paretic legs and those who are bed-ridden (  ) (B). This has been demonstrated for fraxiparin, tinzaparin and certoparin ( ).

Primary prevention

Objectives

The objective of primary prevention is to prevent cerebral ischaemia or transient ischaemic attacks (TIAs) in patients without previous cerebrovascular diseases. In this context, we differentiate between four subgroups of patients:

	Completely healthy persons
	Persons without major pre-existing diseases, but with vascular risk factors
	Persons with asymptomatic stenoses or occlusions of the arteries supplying the brain
	Patients suffering from vascular diseases in other vascular regions (myocardial infarctions, coronary heart disease = CHD or peripheral arterial disease = PAD

Basically, the hazard should be seen in ascending order for the four groups, which should affect prevention strategies. Unfortunately, no comparative investigations on the effect of primary stroke prevention have been made for the different patient populations.

Epidemiology

Depending on the geographic region, at total of 100–700 strokes occur per 100,000 persons and year. Currently, the highest incidences are registered in the East European countries, with relatively low rates in the West European countries, Scandinavia and North America (Khaw 1996).

Tests and examinations

Mandatory:

Determination of vascular risk factors (blood pressure, blood sugar, cholesterol, and, if appropriate, tests for LDL and HDL), ECG, neurological and full physical examination.

Necessary in individual cases:

Ultrasound of the extracranial arteries, echocardiography, CT to exclude clinically silent ischaemia or any subcortical vascular encephalopathy in patients suffering from arterial hypertension for many years.

Therapy

Recommended treatment

	Includes a "healthy lifestyle" with at least 30 minutes of sport three times a week and a diet rich in fruits and vegetables and/or a Mediterranean diet (A). Cardiovascular risk factors (blood pressure, blood sugar, disorders of lipid metabolism) should be checked regularly and then treated appropriately (B).
	Patients suffering from arterial hypertension (BP systolic > 140 mmHg, diastolic > 90 mmHg, diabetics: BP systolic > 130 mmHg, diastolic > 85 mmHg) have to be with treated by means of diet (DASH diet, low-sodium diet), endurance sport and/or antihypertensives (A). In this context, the preventative effect of antihypertensives is greater, the stronger the blood-pressure-lowering effect (A). The individual antihypertensives differ only slightly in their effectiveness in preventing stroke (A). Alpha blockers are less effective than other antihypertensives (B).
	Smokers should stop smoking. Pharmacological (nicotine dermal patch, nicotine chewing gum, selective serotonin reuptake inhibitors or buproprion) or non-pharmacological aids (behavioural therapy, group therapy; B) have proved effective.
	Patients suffering from coronary heart disease or with a myocardial infarction in their history and an LDL > 100 mg% should be treated with a statin (A). In persons without CHD and one or no vascular risk factors, a statin should be given when their LDL levels > 160 mg%, with moderate risk and LDL > 130 g % and > 100 mg% and several vascular risk factors. The best evidence exists for simvastatin, pravastatin and atorvastatin.
	Diabetics should be treated with diet, regular exercise, antidiabetics and insulin, as needed (B). Normoglycaemic values should be striven for. In diabetics, the importance of antihypertensive treatment with ACE inhibitors or sartans and the administration of statins is particularly high with regard to stroke prevention (B).
	Patients suffering from persistent or paroxysmal atrial fibrillation and concomitant vascular risk factors (hypertension, coronary heart disease, heart failure, age over 75 years) should receive oral anticoagulation with a target INR of 2.0–3.0 (A). In patients aged over 75 years, an INR of about 2.0 should be targeted. No anticoagulation or platelet aggregation inhibition is necessary in patients aged under 65 years with so-called lone atrial fibrillation, a rare form of atrial fibrillation, and lacking vascular risk factors. Acetylsalicylic acid (100–300 mg) is recommended for patients aged over 65 years with atrial fibrillation and without any vascular risk factors. ASA is likewise used in patients with contraindications to oral anticoagulants such as cerebral microangiopathy, incipient dementia and an elevated risk of falls.
	Oral anticoagulation for stroke prevention for longer than 3 months is not necessary in patients who have received heart valve replacement with a bioprosthesis.
	An asymptomatic patent foramen ovale with or without atrial septal aneurysm (ASA) does not require treatment (A).
	Acetylsalicylic acid is not effective for primary stroke prevention in men (A). It prevents strokes, but not myocardial infarctions in women aged > 45 years with vascular risk factors (B).
	Surgery for asymptomatic carotid artery stenosis with a degree of stenosis > 60% based on Doppler or duplex sonographic criteria significantly reduces stroke risk. However, this is only true when the combined 30-day mortality and morbidity of the intervention is less than 3% (A) (  ). The patient's life expectancy should be > 5 years.

Non-recommended treatments

	Patients should not drink alcohol for primary prophylaxis (C).
	Postmenopausal hormone replacement elevates the risk of stroke (  ) (A).
	Vitamins, particularly vitamin E, A and C, are not effective as primary prophylactics (A).
	Garlic preparations and so-called nootropics are not effective for stroke prophylaxis (B).
	Polypragmatic therapies with combinations of vitamins, ASA, statins, folic acid, trace elements are not recommended (B). On the contrary, there are reports that antioxidants (vitamin E and C) can negatively affect the efficacy of statins.
	Platelet aggregation inhibitors such as clopidogrel, ticlopidine or the combination of ASA plus dipyridamole should not be used for primary prophylaxis (B).
	ASA and oral anticoagulation should not be used in combination. The combination of 325 mg ASA/day and anticoagulation with an INR of 1.2–1.5 provides no advantage in patients suffering from atrial fibrillation (B).
	Surgery of asymptomatic carotid artery stenosis by surgeons with a complication rate of > 3% (B).
	There is currently no sufficiently valid evidence that patients suffering from asymptomatic carotid artery stenoses profit from PTA or stenting (C).
	Anticoagulation is not indicated in patients suffering from mitral valve prolapse syndrome (A).
	Treating elevated homocysteine levels with B vitamins and folic acid is not recommended (B).

Identification and treatment of vascular risk factors

Arterial hypertension

The treatment of arterial hypertension ranks highest in importance for primary stroke prevention. According to numerous high-quality studies, treatment of high blood pressure leads to a marked risk reduction for both ischaemic and also haemorrhagic strokes. Even a slight and easily achievable lowering of systolic blood pressure by 5–6 mmHg and/or of diastolic blood pressure by 2–3 mmHg leads to an approx. 40% relative risk reduction (Collins et al. 1990). The absolute RR is around 0.5% annually (NNT = 200), meaning that 200 patients with hypertension have to be treated to prevent one stroke. This effect has been observed across all age groups and hypertensive forms, also in > 80-year-olds and patients with isolated systolic hypertension (Staessen et al. 2000, Staessen et al. 2001). According to a recommendation by the WHO and the German Hypertension League, when antihypertensive treatment should start is dependent on other risk factors, end-organ damage and secondary/concomitant diseases. Table 1 provides an overview of the vascular risk factors on which the therapeutic recommendations are based.

The target range for blood pressure lowering also depends on the risk profile, and is generally lower in diabetics. The minimum targets are considered to be upper limits of systolic < 140 mmHg and diastolic < 90 mmHg. As a rule, the preventative efficacy increases in a linear fashion with the extent of blood-pressure lowering. In other words, lowering it to the optimal blood pressure range (< 120/80 mmHg) is recommended when the treatment is well tolerated.

Under all circumstances, the relevance of non-medical measures should be clearly elucidated to every patient before initiating drug therapy. Non-medical measures should always be incorporated into therapy and are particularly effective in younger persons. With regard to drug antihypertensive therapy, there are no validated differences between the following 5 classes of drugs (Droste et al. 2003, Group 2003, International Society of Hypertension Writing Group 2003) (  ):

	Angiotensin-converting enzyme (ACE) inhibitors,
	Angiotensin (AT) receptor blockers
	Beta-blockers,
	Calcium antagonists,
	Diuretics.

A comparison of the so-called conventional blood pressure-lowering drugs (atenolol, metoprolol, pindolol, hydrochlorothiazide plus amiloride) with newer drugs (enalapril, lisinopril, felodipine, isradipine) produced no significant differences in stroke rate in elderly persons (Hansson et al. 1999). Therefore, representatives of this class of drugs can all be regarded as drugs of choice. There was a trend showing a minor superiority in the prevention of cerebrovascular events in favour of AT blockers and calcium antagonists compared to the other classes of drugs (Staessen et al. 2001). Losartan is more effective than atenolol (Dahlof et al. 2002, Lindholm et al. 2002). Nevertheless, no convincing differences in overall mortality or cardiovascular events have been revealed to date. By contrast, the preventative efficacy of alpha-receptor blockers is clearly worse, which is why this class of drugs cannot be regarded as first-line therapy (The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group 2000) ( ). According to current recommendations, primary combination therapy (e.g. ACE inhibitors or ß-blocker plus diuretics) is equivalent to staged therapy (start monotherapy -> other monotherapy).

Dietary measures (low-sodium diet and a diet rich in fruits, vegetables, low-fat milk, poultry, fish and cereals) can effectively lower blood pressure, as the DASH study (Sacks et al. 2001) has shown. In this study, a special DASH diet and table salt reduction led to a mean reduction in blood pressure of around 11 mmHg. This study did not demonstrate an equivalent primary preventative impact on cardiovascular and cerebrovascular events (however, the sample sizes were not large enough).

Smoking cessation

Smoking increases the stroke risk by a factor 1.8 (Goldstein et al. 2001). Randomized studies on the effect of the smoking cessation are lacking. Epidemiological studies have shown that nicotine abstinence can substantially reduce an elevated stroke risk. Remarkably, observations showed that the vascular risk was reduced by half after 12-month abstinence; and after a further 5 years, the vascular risk profile was just slightly above that of a non-smoker (Kawachi et al. 1993, Wilson et al. 1985, Wilson et al. 1997). Surveys have shown that around 70% of all smokers would like to stop, 30% have previously undertaken at least one unsuccessful attempt to quit, but only very few of them achieve this goal. Only around 3% of smokers achieve permanent abstinence by virtue of sheer will alone. Professional help can raise this success rate. A brief educational talk with the physician including recommendations to quit smoking, statistically only leads to abstinence in 5%. Comprehensive informational material and counselling by a specialist can raise this rate to 10%. Alongside concomitant behavioural interventions, nicotine replacement therapy leads to a further doubling of this success rate and should be considered in patients with sufficient motivation and previously failed attempts at abstinence. Nicotine replacement therapy can be given in the form of dermal patch, nasal spray or chewing gum. Nicotine abstinence is a low-cost and effective instrument for primary prevention, leads to a significant risk reduction and should be proposed to every smoker ( ). In sufficiently motivated patients, nicotine replacement therapy should be given consideration.

Statin therapy for hypercholesterolaemia

Vascular primary prevention studies have shown that statin therapy can achieve a reduction in mortality and the myocardial infarction rate by 30–40% and a reduction in stroke rate by 11–30%. In the ALLHAT study, a non-significant stroke reduction of 11 % was achieved with pravastatin versus placebo (The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group 2002). In the Heart Protection Study, a significant 25% risk reduction was demonstrated (Heart Protection Study Collaborative Group 2002). An updated overview established the average RR for strokes through statin therapy at 21 % (Amarenco et al. 2004). The results of studies on the treatment of hypercholesterolaemia exclusively designed for primary stroke prevention, however, are still outstanding. In the CARDS study, 2838 diabetics were treated with either atorvastatin 10 mg or placebo. The risk reduction in relation to stroke was 48 %, where the absolute numbers were 39 strokes in the placebo group and 21 in the atorvastatin group (Colhoun et al. 2004). According to current guidelines, pharmacological lipid lowering should be made dependent on LDL levels and individual risk profile. When additional risk indicators are lacking, LDL levels should be < 130 mg/dL or, if necessary, achieved therapeutically. In the presence of diabetes mellitus (Colhoun et al. 2004), a high vascular risk profile or a CHD, the LDL level should be < 100 mg/dL. Meanwhile, it can be regarded as validated that the statin effect is also based on cholesterol-independent effects. Anti-inflammatory, plaque-stabilizing, vasodilatatory and blood-pressure lowering effects have been described in this context.

Antidiabetic therapy

Although diabetes mellitus is a relevant and independent risk factor for stroke, so far no primary prevention study has demonstrated that strict antidiabetic therapy significantly reduces the risk of strokes or other macrovascular complications. The UKPDS study, the latest randomized study on this subject, found, by contrast, that the therapy group had a 25% risk reduction for secondary microvascular diseases (Stratton et al. 2000). In the Steno-2 study, intensified antidiabetic treatment led to a 50% reduction in cardiovascular complications compared to conventional guideline-orientated therapy (Gaede et al. 2003). Lowering the antihypertensive target range can also virtually halve the stroke risk, irrespective of any antidiabetic therapy. Generally, blood pressure control to values < 130/85 mmHg is recommended. In this context, diabetics should receive therapy of the renin-angiotensin-aldosterone system (RAAS), which means that ACE inhibitors and AT-1 blockers should be preferred.

Weight reduction

Reducing obesity has an intermediate effect on stroke risk in that it positively influences other risk factors such as arterial hypertension, hypercholesterolaemia and diabetes mellitus (). However, no results from randomized studies are available on this subject.

Physical activity

Similar to the elimination of obesity, physical activity has effects on the stroke risk, especially indirect effects, achieved by the modification of other risk factors such as arterial hypertension, hypercholesterolaemia and diabetes mellitus. Moreover, beneficial effects on blood rheology and platelet reactivity have been described. Several studies have found that regular physical activity leads to a gender-independent relative risk reduction of around 40–60% (Abbott et al. 1994, Kiely et al. 1994, Lee et al. 1999). A prerequisite for this activity was that it either accelerated the heart rate or increased sweat production. Here, both the risk for ischaemic strokes as well as for cerebral haemorrhage was reduced, which was particularly attributable to the associated lowering of blood pressure. It was remarkable that no linear dose-effect relationship was obvious, but rather a constant class effect was observed. Thus, the risk reductions achieved in the Physicians Health Study and the Framingham Study with once-weekly and minor physical exercise, respectively, were similarly as effective as several times weekly or vigorous exercise.

Vitamin therapy for hyperhomocysteinaemia

Numerous studies have shown that serum homocysteine levels can be lowered by dietary modifications involving elevated nutritional supplements with vitamin B6, B12 and folic acid or by their direct intake. One large-scale study showed that enriching cereal products with folic acid was able to increase serum folic acid levels by 60% and lower homocysteine levels by 10–15%. Conversely, it has so far not been demonstrated that lowering of homocysteine serum levels can also reduce cerebro- or cardiovascular risk.

Other risk factors

Female sex hormones, whether given for contraception or postmenopausal hormone replacement therapy (HRT), increase the risk of vascular events including stroke. This is also true for oestrogen replacement in hysterectomised women. HRT has no protective effect on cardiovascular and cerebrovascular morbidity and mortality (  ) (Anderson et al. 2004, Burry 2002, Grady et al. 2002).

Migraine is a risk factor for stroke (Diener et al. 2004, Merikangas et al. 1997). However, the risk is only elevated in women suffering from migraines with aura and hypertension who also smoke and take the pill. There are no prospective prophylaxis studies available. Nevertheless, female risk patients must treat their risk factors.

With regard to primary stroke prevention, no data are available on the following, as yet unvalidated risk factors: obstructive sleep apnoea syndrome, chronic infection, chronic inflammation and depression.

Platelet aggregation inhibitors

Two large studies examined the primary prophylactic use of aspirin (Peto et al. 1988, The Steering Committee of the Physicians' Health Study Research Group 1988). A significant risk reduction for myocardial infarctions was found, but no risk reduction was shown for cerebral infarctions (  ). Intracranial bleedings were more common under ASA. Neither was the Nurses Health Study able to show any advantage of ASA for stroke prevention in women (Iso et al. 1999). The Women's Health Study did however reveal a benefit for ASA for primary prevention in women aged > 45 years (RRR = 17 %) (Ridtker et al. 2005). A large meta-analysis of five studies with over 250,000 persons receiving 75–650 mg ASA/day (Hart et al. 2000) was not able to demonstrate that ASA had any advantage, given that the annual stroke risk in these studies was very low (0.3%) and a relative risk elevation of 8% was observed for intracerebral bleedings. In the Primary Prevention Project study, diabetics showed no benefit from the prophylactic administration of ASA (Collaborative Group of the Primary Prevention Project, PPP, 2001).

Atrial fibrillation (AF)

A meta-analysis of five randomized studies investigating the primary prevention of AF showed that oral anticoagulation with a target INR of 2.0–3.0 achieved a 70% RR compared to placebo treatment (  ) (Hart et al. 1999). The absolute RR achieved through effective anticoagulation is around 3%/year, which is equivalent to an NNT of 33. Low-grade anticoagulation with a target INR of 1.5–1.9 has virtually no benefit. Even stronger anticoagulation with a target INR of 3.0–3.9 only led to an around 40% reduction of all stroke syndromes due to its elevated rate of cerebral haemorrhage. An exponential increase in the cerebral haemorrhage risk has been observed from an INR > 4.5: every ~0.5-point INR elevation from this value (i.e. to 5.0–5.5 etc.) nearly doubled the risk for cerebral haemorrhage. ASA, administered in the studies at a dosage of 75 or 325 mg/day, likewise exerted a stroke-protective effect; the risk reduction however was only around 20% (Hart et al. 1999).

Given that the stroke risk indeed is strongly dependent on AF type and vascular risk profile, a diversified primary prevention strategy is recommended (see Table 2). In patients < 65 years without any additional risk factors the stroke risk is low, i.e. there is no validated indication for an antihaemostatic therapy; optionally ASA therapy can be given. Patients < 65 with risk factors and those aged 65–75 years without risk factors have an intermediate risk and should be treated at least with ASA. By contrast, patients with a high thromboembolism risk should receive permanent and rigorous oral anticoagulation. It is currently unclear how patients aged > 80 years with AF and further risk factors should be treated. The dilemma is that these patients have an elevated thromboembolism risk, on the one hand (which requires rigorous anticoagulation), but, at the same time, they show an elevated risk of cerebral haemorrhage (which represents a relative contraindication for anticoagulation). Fearing a higher bleeding rate, elderly patients are frequently not given any anticoagulation. Hart and Halperin (2001) recommend anticoagulation with a target INR of 2–3 in patients up to 75 years and a target INR of just 2.0 from the 75th year of life onward. Modern thrombin antagonists like melagatran have an efficacy comparable to vitamin K antagonists and a lower rate of mild haemorrhagic complications ( ) (Albers et al. 2005). Serious haemorrhagic complications occur with an equal frequency. The advantages of this substance class include the fact that monitoring of coagulation is not necessary and the fixed oral dose (36 mg bid). During treatment with melagatran, approx. 6% of the patients exhibit an increase in GPT. Therefore, liver values should be monitored during the first 12 months of treatment.

Primary prevention in other cardiac diseases

Patients with congenital or acquired valvular defects or with mechanical artificial valves experience a preventative effect from oral anticoagulation (Cannegieter et al. 1995, Salem et al. 1998). The annual stroke risk is 1–4% for mechanical prostheses and 0.2–2.9% for bioprostheses. An INR of 2.5–3.5 is recommended–an empirically good compromise between optimally effective thrombosis prophylaxis and prevention of haemorrhagic complications. Patients with biovalves in the mitral position are anticoagulated for 3 months and, afterwards, treated with ASA. In patients with acute myocardial infarctions, strokes occur in the first 6 weeks in approx. 2.5% of the cases. Permanent anticoagulation should be given in myocardial infarction patients with poor ventricular function and concomitant atrial fibrillation (Hardman et Cowie 1999).

The relevance of PFO (persistent patent foramen ovale), which is present in 20–25% of all persons, has currently not yet been conclusively clarified. The stroke risk is not elevated in patients with isolated PFO. An elevated stroke risk exists only in patients with a septal aneurysm. Routine anticoagulation should also be rejected for primary prevention, as should any and all surgical or interventional (umbrella occlusion) therapies. The singular exceptions would constitute large defects with impairment of cardiac haemodynamics. In these cases, a correction is indicated on cardiological grounds, not for cerebrovascular prevention.

Patients suffering from mitral valve prolapse do not have an elevated stroke risk and therefore do not require any drug prophylaxis ( ) (Gilon et al. 1999).

Surgery for asymptomatic carotid artery stenosis

At the beginning of the 1990s, smaller randomized prospective studies were published on surgery for asymptomatic carotid artery stenoses (Hobson et al. 1993, Mayo Asymptomatic Carotid Endarterectomy Study Group 1992, The Casanova Study Group 1991). They demonstrated no benefit for surgery. However, the two largest studies on this subject to date, ACAS (Asymptomatic Carotid Atherosclerosis Study) in North America with 1600 patients (Executive Committee for the Asymptomatic Carotid Atherosclerosis Study 1995) and ACST (Asymptomatic Carotid Surgery Trial) from Europe with 3100 patients (Halliday et al. 2004), were able to consistently show a primary prophylactic effect. The absolute RR extrapolated to a period of 5 years was around 5%, which is equivalent to an annual risk reduction of 1 % (NNT = 40/5 years). In the larger ACST study, the following patient subgroups particularly profited from surgery:

	Men (absolute RR over 5 years 8.2%).
	Younger patients< 65 years (7.8%).
	Patients suffering from only moderate stenoses of 60–80% (7.4%).
	Patients suffering from strongly elevated serum cholesterol > 250 mg/dL (11.4%).

By contrast, no differences were shown with regard to patients' blood pressure or the ultrasound morphology of the plaques. One limitation that should be noted is that the surgeons in the two studies were selected according to stricter criteria. In the ACAS study, around 40% of all applicants were rejected as surgeons for study because of too high complication rates (Moore et al. 1991). This selection process led to very low perioperative complication rates of 2.7% (ACAS) and 3.1% (ACST). Thus, it is doubtful that these numbers can be viewed as representative for the totality of all surgeons. Rather, it must be assumed that the complication rates for unselected surgeons will be higher by a factor of 2–3 (Bond et al. 2003, Bond et al. 2004). The latter implies that carotid surgery thus either yields no benefit (starting from > 4% complications) or may even be harmful (starting from > 6% complications). To date, no prospective randomized studies are available on stent-supported balloon angioplasty. Case series suggest however that their periprocedural complication rate is similarly as high as with surgery. There is presently no evidence that endovascular treatment of asymptomatic carotid artery stenoses yields a reduction of stroke risk (Table 3).

Secondary prevention

The objective of any secondary prevention is to prevent recurrent cerebral ischaemia (TIA or stroke) after the initial event. Data on the prevention of further events (so-called tertiary prevention) have mostly been retrospectively concluded from results for secondary prevention; specific studies on this subject are not available to date.

Epidemiology

Around 80–85% of patients survive their first stroke in the acute phase (Grau et al. 2001, Wolf et al. 1992). Of these patients, 8–15% suffer their second event within the first year. In this context, the risk is greatest during the first few weeks and continues to decline steadily proportionate to the time elapsed since the index event (Hill et al. 2004, Johnston et al. 2000, Lovett et al. 2004, Weimar et al. 2002). Patients suffering from multiple vascular risk factors or those with concomitant CHD or PAOD are particularly at risk. Patients suffering from cerebral symptoms are particularly at risk for TIAs compared to those with retinal symptoms (amaurosis fugax) as well as patients over 60 years with duration of symptoms longer than 10 min and symptoms involving paralysis or speech disorders. The greatest risk exists during the first three days after a TIA.

Tests and examinations

Obligatory

Neurological and full physical examination, CT or MRI (DD ischaemia, bleeding, SAB etc.), ultrasound examination of the vessels supplying the brain, (if the findings are unclear: CTA or MRA), lab work, ECG, echocardiography (in patients with territorial infarction).

Optional

24-hour ECG, 24-hour blood pressure measurement, special laboratory tests (exclusion of vasculitis and coagulation disorders).

Treatment of risk factors

	Rigorous treatment of arterial hypertension reduces the stroke risk (  ) (A). The combination of perindopril plus indapamide is significantly more effective than placebo ( ) and eprosartan is significantly more effective than the calcium antagonist nitrendipine ( ). In post-stroke patients, ramipril reduced vascular endpoints.
	However, all antihypertensives are probably effective in secondary stroke prevention (B).
	The treatment of diabetes mellitus reduced the stroke risk (C), although this has not been previously well investigated in prospective studies.
	Statins should be used in patients suffering from focal cerebral ischaemia and CHD independent of their baseline levels of LDL cholesterol (  ) (A). Target values for LDL should be between 70 and 100 mg%.
	Simvastatin (40 mg) can be given to patients suffering from focal cerebral ischaemia without CHD, which basically lowers the general vascular myocardial infarction risk ( ) (B). Probably, the other statins are also effective (C).
	The treatment of hyperhomocysteinaemia with Vitamin B6, B12 and folic acid is not effective in secondary stroke prevention () (B).
	Postmenopausal hormone replacement is not effective in secondary stroke prevention () (B).

Hypertension

The PROGRESS study (Progress Collaborative Group 2001) was the first large randomized study on treatment with antihypertensives for secondary prevention.

In this study, 6105 patients were treated with either the ACE inhibitors perindopril and indapamide or placebo (started approx. 2 months) after a stroke or TIA. After a four-year follow-up period, blood-pressure lowering treatment had led to a reduction of blood pressure by 9/4 mmHg. The absolute risk reduction for suffering a stroke was 4% (10 v 14%), which translates to a significant relative risk reduction of 28% (p < 0.0001). The rate vascular of events was also lowered by 26%. Interestingly, hypertensive and non-hypertensive patients benefited from treatment to an equal extent (the hypertensive/non-hypertensive limit was set at an unusually high 160/90 mmHg). The combination of ACE inhibitors and diuretic reduced blood pressure by 12/5 mmHg and the stroke rate by 43%. Perindopril alone, however, was not effective. This study indicates that all patients with a post-cerebrovascular event (also normotensives) can benefit from the administration of perindopril in combination with the diuretic indapamide. A previous meta-analysis (Gueyffier et al. 1999, The INDANA Project Collaborators 1997) with 6772 patients and a medium-term follow-up of 1.8 years found that antihypertensive treatment had a lesser effect than primary prevention (237 v 270 strokes). Nevertheless, it remains unclear whether this result depended on blood-pressure lowering only or on the specific therapy with an ACE inhibitor and a diuretic. The MOSES study showed that antihypertensive therapy with the angiotensin-receptor blocker eprosartan was significantly more effective than the calcium antagonist nitrendipine in stroke patients (Schrader et al. 2005). An identical blood pressure lowering effect was achieved with the two compounds, implying that sartans may additionally possess pleiotropic properties. A similar tendency was demonstrated in the ACCESS study, where candesartan led to a more pronounced reduction of vascular events than placebo (Schrader et al. 2003).

Hypercholesterolaemia

The Heart Protection Study (Heart Protection Study Collaborative Group 2002) investigated statins. In this context, a group of 20,536 high-risk patients was shown to have a 25% risk reduction for strokes under treatment with 40 mg simvastatin compared to placebo ( ). The prevetive effect was independent of initial cholesterol levels and of the extent of cholesterol lowering. A large multicentre trial with atorvastatin (SPARCL) for secondary prevention of cerebrovascular events was recently finished. At this time, routine secondary prophylaxis with statins can be recommended for patients suffering from cerebrovascular diseases and vascular risk factors even with normal serum cholesterol. Despite very promising approaches however, a euphoric stance towards the use of lipid-lowering drugs in stroke prevention is not presently justified, given that a secondary preventative study with the fibrate gemfibrozil demonstrated a reduction of myocardial infarction rate, but not stroke frequency (Rubins et al. 1999).

Other therapies

Currently, several prospective studies on the use of vitamins (E, B6 and folic acid) have not yet been concluded (Hankey et Eikelboom 1999). The VISP study showed that stroke patients with elevated homocysteine have no benefit from therapy with B vitamins and folic acid (Toole et al. 2004). A second study with a longer follow-up period is still ongoing. The notion that postmenopausal hormone replacement therapy (HRT) can have a cardiovascular protective effect is unsupportable in the secondary prophylaxis of cerebrovascular diseases. A study by Viscoli et al. (2001) showed that female patients suffer an increased rate of fatal strokes from HRT and have a worse prognosis with regard to hindering a nonfatal stroke.

The authors conclude that post-stroke HRT not only is not helpful, but also is contraindicated in the light of the negative results.

Platelet aggregation inhibitors

	In patients suffering from focal ischaemia, platelet aggregation inhibitors are effective in secondary prevention (  ) (A). This applies to ASA (50–150 mg), ASA plus dipyridamole and clopidogrel (75 mg) (A).
	In patients after TIA and ischaemic stroke and with a low risk of recurrence (< 4%/year), the daily administration of 50–150 mg acetylsalicylic acid is recommended (B). In patients with a high risk of recurrence (· 4% per year), the twice-daily administration of the fixed combination of 25 mg acetylsalicylic acid plus 200 mg sustained-release dipyridamole is recommended (B).
	In patients with a high risk of recurrence (4% per year) and additional PAOD, clopidogrel 75 mg is recommended (C).
	In patients with contraindications to or intolerance to ASA, clopidogrel is recommended (A).
	In patients who develop gastric or duodenal ulcers under ASA prophylaxis, ASA administration in combination with a proton pump inhibitor is recommended after a period off drug (B). At doses > 150 mg, ASA leads to an elevated risk of haemorrhagic complications ( ).
	The combination of 75 mg ASA and 75 mg clopidogrel is not more effective than monotherapy with clopidogrel, and does lead to increased haemorrhagic complications (  ).
	The duration of treatment with platelet aggregation inhibitors has not previously been examined beyond of the 4th year after the initial event. In theory, this prophylaxis should be given for life, if tolerated (C).
	If a recurrent ischaemic event occurs under ASA, the pathophysiology and risk of recurrence should be re-evaluated. If a cardiac embolic source is established, oral anticoagulation is given. If the risk of recurrence has not changed (4%/year), prophylaxis with ASA should be continued (C). Otherwise see above.
	Patients suffering from TIA or a stroke and acute coronary syndrome should be treated with a combination of 75 mg clopidogrel and 75 mg ASA for a period of 3 months ( ) (C).

The emphasis in secondary stroke prevention to date has been on platelet aggregation inhibitors. Several meta-analyses have shown that platelet aggregation inhibitors can make a major contribution to stroke prevention (Antiplatelet Trialists' Collaboration 1994, Antithrombotic Trialists' Collaboration 2002, Patrono et al. 1998). The only area where a lack of clarity prevails is "which drugs should be used at which dosage." Meta-analyses have shown that platelet aggregation inhibitors reduce the risk of a nonfatal stroke by 23% (from 10.8% to 8.3% over 3 years) in patients after TIA or stroke (Antithrombotic Trialists' Collaboration 2002). The combined vascular endpoint (stroke, myocardial infarction, vascular death) is reduced by 17% (from 21.4% to 17.8% over 29 months;   ). The actual benefit of platelet aggregation inhibition might actually even be higher, since the elderly, patients suffering from completely different causes, e.g. atrial fibrillation and carotid artery stenoses, included in the comprehensive stroke studies of the meta-analysis were those who would benefit even more from other forms of secondary prevention.

A total of 11 placebo-controlled studies have been conducted on ASA for secondary prevention after TIA or stroke. One meta-analysis produced a relative risk reduction of around 13% (95% confidence interval 6–19%) for a combined vascular endpoint (vascular death, stroke, myocardial infarction; Algra et van Gijn 1999). Various meta-analyses found no difference between the different dose ranges (Algra et van Gijn, 1999, Antithrombotic Trialists' Collaboration 2002, Diener 1998). The FDA has established that any aspirin dose between 50 and 325 mg can be recommended (Department of Health and Human Services and Food and Drug Administration 1998). As in most European countries, therapy with 100 mg ASA per day has currently become established in Germany. In this context, it is important to know that subjective gastrointestinal side effects (such as nausea, dyspepsia etc.) as well as haemorrhagic complications are dose-dependent (Topol et al. 2003, Yusuf et al. 2001). At ASA doses of > 150 mg/day, the risk of haemorrhagic complications increases significantly (Topoi et al. 2003). Clopidogrel can be given to patients who develop side effects under ASA (see below). If a patient develops a gastric or duodenal ulcer under ASA, prophylaxis with ASA in combination with a proton pump inhibitors continued after a waiting period leads to fewer haemorrhagic complications than prophylaxis with clopidogrel (Chan et al. 2005) ( ).

Another platelet aggregation inhibitor is clopidogrel. This drug has been examined with regard to its post-stroke prophylactic efficacy in the CAPRIE study. CAPRIE was a double-blind, randomized study with around 20,000 patients, in whom 75 mg clopidogrel were compared with 325 mg ASA (CAPRIE Steering Committee 1996). Qualifying events were stroke, myocardial infarction or symptomatic PAD. The primary endpoint was a recurrent vascular event (myocardial infarction, stroke or vascular death). Clopidogrel lowered this combined endpoint by relative 8.7% (p < 0.043). The absolute annual risk reduction was 0.51 %. The three patient subgroups in the study (myocardial infarction, stroke and peripheral arterial disease) profited in different ways. For example, there was a greater risk reduction with clopidogrel in patients suffering from PAD treated (23.8%) and/or with PAD plus stroke plus myocardial infarction (22.7%). The safety of clopidogrel is good. Severe neutropenia was observed in only 0.1 %. Cases with thrombotic-thrombopenic purpura have also been described (Bennett et al. 2000). Their incidence was equivalent to the spontaneous incidence of TTP in the normal population. The gastrointestinal bleeding rate was significantly lower in the clopidogrel group than in the ASA group (1.99 v 2.66%). Gastrointestinal side effects were significantly less frequent in patients taking clopidogrel compared to ASA (15% v 17.6%).

The recently published MATCH study compared the prophylactic efficacy of clopidogrel versus the combination of 75 mg clopidogrel plus 75 mg ASA in high-risk patients with previous TIA or ischaemic stroke. Primary endpoint was the occurrence of myocardial infarction, stroke or vascular death and/or a hospitalization due to a recurrent vascular event. During the 18-month follow-up period, no statistically significant difference emerged for this endpoint. However, the haemorrhagic complication rate was significantly different; life-threatening haemorrhagic complications were significantly more frequent with the combination (2.6% v 1.3%; Diener et al. 2004).

Dipyridamole is the third clinically relevant platelet aggregation inhibitor. In 1987, one of the first placebo-controlled European studies (ESPS 1) published its results on 2500 patients who had suffered a stroke or a TIA (The ESPS Group 1987). There was a group of patients who received 990 mg ASA per day and 225 mg dipyridamole, in addition to a group of patients who received placebo. The primary endpoint was stroke or all-cause death. Within two years, this endpoint was lowered by 33% in the treatment group. Four additional studies with smaller sample size and non sustained-release dipyridamole were negative (American-Canadian Co-surgical Study Group 1985, Bousser et al. 1983, Guiraud-Chaumeil et al. 1982). The largest study to date has been the ESPS-2 with 6602 patients (Diener et al. 1996, Diener et al. 1997). This study had four arms: ASA (25 mg bid), sustained-release dipyridamole (200 mg bid), ASA plus sustained-release dipyridamole (25 mg + 200 mg bid) and placebo.

The qualifying event was stroke or TIA. The primary endpoint was stroke and/or death within 2 years. With regard to the endpoint "recurrent stroke", the combination treatment produced a relative risk reduction of 23% (3% absolute) compared to ASA, compared to placebo a 37% relative risk reduction (5.8% absolute). ASA alone led to a stroke risk reduction (RR) of 18% (2.9% absolute) and dipyridamole alone to an RR of 16% (2.6% absolute). With regard to the endpoint "stroke and death", the risk reduction was 13% (2.6%), 24% (5.6%), 13% (3%) and 15% (3.5%). Major haemorrhagic complications of all kinds occurred in 8.7% of patients receiving the combination and/or in 8.2% of those receiving ASA alone. The bleeding rate (severe bleedings) for dipyridamole was 4.7% and for placebo 4.5%. Headaches were the reason for drop-outs in patients under combination therapy at 8.1%, in 8% with dipyridamole alone, in 1.9% with ASA alone and in 2.4% with placebo. Cardiac events were not more frequent in the dipyridamole group than in the groups treated with ASA (Diener et al. 2001). The fixed combination of 25 mg ASA plus 200 mg sustained-release dipyridamole is approved in Germany.

No direct comparisons exist between clopidogrel and the combination of ASA plus dipyridamole, but are currently being conducted (PRoFESS). For relative and absolute risk reductions, see Table 4.

Risk model for identifying patients with a high recurrent stroke risk

Based on a post-hoc subgroup analysis of the CAPRIE study (Ringleb et al. 2004), the risk factors and concomitant diseases were identified in a logistic regression analysis predictive of stroke recurrence. These factors were used to develop a predictive model in Essen. Table 5 shows the individual factors and their weighting. The maximal achievable score is 10. A linear increase in the frequency of recurrent stroke is given up to a score of 7. The total population of patients suffering from stroke in the CAPRIE study and in ESPPS-2 study can be subdivided with a risk of recurrence of 4 %/year. In the patient group with a low risk of recurrence (0–2 points), there was no difference in efficacy between ASA and clopidogrel. In the high-risk patients (3–6 points), clopidogrel was significantly more effective than ASA (Diener et al. 2005 b). The number of patients with a score> 6 was low, and this led to wide confidence intervals. A post-hoc-analysis of the ESPS-2 study using the Essen Risk Score demonstrated, from a risk score of 3 points, that the combination therapy of ASA plus dipyridamole was clearly superior to ASA monotherapy (Diener et al. 2005 a). This analysis showed that secondary prevention strategies based on the risk of recurrence are sensible. Nevertheless, these calculations have to be validated in a prospective study.

Glycoprotein llb/llla antagonists

Glycoprotein (GP) llb/llla antagonists should not be used for secondary stroke prevention (A) ((). They are not more effective than acetylsalicylic acid, while their use is associated with a significantly elevated bleeding risk.

Glycoprotein llb/llla antagonists belong to the family of plasma membrane receptors (integrins). They are located only on the platelets and their precursors. Inhibition of these receptors prevents the formation of fibrinogen bridges and platelet aggregation. Three intravenous GP llb/llla antagonists are available: abciximab, eptifibatide and tirofiban. They are effective and reduce early mortality in acute coronary syndromes (Topol et al. 1999). In stroke patients, abciximab has produced initial data for safe use (Burton 2003); the same applies to tirofiban (Burton 2003, Junghans et al. 2001, Seitz et al. 2003) and in combination with rtPA (Seitz et al. 2003) with potential efficacy–further studies are ongoing (SATIS).

All studies that have examined oral glycoprotein llb/llla inhibitors in stroke patients had to be prematurely terminated because of elevated bleeding rates (BRAVO; Topol et al. 2003). As a consequence, there are currently no further studies being conducted on secondary prophylaxis of stroke.

Anticoagulation

Early secondary prophylaxis

	In patients with TIA or ischaemic stroke, PTT-based heparinisation is not indicated for the treatment of the stroke (A) ().
	This is also true for patients suffering from non-rheumatic atrial fibrillation (ß) (A).
	In special indications (see below), the risk-benefit ratio must be weighed up very carefully.

Various antithrombotic agents have been recommended for the early phase of ischaemic stroke, especially when thrombolysis is not possible. They are intended to prevent 1) progression of acute thromboembolic events and/or recurrent thromboembolic events, and 2) minimize venous thromboembolic complications (pulmonary embolism/deep vein thrombosis). Older studies in this area with conflicting findings have not been conclusively able to differentiate between different stroke types with varying degrees of risk, among other things, because of their often deficient designs and too small sample sizes. A large meta-analysis has shown that early anticoagulation in patients suffering from ischaemic infarctions is not effective for secondary prophylaxis. In spite of this, acute anticoagulation with heparin/heparinoids at varying dosage continues to enjoy great popularity in North America (Adams 2002) and in Germany (Daffertshofer et al. 2003).

Nevertheless, various controversies have led to a reduction in their use in most German hospitals (Daffertshofer et al. 2003, Grips et al. 2003, Hamann et Diener 2001). From the point of view of evidence-based medicine, there are no indications for acute PPT-guided i.v. heparin. Neither is there any indication for subcutaneous heparin/heparinoid–administered within the first 48 hours: IST (1997) did not show any protective effect for 5000 U or for 12500 U (bid) unfractionated heparin compared to placebo and/or aspirin 300 mg, even when both dosage arms were analyzed together: A significant reduction of stroke recurrences (from 3.8% to 2.9%) compared to placebo observed 14 days after initiation of therapy was neutralized by a significant increase in haemorrhagic strokes (from 0.4 to 1.2%; International Stroke Trial Collaborative Group 1997). The same was found for a subgroup of patients suffering from non-valvular atrial fibrillation as the ischaemic cause (Saxena et al. 2001). In TOAST (The Publications Committee for the Trial ORG 10172 In Acute Stroke Treatment [TOAST] Investigators 1998), heparinoids produced a rate of 1.1 %/week in 7 of 628 patients suffering from a second event within 7 days compared to only 0.6% second strokes per week in CAST (Chinese Acute Stroke Trial Collaborative Group 1997), but 2.2% in the IST control group. Based on the comparable individual evaluation of these studies, a minor effect is thus conceivable, but has not been demonstrated by either the studies themselves or by a systematic meta-analysis. After a rather positive estimation for heparin showing approx. 50% efficacy for heparin treatment in the prevention of strokes, Swanson concluded that 100 patients a week have to be treated to prevent a stroke (Swanson 1999). Nevertheless, the same author also made the understandable statement that the complications of heparin therapy (1.93% major bleeding under heparin v 0.44% in the controls) indicate that heparin should not be administered after acute strokes.

Interestingly, a recently published Scandinavian study showed that aspirin is equivalent to heparin administration (15000 IU LMWH subcutaneous) in "acute cardioembolic stroke" and might encourage further doubts about the benefit of heparin therapy (Berge et al. 2000). Pragmatic indications that have not been well investigated include:

	Detection of cardiac thrombus (e.g. by TEE),
	Floating thrombus in the aortic arch or in the vessels supplying the brain,
	Dissection of the internal carotid artery or vertebral artery
	Mechanical heart valves.

Prophylaxis of pulmonary embolism (PE) / deep vein thrombosis (DVT)

Low-dose heparin and low-molecular-weight heparins reduce the risk of deep vein leg thrombosis in stroke patients who have paretic leg and are bed-ridden (  ) (B). This has been demonstrated for fraxiparin, tinzaparin and certoparin ( ).

DVT and LE are known complications of stroke, exhibiting a mortality of up to 5% (Antiplatelet Trialists 1994). The efficacy of antithrombotic prophylaxis has, however, been less well investigated than in other diseases. Sandercock et al. (1993) compared 10 studies in a meta-analysis and described a reduction of DVT by 80% and of PE by 58%. The IST study also found a significant reduction in PE from 0.8 to 0.5% under therapy with s.c. heparin (p < 0.05), while ASA was not effective (International Stroke Trial Collaborative Group 1997). A more recent meta-analysis by Bath et al. (2003), however, did demonstrate that a significant reduction in DVT and PE was associated with a significant increase in haemorrhagic complications (Bath et al. 2000). Despite this evidence, the ACCP recommended in 2004 that stroke patients with limited mobility be given a low-dosed s.c. heparin or heparinoid alongside ASA in patients without thrombolytic therapy (if necessary, stopping for 24 hours) in stroke patients with ICH, compression stockings at the beginning, and s.c. heparin/heparinoids starting from the second day (C).

Anticoagulation of non-cardiogenic cerebral ischaemia

Oral anticoagulation after an ischaemic stroke is not more effective than the administration of ASA. It leads to an elevated rate of haemorrhagic complications and can therefore not be recommended (A) ().

New data show that oral anticoagulation is not effective in the prevention of secondary events of vascular thromboembolic causes. The SPIRIT study investigated high-dose anticoagulation with 30 mg ASA per day and an INR of 3–4.5 (The Stroke Prevention in Reversible Ischaemia Trial [SPIRIT] Study Group 1997) in patients without a cardioembolic cause of their stroke. The study was stopped due to an elevated bleeding rate under oral anticoagulation. Afterwards, many authors expressed the opinion that oral anticoagulation is not indicated in stroke of non-cardiac origin. The WARSS study showed the same rate of ischaemic events and bleeds under ASA as under oral anticoagulation (INR 1.4–2.8) in patients after ischaemic stroke and exclusion of any cardiac embolic source (Mohr et al. 2001). This difference can be explained by the different intensities of anticoagulation. If strong anticoagulation is given, as was the case in the SPIRIT trial, then a markedly higher rate of bleeding occurs. If an INR around 2 is selected, the bleeding rates are comparable with those under ASA. Moreover, serious haemorrhagic complications were observed in 1.5% of the WARSS study patients under ASA. Subgroup analyses (PFO, antiphospholipid antibody syndrome or ischaemic posterior circulation) yielded no benefit of anticoagulation compared to ASA (Homma et al. 2002).

Careful anticoagulation for stroke of non-cardiac origin was confirmed by a recently published paper from Holland where a target INR of 2.5–3.5 for strokes of presumably arterial origin, particularly in elderly patients and early initiation of therapy, led to an elevated bleeding risk (Torn et al. 2001). In this retrospective analysis, the bleeding rate was high, at 3.9%, and comparable with the high bleeding rate seen in the SPIRIT trial. Importantly, a marked reduction in serious complications can be achieved a by special anticoagulation education and monitoring (Ansell et al. 2001). Specific guidelines for handling anticoagulation and the respective problems were recently published in Chest and can be recommended (Singer et al. 2004). When informing patients, a frequency of approx. 2% should be assumed for serious haemorrhagic complications (including intracerebral bleeds) and 0.5% for anticoagulation-related deaths per year.

The WASID study compared ASA 1300 mg with oral anticoagulation (INR 2–3) in patients with symptomatic intracranial stenoses. There was no significant difference between the treatment arms (Chimowitz et al. 2005). The European-Australian Stroke Prevention Trial compared anticoagulation (INR 2–3) with aspirin (30–325 mg) or aspirin plus dipyridamole (Algra et al. 2003).

Anticoagulation for cardiogenic thromboembolic events

	In patients with a cardiac embolic source, in particular with atrial fibrillation, oral anticoagulation with INR levels of 3.0 is recommended ( ) (A).
	After TIA and mild ischaemic stroke and atrial fibrillation, oral anticoagulation can be initiated within 3–5 days (C).
	In patients with mechanical heart valves, anticoagulation is continued with INR levels between 2.0 and 3.5 () (C).
	In patients with biological valves, temporary anticoagulation for 3 months is recommended () (C).

The evidence on oral anticoagulation is more reliable, although it is less conclusive than for primary prevention in this situation. In 1993, the European Atrial Fibrillation Trial (1993) conducted a small randomized study investigating the efficacy of oral anticoagulation (INR 3.0–4.5) in post-stroke patients with atrial fibrillation. In recurrent stroke, a 70% risk reduction was achieved under oral anticoagulation compared to 15% under ASA; notably however, most patients were included several weeks (to 3 months) after the qualifying event. A meta-analysis of 21 studies on early anticoagulation in over 23,000 patients produced no additional benefit (Hart et al. 2002). The data show that 80 strokes per 1000 treated patients were avoided - the haemorrhagic complication rate was 20/1000. The ATRIA study confirmed these numbers (Go et al. 2000). In contraindications to oral anticoagulation, ASA (dose 100–300 mg) can be recommended, as for primary prevention (B). Risk stratifications here make sense. Post-hoc analyses have shown that the best ratio between reduction of ischaemic events and the prevention of haemorrhagic complications are achieved at an INR of 3.0 (Gorter for the Stroke Prevention in Reversible Ischaemia Trial [SPIRIT] and European Atrial Fibrillation Trial [EAFT] Study Groups 1999).

For secondary prevention in post-TIA and stroke patients with atrial fibrillation, 36 mg melagatran bid were equally effective as oral anticoagulation with warfarin at a INR of 2.0–3.0 (  ). Also, the haemorrhagic complications were identical (Albers et al. 2005, Executive Steering Committee on behalf of the SPORTIF III Investigators 2003).

Carotid artery–TEA and Stent

	Patients with high-grade symptomatic carotid artery stenoses should undergo an endarterectomy (  ) (A). The benefit of surgery increases with the degree of stenosis between 70 and 95 %. The benefit of surgery is less at a degree of stenosis of 50-70 %, in subtotal stenoses, in women and when the operation is conducted beyond the 2nd week after the index event ( ) (B).
	The benefit of surgery is lost at complication rates of > 6 % (  ).
	The period between event and operation should be bridged with platelet aggregation inhibitors. ASA should continue to be given before, during and after the intervention (B). Clopidogrel should be replaced with ASA at the latest 5 days before surgery (C).
	Neurosonographic procedures, MRA or CT angiography are sufficient for verifying the diagnosis of carotid artery stenosis ( ). A DSA is usually not necessary (B).
	The CARESS study found that the combination of clopidogrel plus ASA achieved a significant reduction of asymptomatic microembolisms in high-grade symptomatic carotid stenosis.
	At present, carotid angioplasty with stenting is not a routine procedure. Stenting should be considered in patients suffering from recurrent stenoses after TEA, high-grade stenoses after radiotherapy or high-carotid stenoses poorly accessible to surgical intervention (C). Isolated angioplasty without stenting should no longer be conducted given its high restenosis rate (C).
	Prophylaxis with clopidogrel (75 mg) plus ASA (100 mg) should be given before, during and after stenting for 1–3 months (B).

Two large randomized landmark studies have shed light on the indication for surgery of symptomatic carotid artery stenoses (Barnett et al. 1998, European Carotid Surgery Trialists' Collaborative Group 1991 and 1998, Ferguson et al. 1999, Rothwell et al. 1999). If both studies are assessed together, a relative risk reduction of 60–80% is found for surgery, compared with medical therapy alone in patients with over 70% symptomatic carotid artery stenosis. Patients with less than 50% carotid artery stenosis do not benefit from surgery. In 50–69% stenosis, the advantage of surgery is very small and only men benefit. The perioperative complications should be less than 5.8–7% (30-day complication rate). Postoperatively lower ASA doses (81 or 325 mg) are to be preferred over the higher doses (650 or 1300 mg) commonly given in North America, (Taylor et al. 1999). Perioperative, prophylaxis with ASA should be continued. The benefit of surgery is lost when the intervention is performed later than 14 days after the initial event (Rothwell et al, 2004).

The CARESS study found that the combination of clopidogrel plus ASA achieved a significant reduction of asymptomatic microembolisms in high-grade symptomatic carotid artery stenoses (Markus et al. 2005).

The data on stenting with or without balloon dilatation of the carotid artery stenosis have not been verified. Initial results suggest that similar outcomes to surgery are achieved (CAVATAS Investigators 2001, Yadav et al. 2004). Long-term follow-up data are not available. Therefore, patients suffering from high-grade stenoses should preferably be randomized in the SPACE study. The benefit of protection systems during stenting has not so far been demonstrated. All data on this issue derive from non-randomized studies.

Intracranial stenoses

In patients suffering from high-grade intracranial stenoses or occlusions, anticoagulation with an INR of 3.0 is not more effective than administration of 1300 mg ASA ( ), but leads to an increased rate of haemorrhagic complications and can therefore not be recommended. In view of the poor tolerability of 1300 mg ASA we recommend prophylaxis with 100–300 mg ASA (C).

In patients with recurrent events, stent implantation can be considered (C). Afterwards, they should receive 75 mg clopidogrel and 100 mg ASA over a period of 1–3 months (C).

The WASID-II study enrolled 569 patients suffering from intracranial stenoses who were treated with either 1300 mg ASA or oral anticoagulation (INR 2–3). The study was terminated prematurely due to the elevated bleeding rate in the warfarin therapy arm (Chimowitz et al. 2005). Therefore, prophylaxis should be given with ASA. In view of the rate of intolerances at a dose of 1300 mg ASA, we recommend a lower dose. Predictors for a recurrent ischaemic event were the extent of stenosis, stenosis in the vertebro-basilar region, and female sex. Contrary to expectation, there is no benefit from maintaining blood pressure > 140/90 mmHg. If further ischaemic events occur under ASA administration, stent implantation can be considered.

Patent foramen ovale (PFO)

Patients suffering from cryptogenic stroke and PFO with or without atrial septal aneurysm.

	In patients with PFO alone, irrespective of the size, and first cerebral ischaemic event, prophylaxis with ASA (100–300 mg) is given (B).
	If there is a recurrence under ASA or a PFO exists concomitant with atrial septal aneurysm, oral anticoagulation with an INR of 2.0–3.0 (for at least 2 years) is recommended (C).
	If there is a further recurrence or contraindications to oral anticoagulation, interventional PFO occlusion should be performed (C).

Especially in younger stroke patients, it is often a question of whether to treat patent foramen ovale (PFO). There is currently little published, evidence-based data on this. A large European multicentre trial on secondary prevention under ASA (325 mg) produced a very low risk of recurrence, which does not justify surgical intervention or the placement of a closing device (Mas et al. 2001). A recent recommendation by the American Academy of Neurology (Messe et al. 2004) established that a PFO is not associated with an elevated risk for death or stroke ( ). Only patients suffering from an additional intraseptal aneurysm had an elevated stroke risk. The European multicentre trial on natural recurrence rates under ASA (325 mg/day) produced a low recurrence rate of 0.6% per year for pure PFO (Mas et al. 2001) and a risk of 6% per year for PFO and intraseptal aneurysm. We recommend an ASA dose of 100–300 mg/day. With regard to recurrent strokes, the PICSS study produced no differences between anticoagulation with warfarin and the administration of 325 mg ASA (Homma et al. 2002).

Many Cardiology centres support the implantation of so-called PFO closure devices in patients suffering cryptogenic stroke. This technically elegant method of mechanical PFO occlusion must be viewed critically, not only given the low natural recurrence rate under ASA, but also because the early publications (Windecker et al. 2000) report that there is an astonishingly high recurrence rate of approx. 3.4% recurrences per year. A review of 16 published studies yielded a risk of 1.5–7.9% for complications relating to catheder occlusion with a one-year recurrence rate for strokes of 0–4.9%, while the 1-year-stroke recurrence risk under conservative therapy was 3.8–12% (Khairy et al. 2003). This comparison is complicated by the fact that the review counted global TIA, minor and major stroke as recurrences, whereas the complications were divided into severe (death, major bleeding, cardiac surgical revision and pulmonary embolism–1.5%) and milder (arrhythmias, device fracture, embolisation, thrombosis and air embolism–7.9%). Here, the latter, milder complications do indeed appear threatening. One Italian publication reported a low recurrence rate with PFO occlusions of 22% (1 month) to 9% (12 months), verifiable right-left shunt, 8% atrial fibrillation and impairment from nickel toxicity in 6% of the patients (Anzola et al. 2004). The only group of patients that have a clearly elevated stroke risk are those with a PFO and an intraseptal aneurysm (Messe et al. 2004). Several multicentre trials are presently comparing catheder occlusion versus conservative therapy. No validated therapeutic recommendation can be given until these date become available.

Further therapeutic recommendations

Ineffective therapies

	Extra-intracranial bypass (); except for rare special indications such as bilateral carotid artery occlusions with insufficient collateralization or Moyamoya syndrome)
	Combination of anticoagulation and platelet aggregation inhibition ()
	Anticoagulation with an INR of > 3.5 after a stroke
	Carotid artery surgery or stenting in < 50% stenoses
	Carotid artery TEA in patients without prophylactic potential, uncontrolled hypertension or short life expectancy

Consensus procedure:

Modified Delphi Technique. Discussed and approved at a meeting of the Consensus Group on 21 December 2004 in Frankfurt, Germany. Revised by the Guidelines Commission of the DGN.

Co-operative partners and sponsors

This Guideline is issued jointly by the German Neurological Society (Deutsche Gesellschaft für Neurologie; DGN) and the German Stroke Society (Deutsche Schlaganfallgesellschaft; DSG). The Guideline was written without industry support or influence. The costs were borne by the German Neurological Society.

Validity:

Issued on September 2005

Expert committee

Diener HC, Department of Neurology, Essen University Hospital

Allenberg J-R, Vascular Surgery, University of Heidelberg

Bode C, Cardiology, University of Freiburg

Busse O, Neurological Clinic, Minden

Forsting F, Neuroradiology, Essen University Hospital

Grau AJ, Department of Neurology, Ludwigshafen Hospital

Haberl RL, Neurological Clinic, Munich-Harlaching

Hacke W, Department of Neurology, University of Heidelberg

Hamann GF, Neurology, Dr Horst Schmidt Hospital, Wiesbaden

Hennerici M. Department of Neurology, Heidelberg University Hospital, Mannheim Faculty

Grond M, Neurological Clinic, Siegen

Ringelstein B, Department of Neurology, Münster University Hospital

Ringleb PA, Department of Neurology, Heidelberg University Hospital

Chairman: Professor Hans-Christoph Diener, Department of Neurology, Essen University Hospital, Hufelandstrasse 55, 45147 Essen, Germany; e-mail: hans.diener@uni-duisburg-essen.de

Literature