SUBARACHNOID HEAMORRHAGE
The sudden onset of a severe headache in a patient should be regarded as subarachnoid haemorrhage until proven otherwise.
Subarachnoid haemorrhage occurs when bleeding is primarily within the subarachnoid space rather than into the brain itself. It represents about 5–10% of all non-traumatic intracranial haemorrhage with an incidence of approximately 15 per 100000 population.
Apoplectic death has been mentioned in the earliest medical writings but its relationship to intracranial haemorrhage and cerebral aneurysm was not established until the latter part of the seventeenth century. The introduction of cerebral angiography by Moniz and Lima in Lisbon in 1927 allowed the diagnosis of cerebral aneurysm to be made in living patients who had sustained subarachnoid haemorrhage. Pioneering surgery in the 1930s and 1940s, by Krayenbuhl in Switzerland and Dandy in North America, showed that aneurysms could be treated operatively, although at that time with considerable morbidity and mortality. Consequent improvements in microsurgical techniques and neuroanaesthesia have considerably improved the safety of surgery.
Causes of subarachnoid haemorrhage
The most common cause of subarachnoid haemorrhage in adults is rupture of a berry aneurysm (70%). Subarachnoid haemorrhage in children is much less common than in the adult population and the most common paediatric cause is rupture of an arteriovenous malformation. Rare causes of subarachnoid haemorrhage include bleeding from a tumour, bleeding disorders, blood dyscrasias and rupture of a spinal arteriovenous malformation The aetiology of subarachnoid haemorrhage remains undiscovered in approximately 15% of cases after thorough clinical and radiographic study. These patients often have associated intracranial vascular atherosclerosis and hypertension.
Subarachnoid haemorrhage— presenting features
Headache
The sudden onset of a severe headache of a type not previously experienced by the patient is the hallmark of subarachnoid haemorrhage. A relatively small leak from an aneurysm may result in a minor headache, sometimes referred to as the ‘sentinel headache’, as this may be the warning episode of a subsequent major haemorrhage from the aneurysm. Naturally, recognition of a possible minor ‘warning’ haemorrhage is essential to avert a possible later catastrophic bleed, although many are only recognized in retrospect.
Diminished conscious state
Most patients have some deterioration of their conscious state following subarachnoid haemorrhage. This varies from only a slight change when the haemorrhage has been minor to apoplectic death resulting from massive haemorrhage. It is a common cause of sudden death in young adults.
Meningism
Blood in the subarachnoid cerebrospinal fluid will cause the features of meningism—headache, neck stiffness, photophobia, fever and vomiting. Irritation of the nerve roots of the cauda equina, which occurs when the blood extends down to the lumbar theca, may result in sciatica-type pain and low back discomfort.
Focal neurological signs
Focal neurological signs may occur in subarachnoid haemorrhage due to concomitant intracerebral haemorrhage, the local pressure effects of the aneurysm itself, or cerebral vasospasm. A cerebral aneurysm usually lies within the subarachnoid cisterns but the aneurysm may become adherent to the adjacent brain parenchyma due to adhesions, frequently resulting from previous leakage of blood. A haemorrhage from an aneurysm in these circumstances may also extend into the brain and the position of the intracerebral haematoma will determine the type of neurological deficit. A middle cerebral artery aneurysm frequently ruptures into the temporal lobe, resulting in hemiparesis and aphasia if the dominant hemisphere is involved (Fig. 9.1). Anterior communicating artery aneurysms may haemorrhage into the frontal lobes with subsequent akinetic mutism (Fig. 9.2). Defective conjugate ocular movement may result from haemorrhage into a frontal lobe, persistent deviation usually being towards the side of the lesion and purposeful gaze defective away from that side.
Fig. 9.2 Frontal intracerebral haematoma with blood in the Sylvian fissure and ventricles from a ruptured anterior communicating artery aneurysm.
Occasionally, an aneurysm may also rupture into the subdural space, resulting in a subdural haematoma and brain compression causing lateralizing neurological signs. An arteriovenous malformation usually lies at least partially within the brain parenchyma, so that when it ruptures intracerebral bleeding is frequently associated with the subarachnoid haemorrhage. Focal neurological signs may result from the position of the aneurysm itself. An aneurysm arising from the internal carotid artery at the origin of the posterior communicating artery (known as a posterior communicating artery aneurysm) may cause pressure on the 3rd cranial nerve. Patients with an enlarging aneurysm in this position may present with features of a 3rd cranial nerve palsy (ptosis, pupil dilatation, extraocular
muscle palsy) prior to a subarachnoid haemorrhage. It is vital that the correct diagnosis of an enlarging cerebral aneurysm is made in this situation, so as to avoid the possible catastrophic effects of subarachnoid haemorrhage. The major differential diagnosis of the aetiology of an apparently isolated 3rd cranial nerve palsy is an ischaemic lesion such as those resulting from diabetes mellitus or atherosclerosis. Pupil size is a useful guide in differentiating between these causes. The pupil is usually dilated, with an expanding aneurysm which compresses the superior aspect of the nerve that contains the parasympathetic pupillary fibres arising from the nucleus of Edinger–Westphal in the midbrain. An expanding aneurysm usually results in more pain than the ischaemia associated with diabetes mellitus, although this is an unreliable guide. If there is any doubt about the cause of the 3rd nerve palsy then angiography must be performed expeditiously. In a patient with impaired conscious state, or in one with other abnormal neurological signs suggesting a massive haemorrhage, 3rd cranial nerve palsy may be secondary to temporal lobe herniation. A giant aneurysm (defined as larger than 2.5 cm in diameter) may cause compression of adjacent neural structures resulting in focal signs (Fig. 9.3). A large aneurysm of the internal carotid artery or anterior communicating artery will cause compression of the optic nerve or chiasm, respectively, resulting in visual failure. Large vertebrobasilar aneurysms may cause brainstem compression.
Cerebral vasospasm following subarachnoid haemorrhage does not usually result in clinical manifestations for 2 or 3 days after the initial bleed so that, although it may be the cause of subsequent focal signs resulting from brain ischaemia, it is not the cause of focal signs immediately after the haemorrhage.
Optic fundi
Mild papilloedema is common within the first few days of haemorrhage because of the sudden elevation of intracranial pressure resulting from hydrocephalus or cerebral oedema. A transient communicating hydrocephalus often occurs after subarachnoid haemorrhage due to blood blocking the arachnoid villi. In about 10% of cases the hydrocephalus persists and is severe enough to require a CSF shunt.
Ophthalmoscopy may reveal fundal haemorrhages, particularly in severe subarachnoid haemorrhage. Small, scattered retinal haemorrhages usually resolve satisfactorily, although the large subhyaloid haemorrhages may break into the vitreous, resulting in permanent visual defect.
Clinical assessment
The diagnosis is usually obvious when the history is obtained from the patient, relative or friend. The classic sudden onset of severe headache with features of meningism and decreased conscious state is characteristic of a subarachnoid haemorrhage. However, difficulty may occur when the haemorrhage has been minor and, tragically, a subarachnoid haemorrhage may be misdiagnosed as either migraine or tension headache. A full neurological examination should be performed with particular attention given to the presence of neck stiffness, altered conscious state, pupillary status and fundal haemorrhage. Clinical grading systems have been based on the severity of the headache and neck stiffness and on the level of conscious state. The two major systems are the Hunt and Hess classification and the World Federation of Neurological Surgeons (WFNS) system (Table 9.3).
Investigations
The major differential diagnosis is meningitis, although a minor haemorrhage is often misdiagnosed as migraine. Confirmation of the clinical diagnosis of subarachnoid haemorrhage should be undertaken as soon as possible. Computerized tomography (CT) scanning (Fig. 9.4) is the best initial investigation as it will confirm the diagnosis in over 85% of cases. It will also provide additional information on associated pathology such as intracerebral haemorrhage and hydrocephalus, and on the position of the haemorrhage, which is helpful if there is more than one aneurysm. Arteriovenous malformation causing subarachnoid haemorrhage can frequently be diagnosed on the CT scan. If there is any doubt that subarachnoid blood is present on the CT scan, as may occur following more minor haemorrhages, a lumbar puncture is essential. The presence of xanthochromia (yellow staining) in the CSF will confirm subarachnoid haemorrhage.
Fig. 9.4 Blood in the Sylvian fissure and basal cisterns indicative of subarachnoid haemorrhage.
Cerebral angiography will confirm the cause of the subarachnoid haemorrhage and will determine the subsequent treatment. Intra-arterial digital subtraction angiography has considerably reduced the risks of conventional angiography and should be undertaken as soon as the diagnosis has been confirmed and it is clear that the patient will survive the initial haemorrhage.
Cerebral aneurysm
Cerebral aneurysms are the most common cause of subarachnoid haemorrhage in the adult population, with a maximal incidence in the 4th and 5th decades of life, although they can occur at any age.
Surgical anatomy
The great majority of aneurysms arise at the branch points of two vessels, usually at an acute angle, and are situated mainly on the circle of Willis and the trunks of the large arteries which supply it. Afew arise from its immediate branches but aneurysms on peripheral vessels are rare (Fig. 9.5). The majority of aneurysms occur in constant positions on the circle of Willis and about 85% occur on the anterior half of the circle. Aneurysms arise at approximately equal frequency from the internal carotid artery, anterior communicating artery and middle cerebral artery. Those associated with the internal carotid artery most frequently arise at the origin of the posterior communicating artery (the so-called posterior communicating artery aneurysm), less frequently at the terminal bifurcation, and occasionally at the origin of the ophthalmic artery, the anterior choroidal artery or in the cavernous sinus. Middle cerebral artery aneurysms arise from the middle cerebral artery at its bifurcation or trifurcation in the Sylvian fissure (Fig. 9.6). Less commonly an aneurysm may arise from the pericallosal artery at the genu of the corpus callosum.
Fig. 9.5 Usual sites of cerebral aneurysms.
Fig. 9.6 (a) Anterior cerebral artery aneurysm. (b) Middle cerebral artery aneurysm. (c) Posterior communicating artery aneurysm. (d) Terminal internal carotid artery aneurysm.
Approximately 15% of aneurysms arise from the posterior half of the circle of Willis, the most common position being the basilar artery, most frequently at the terminal bifurcation into the posterior cerebral arteries. However, an aneurysm may arise from any of the main branches of the vertebral or basilar arteries, in particular the posterior inferior cerebellar artery, anterior inferior cerebellar artery or superior cerebellar artery (Fig. 9.7).
Fig. 9.7 (a) Terminal basilar artery aneurysm. (b) Aneurysm arising from junction of basilar artery and superior cerebellar artery. (c) Posterior inferior cerebellar artery aneurysm.
Multiple aneurysms occur in more than one position in approximately 15% of cases.
Pathogenesis of cerebral aneurysms
The common type of cerebral aneurysm resulting in a subarachnoid haemorrhage is a saccular aneurysm, which is also known as a berry or congenital aneurysm. Fusiform aneurysms occur in the intracranial circulation, particularly the vertebrobasilar arteries or internal carotid arteries, and are due to diffuse atheromatous degeneration of the arterial wall, frequently associated with hypertension. Mycotic aneurysms result from septic emboli. They may be situated on peripheral vessels, are frequently multiple and have a high risk of haemorrhage. The saccular or berry aneurysm arises at the junction of vessels where there is a congenital deficiency in the muscle coat. The elastic layer in cerebral arteries, in contrast with arteries elsewhere, is limited to the internal lamina, making these vessels more susceptible to weakening effects of degeneration. Fragmentation and dissolution of the internal elastic membrane occurs at the site of aneurysm development. The combination of the muscle defect and the discontinuity of the underlying internal elastic membrane is probably necessary for the formation of a saccular aneurysm. Other factors that increase the risk of aneurysm formation include atheroma and hypertension. There is an increased incidence of atheroma in the vessels of the circle of Willis and hypertension in patients with ruptured aneurysms. It is probable that these factors play a role in the growth of the aneurysm and its subsequent rupture in some patients.
Related diseases
There is no definite hereditary basis to the development of intracranial aneurysms, although an epidemiology study has shown an increased incidence of approximately seven-fold in first-degree relatives of patients who have had an aneurysmal subarachnoid haemorrhage, with a lifetime risk of 2–5% of developing an aneurysmal subarachnoid haemorrhage. Aneurysms do occur in association with hereditary syndromes such as Ehlers–Danlos syndrome, coarctation of the aorta and polycystic kidney disease.
Management of ruptured cerebral aneurysm
The management of patients following rupture of a cerebral aneurysm is determined by three major factors.
1 Severity of the initial haemorrhage.
2 Rebleeding of the aneurysm.
3 Cerebral vasospasm.
Severity of the initial haemorrhage
About 30% of all patients suffering a subarachnoid haemorrhage from a ruptured aneurysm either have an apoplectic death or are deeply comatose as a result of the initial haemorrhage.
Rebleeding
This occurs in about 50% of patients within 6 weeks and 25% of patients within 2 weeks of the initial haemorrhage. About half the patients that have a subsequent haemorrhage will die as a result of the rebleed. After the first year the risk of a further haemorrhage from the aneurysm is about 2–3% per year.
The only certain way to prevent an aneurysm rebleeding is to exclude it from the circulation. Antifibrinolytic agents (such as epsilon amino caproic acid or tranexamic acid) decrease the risks of rebleeding but, as they are associated with increased incidence of thrombosis (such as deep vein thrombosis and pulmonary embolus) and an increased risk of cerebral thrombosis associated with vasospasm, these agents are now rarely used.
Cerebral vasospasm
Angiographic vasospasm (Fig. 9.8) occurs in about 50% of patients following subarachnoid haemorrhage and in 25% it results in serious neurological complications. There is a direct correlation between the amount of blood noted in the basal cisterns on the CT scan, the risk of developing vasospasm and its severity. Although the spasm may principally affect the vessels most adjacent to the ruptured aneurysm, generalized vasospasm occurs frequently. The clinical manifestations resulting from vasospasm will be determined by the vessels which are most severely affected. Spasm of the internal carotid artery and middle cerebral arteries produces hemiparesis and aphasia in the dominant hemisphere. Vasospasm of the anterior cerebral vessels causes paralysis of the lower limbs and akinetic mutism. Severe vasospasm may cause widespread cerebral ischaemia so that the patient may become obtunded; if the vasospasm is sufficiently severe it will result in death. Vasospasm does not usually occur until 2 or 3 days after the initial haemorrhage and its onset is rarely delayed beyond 14 days.
Fig. 9.8 Spasm of the anterior and middle cerebral arteries following subarachnoid haemorrhage from an anterior communicating artery aneurysm. (initial angiogram, on the 3rd day and after treatment)
Until recently there has beeo satisfactory treatment for established cerebral vasospasm. If the aneurysm has been surgically occluded from the circulation then hypertensive therapy combined with hypervolaemia may overcome the hypoperfusion due to narrowing of the cerebral blood vessels and reverse the ischaemic effects. Calcium channel blocking agents such as nimodopine and nifedipine are frequently used in subarachnoid haemorrhage to prevent and treat vasospasm, although there is still doubt as to their effectiveness.
Following the angiogram that confirms a cerebral aneurysm, a decision is then made as to the definitive treatment of the aneurysm. This will involve either:
• surgery with clipping of the aneurysm or
• endovascular obliteration of the aneurysm.
Surgery for ruptured aneurysm
The timing of the operation is critical in obtaining optimal results following subarachnoid haemorrhage. Although better operative results may be achieved when the surgery is delayed, the longer the operation is deferred the greater the risk that the aneurysm will rebleed. In general, the operation is performed as soon as possible after the cerebral angiogram. In the past surgery was avoided when the patient had clinical or angiographically severe vasospasm, but it is now recognized that it is best to clip the aneurysm even in the presence of clinical or radiological vasospasm as with the aneurysm excluded from the circulation the spasm can be treated using hypertensive hypervolaemic therapy and endovascular techniques.
Surgery is usually not performed on patients who are comatosed or have features of decerebrate posturing response, unless the CT scan shows a large intracerebral haematoma resulting from the ruptured aneurysm which needs to be evacuated, or hydrocephalus as a cause of the poor neurological state.
Preoperative management
In those patients in whom it has been elected for some reason to delay surgery, the management should include careful attention to the following.
• Posture. The patient should lie flat in a quiet room with subdued lighting. Every attempt should be made to avoid environmental situations which could cause sudden elevation of the patient’s blood pressure and thus increase the risk of rupture of the aneurysm. Sedation usingbarbiturates or diazepam may be necessary if the patient is agitated.
• Blood pressure control. The blood pressure is frequently elevated immediately after the haemorrhage and should be carefully controlled. Initially this should be done using intravenous medication and utilizing vasodilating agents (such as hydralazine or glyceryl trinitrate) and betablockers. Although it is essential to control high blood pressure, as this may lead to rupture of the aneurysm, hypotension may result in cerebral ischaemia, particularly when vasospasm is present. The appropriate desirable blood pressure will depend upon the premorbid level.
• Fluids and electrolytes. Correct hydration is essential to avoid electrolyte disturbance; in addition, overhydration may precipitate cerebral oedema and insufficient fluids may increase the risk of cerebral thrombosis associated with vasospasm. Electrolyte disturbances may also occur following subarachnoid haemorrhage due to inappropriate antidiuretic hormone (ADH) secretion, which results in hyponatraemia.
• Pain relief. Simple analgesic medication or codeine phosphate is best used for controlling the headaches resulting from subarachnoid haemorrhage.
Surgical procedures
The surgical procedures available are:
• occlusion of the neck of the aneurysm
• reinforcement of the sac of the aneurysm
• proximal ligation of a feeding vessel.
Haemorrhage from an aneurysm is due to rupture of the fundus of the aneurysmal sac. Therefore, the best surgical procedure is to occlude the neck of the aneurysm, thereby isolating the aneurysm from the circulation.
In brief, the operation involves a craniotomy which is usually based on the pterion (pterional craniotomy) for aneurysms of the anterior circulation. This type of craniotomy may also be used for aneurysms arising from the terminal basilar artery, although some surgeons prefer an approach under the temporal lobe via a temporal craniotomy. Microsurgical techniques, utilizing the operating microscope and microneurosurgical instruments, are employed. Access to the basal cisterns may be aided by withdrawing CSF either using a ventricular drain or from the lumbar theca. The arachnoid around the basal cisterns is opened, the neck of the aneurysm identified and dissected and a clip placed across the neck to exclude the aneurysm from the circulation. During the dissection of the aneurysm it is essential that vital adjacent vessels, including the perforating arteries, are not injured, as damage to these vessels may result in severe neurological disability. Occasionally it is not possible to safely place a clip across the neck of the aneurysm, usually as a result of branches of the parent vessel either arising from the aneurysm or being inseparable from the fundus. In this case the wall of the aneurysm may be reinforced by a number of techniques, including wrapping the wall with crushed muscle, gauze or cotton wool or a combination of these. Rapidly solidifying polymer (aneurysm cement) may be poured around the aneurysm to provide it with a solid covering.
Postoperative management
The usual postcraniotomy operative management applies, with special attention to be given to the neurological state, hydration, posture, oxygenation and blood pressure. Anticonvulsant medication is recommended for 3 months to 1 year. Steroid medication is sometimes used in the initial postoperative course to control cerebral oedema, although its effectiveness is not proven. The major specific postoperative problem results from delayed cerebral vasospasm. As indicated previously, prophylactic calcium channel blocking agents may be of use in preventing this complication. The transcranial Doppler, utilizing non-invasive ultrasound, may give useful information on the degree of intracranial vasospasm. Symptomatic vasospasm can be treated using hypervolaemic hypertensive therapy. This treatment entails careful monitoring and requires the transfer of the patient to an intensive care unit. Recently, endovascular techniques to dilate the vessels in spasm or administer intra-arterial papaverine into the intracranial vessels have been used to treat cerebral vasospasm postoperatively with some success.
Endovascular procedures for ruptured aneurysms
Over the past 10 years endovascular techniques (using detachable coils) have been used to obliterate cerebral aneurysms. These have been investigated in international trials, and have proven to be effective in excluding the aneurysm from the circulation.
The technique is usually performed by a specialist interventional radiologist, and almost always under general anaesthesia. As with surgery, it is recommended that the aneurysm is ‘coiled’ as soon as possible after the angiogram has been performed to confirm the presence of an aneurysm (Fig. 9.9).
Fig. 9.9 (a) coils for endovascular treatment of internal carotid artery aneurysm. Figure (c) shows the aneurysm excluded from the circulation following the endovascular insertion of coils.
The recent ISAT trial showed the possible superiority of endovascular coiling over surgery, although there has been some debate as to these findings. The decision as to whether an aneurysm should be ‘clipped’ by a surgeon or ‘coiled’ by an interventional neuroradiologist is best made jointly by the treating specialist, cerebrovascular neurosurgeon and neuroradiologist. The interventional neuroradiologist will base his decision on
• the access to the aneurysm and
• the configuration of the aneurysm.
Access to the aneurysm may be impaired by stenosis or tortuosity within the carotid artery (for anterior circulation aneurysms) and vertebrobasilar artery (for posterior circulation aneurysms). The ‘dome to neck’ ratio is an important consideration in deciding whether the configuration of the aneurysm is appropriate for coiling. In general, most neuroradiologists prefer the ratio to be 2 : 1 or greater. New techniques in interventional radiology including the use of stents and three-dimensional coils have increased the number of aneurysms that can be treated by endovascular techniques. At present, over 80% of terminal basilar aneurysms can be treated by endovascular techniques, but only about half of the anterior circulation aneurysms are amenable to ‘coiling’. Most interventional radiologists do not coil middle cerebral artery aneurysms, as there is difficulty in obliterating the aneurysm whilst maintaining patency of the vessels.
Management of an unruptured aneurysm
Multiple aneurysms occur in 15% of patients who present following subarachnoid haemorrhage. In general, an unruptured aneurysm will be clipped at the same time as the surgery for the ruptured aneurysm, provided it can be performed through the same craniotomy. The indications for surgery are controversial for an unruptured aneurysm occurring in a patient who has suffered a subarachnoid haemorrhage from another aneurysm, or for an unruptured aneurysm found incidentally. The debate regarding the optimal management of patients with an unruptured aneurysm revolves around the relative risk of rupture of the aneurysm vs. the risk of treatment, by either surgery or an endovascular approach. In the past the risk of haemorrhage from an unruptured aneurysm was usually quoted at 2–3% per year.
However, in 1998 the New England Journal of Medicine published a large study of the natural history of unruptured intracranial aneurysms which indicated that the risk of haemorrhage was very much lower, particularly for aneurysms less than 10 mm in diameter and those arising from the middle cerebral artery. There has been considerable debate in the neurosurgical literature regarding the veracity of the so-called ISUIA (International Study of Unruptured Intracranial Aneurysms) study, with some experts questioning the methodology. In general, the risk of rupture will depend on the size of the aneurysm, on the configuration of the aneurysm, in particular if there is a ‘daughter sac’ attached to the fundus, on a positive family history for aneurysmal subarachnoid haemorrhage and on the age of the patient. Symptomatic aneurysms of all sizes should be considered for treatment.
Arteriovenous malformation
The arteriovenous malformation is the most common vascular malformation. Although it accounts for approximately 60% of all subarachnoid haemorrhage in children, by the 3rd decade it is responsible for 20% and by the 5th decade for less than 5%.
Clinical presentation
Haemorrhage. This is the most frequent first symptom of an arteriovenous malformation and, although the bleeding may be subarachnoid, there is commonly an intracerebral component. The arteriovenous fistulous communication results in the development of aneurysms within the lesion, enlargement of the arteries which feed the malformation and, consequently, the possible secondary development of saccular aneurysms on the major feeding vessels. The haemorrhage associated with an arteriovenous malformation may quite often be due to rupture of a saccular aneurysm on the feeding vessel.
Epilepsy. This is the second most common presenting manifestation of an arteriovenous malformation.
Headache. Migraine characteristics are particularly associated with headache due to arteriovenous malformation.
Progressive neurological deficit. For example, a slowly progressive hemiparesis may occur in a large malformation due either to local ischaemia induced by the shunt or to increasing size of the lesion.
Surgical anatomy
Most arteriovenous malformations are situated in the cerebral hemispheres, although they may occur in the posterior fossa involving either the cerebellum or brainstem and they show considerable variation in size. The malformations involving the cerebral hemispheres frequently form a pyramidal mass, the base of which may reach the cortical surface with the apex pointing towards the lateral ventricle. There are frequently multiple, enlarged arteries feeding the malformation and arterialized draining veins extend superficially to the superior sagittal sinus or transverse sinus or deeply into the deep cerebral venous system.
Radiological investigations for arteriovenous malformations (Figs 9.10–9.12)
An arteriovenous malformation is often apparent on the CT scan because of the vivid enhancement of the enlarged feeding vessel and arterialized draining veins after intravenous contrast. Cerebral angiography is best performed using digital subtraction angiographic techniques and is essential for adequate evaluation of the malformation. Precise determination of the position of the major feeding and draining vessels is vital prior to surgery. Magnetic resonance imaging is a valuable aid in determining the exact position of the arteriovenous malformation and the vessels. Preoperative occlusion of accessible major feeding vessels close to the malformation by an interventional radiologist may be useful if the procedure is technically feasible. Aflow-directed catheter is positioned in the artery, which is occluded using cyanoacrylate glue or a polymerizing collagen mixture.
Fig. 9.10 (a) The arteriovenous malformation enhances vividly on the CT scan after intravenous contrast and the major dilated feeding vessels can be seen. (b) The MRI shows the position of the malformation in coronal and axial planes and further information about the feeding vessels and draining veins (c).
Fig. 9.11 Cerebral angiography (digital subtraction angiogram) demonstrates the vascular anatomy of the arteriovenous malformation. (a) The major feeding vessels are shown on the arterial phase. (b) The draining veins are demonstrated on the venous phase.
Management
As with cerebral aneurysms the aim of treatment is to avoid either an initial haemorrhage or rebleed from the malformation. There has been controversy over the risk of haemorrhage and the morbidity and mortality associated with rupture of an arteriovenous malformation. Recent studies have shown that the chance of haemorrhage for both ruptured and unruptured arteriovenous malformations is about 3% each year and that the combined morbidity and mortality of each haemorrhage is at least 40%. However, unlike cerebral aneurysms, haemorrhage from an arteriovenous malformation rarely causes symptomatic vasospasm.
Surgical excision, provided it is technically feasible and would not result in a disabling neurological deficit, should be performed if the malformation has haemorrhaged. Unruptured arteriovenous malformations should be considered for excision if surgery is unlikely to produce significant neurological deficit.
Surgery for arteriovenous malformations
The principles of the operation involve isolation and occlusion of the principal feeding arteries followed by meticulous dissection of the malformation, with occlusion and division of the numerous small feeding vessels. The draining veins should be ligated only after all the feeding vessels have been occluded, since premature obstruction to the arterialized venous outflow will result in a precipitous swelling and rupture of the vascular mass.
The surgical management of giant arteriovenous malformations is fraught with considerable risk. The lesions may be surrounded by chronically ischaemic brain ‘steal’ by the malformation and abrupt occlusion of the shunt through the malformation has led in some cases to oedema and haemorrhage in the adjacent brain, a phenomenon first described by Spetzler and which was called the ‘normal perfusion pressure breakthrough’ theory. Methods that have been employed to avoid this complication include preoperative and intraoperative embolization and staged excision of the malformation.
The use of radiosurgical techniques, involving either the gamma knife (a highly focused cobalt source of irradiation) or stereotactic radiosurgery using a linear accelerator, has been advocated for the treatment of small (less than 3 cm diameter), unruptured and surgically inaccessible arteriovenous malformations with complete angiographic obliteration in greater than 80% of lesions with a diameter of 3 cm or less at 3 years after radiation. However, as the radiotherapy effect is slow, the patient remains at risk from haemorrhage for some years.
Vein of Galen malformation
This unusual malformation results when arteries feed directly into the vein of Galen and produces distinct clinical syndromes depending on the age at which the disease presents.
• Neonates present shortly after birth with cyanosis and heart failure due to the shunt through the malformation.
• Infants and young children present with seizures and hydrocephalus due to obstruction of the cerebral aqueduct.
• Adults may present with multiple subarachnoid haemorrhage.
STROKE
Stroke is the third most common cause of death in most Western countries and the commonest cause of chronic adult disability. In many Western countries, there has been an impressive fall in population-based stroke mortality over the past few decades, averaging 4–5% each year. This has been chiefly attributed to the more effective recognition and treatment of hypertension and other risk factors. Despite this progress, with the increasing life expectancy of the population, a marked increase in stroke prevalence has been predicted. Stroke is generally a disease of ageing, but young adults are also affected, with a somewhat different pathogenetic spectrum. The overall approach to stroke has undergone a dramatic change in recent years, with important recent advances in prevention strategies, the widespread introduction of acute stroke units, the application of new imaging technologies and the introduction of thrombolysis for selected patients with acute ischaemic stroke.
The term ‘stroke’ is used to describe a sudden neurological deficit of vascular aetiology lasting more than 24 hours. A transient ischaemic attack (TIA) indicates a transient neurological deficit of vascular origin lasting less than 24 hours, although many patients with TIAs lasting more than minutes have in fact suffered some neuronal damage and this time definition has been recently challenged. Stroke is classified as cerebral infarction or ischaemic stroke, signifying ischaemic brain damage, or cerebral haemorrhage, where the primary pathology involves vascular rupture and extravasation of blood into the surrounding tissues. The term ‘haemorrhagic infarction’ is used to describe an infarct into which there has been a secondary extravasation of blood. Although subarachnoid haemorrhage (SAH) may not be associated with a focal neurological deficit, it is usually categorized as a stroke subtype.
Stroke prevention
Stroke prevention involves primary and secondary strategies. Primary prevention includes lifestyle modification and treatment of risk factors in individuals who have not experienced cerebrovascular symptoms. Risk factor management can be categorized into the high risk approach (detecting and treating patients at high risk of stroke, such as those with atrial fibrillation or hypertension) and the mass or population approach (such as reducing salt intake and hence attempting to lower blood pressure in the entire population).
Secondary prevention involves the use of strategies in a symptomatic individual after a stroke or TIA, generally tailored to the specific type of cerebrovascular pathology. Most stroke prevention studies have focused on a reduction in the incidence of stroke, neurological disability and other vascular endpoints. Recently, prevention of vascular dementia has also been recognized as an important additional goal.
Primary prevention
Primary prevention strategies target the modifiable risk factors for stroke (Table 10.1).
Secondary prevention
Secondary prevention strategies after stroke or TIA, unlike the primary prevention techniques, are tailored to the underlying stroke pathology. The range of secondary prevention strategies continues to expand (Table 10.2). These include the introduction of a number of new antiplatelet strategies, proof of the efficacy of warfarin in non-valvular atrial fibrillation, clarification of the indications for carotid endarterectomy and the introduction of cerebral angioplasty/stenting.
Carotid endarterectomy
Two large trials (the North American Symptomatic Carotid Endarterectomy Trial—NASCET and the European Carotid Surgery Trial—ECST) showed major benefits for carotid endarterectomy over optimal medical therapy in patients with greater than 70% carotid stenosis and either TIA or non-disabling stroke. In the NASCET study, an absolute risk reduction of 17% over 18 months was achieved, indicating that one stroke could be prevented for every six patients.
carotid endarterectomy
Cerebrovascular angioplasty/stenting
Percutaneous transluminal angioplasty is increasingly used for symptomatic and asymptomatic carotid stenosis, usually combined with endovascular stenting. However, there is limited proof of efficacy and safety compared with endarterectomy. One trial showed that the benefits and risks of surgery and angioplasty/stenting were approximately equivalent. Large trials are now being conducted in patients with symptomatic carotid stenosis, comparing stenting with endarterectomy. Distal protection devices, which trap embolic debris at the time of the procedure, represent an important advance. Stenting has also been used in small series of patients with symptomatic intracranial stenoses, in whom optimal medical therapy has failed.
Acute stroke
The spectrum of transient ischaemic attacks and stroke
Transient ischaemic attacks (TIAs) and completed cerebral infarcts are caused by similar pathological mechanisms, most commonly large vessel atherosclerotic disease, cardioembolism and small vessel lacunar disease. Patients with TIAs and completed infarcts, whether large or small, have a similar prognosis, with a 5–10-fold annual increase in stroke risk. Both conditions should be regarded as medical emergencies. While TIAs by convention last less than 24 hours, most last only minutes. The majority of TIAs lasting more than 1–2 hours produce tissue damage on sensitive brain imaging techniques, such as magnetic resonance imaging. The old 24-hour definition is therefore increasingly criticized and is not clinically useful. The recognition of patients with minor ischaemic deficits presents a vital opportunity for prevention of major stroke. Patients with TIAs should be urgently evaluated within 24 hours of the episode. Investigations would typically include a CT or MR scan to detect infarction or non-vascular pathology, carotid duplex Doppler to diagnose major carotid disease, and ECG, sometimes followed by echocardiography, to diagnose atrial fibrillation or another cardioembolic source.
Carotid-territory TIAs
These are due to transient ischaemia in the retina or cerebral hemisphere. Transient monocular blindness (‘amaurosis fugax’) is due to a transient reduction in retinal perfusion produced by embolism or haemodynamic failure. The patient often describes a shade pulled down over one eye. In clinical practice it is vital to determine whether a visual disturbance is truly monocular, indicating retinal ischaemia, or binocular, often implicating the vertebrobasilar circulation. Hemispheric symptoms most commonly consist of transient dysphasia and varying degrees of hemiparesis or hemisensory disturbance, either singly or in combination.
Vertebrobasilar TIAs
These are often more complex than carotid territory events and usually include two or more of the following symptoms:
• binocular visual disturbance
• vertigo
• diplopia
• ataxia
• bilateral weakness or paraesthesiae
• deafness
• tinnitus
• amnesia.
These symptoms are produced by transient ischaemiaof the brainstem, occipital and medical temporal lobes and upper spinal cord.
Classification and pathogenesis of stroke
Cerebral infarction (ischaemic stroke)
Cerebral infarction accounts for approximately 80% of stroke patients and may be classified according to anatomical location or pathogenesis. It is useful to incorporate both classifications when considering stroke in a particular patient.
Anatomical classification
The anatomical location refers to the specific arterial territory (e.g. internal carotid vs. vertebrobasilar, anterior cerebral vs. middle cerebral) or specific location within the brain (e.g. lateral medullary syndrome, ventral pontine infarction or internal capsular infarction). Infarction most commonly occurs in the middle cerebral arterial territory and can be classified as cortical or deep (subcortical). The cortical middle cerebral syndromes depend on whether a small branch has been occluded, or whether one or both of the main two divisions of the middle cerebral artery is involved, the superior or inferior division. Subcortical infarcts occur in the territory of the deep perforating vessels supplying the internal capsule, thalamus, basal ganglia and brainstem. The occlusion of a single perforating vessel produces a small deep infarct, less than 1.5 cm in diameter, called a lacunar infarct, particularly if associated with one of the five classical clinical syndromes (see below). The obstruction of the origins of several of the deep perforating branches can produce a larger subcortical infarct termed a striatocapsular infarct.
Pathogenetic classification
Greater emphasis is now placed on the pathogenesis of cerebral infarction, as this is useful for the selection of secondary prevention therapies. This classification is often referred to as the TOAST system, after the TOAST trial.
1 Large artery atherosclerosis.
2 Cardiogenic embolism.
3 Lacunar infarction.
4 Rare causes (e.g. dissection, vasculitis, prothrombotic states).
5 Unclassified:
• despite adequate investigation
• due to inadequate investigation.
Large artery atherosclerosis (Figs 10.1–10.3)
The development of extracranial atherosclerotic plaque produces a progressive arterial stenosis. Subsequent plaque complications include ulceration, intraplaque haemorrhage and superimposed platelet–fibrin thrombus formation. Stroke is most often due to the development of thrombus (Fig. 10.4) followed by propagation and distal thromboembolism into the intracranial vessels, sometimes embolism composed of atheromatous debris, or haemodynamic failure due to the reduction of cerebral perfusion in the arterial border zones (borderzone or watershed infarction). Primary intracranial atherosclerosis and atherothrombosis is rare in Caucasian populations, but more common in African, African-American and Asian stroke patients (Fig. 10.5).
Clinical features include demonstration of relevant arterial pathology with 50% or greater stenosis and the absence of a cardiac source. Anterior circulation infarcts typically involve the cerebral cortex. Prodromal TIAs in the same arterial territory are another pointer
.
Fig. 10.1 Arterial digital subtraction angiogram demonstrating severe, proximal internal carotid artery stenosis .
Fig 10.2. carotid duplex ultrasound: high-grade stenosis (a) and occlusion (b) of internal carotid artery
Fig. 10.3 Diffusion-weighted imaging scan (left image) showing large acute middle cerebral artery (MCA) cortical infarct. Magnetic resonance angiography (right image) shows lack of flow in the left MCA.
Cardiogenic embolism
Avariety of cardiac diseases affecting the cardiac walls, valves or chambers can lead to cerebral embolism. These include:
• non-valvular atrial fibrillation (the most common)
• valvular heart disease
• myocardial infarction with ventricular thrombus formation
• post-cardiac surgery (valvular surgery or coronary artery bypass grafts)
• prosthetic cardiac valves
• infective endocarditis
• atrial myxoma
• cardiomyopathy
• septal defect with paradoxical embolism.
Clinical features include delineation of a cardiac source and lack of evidence of large artery disease. Similarly, the cerebral cortex is usually involved.
Lacunar infarction (Fig. 10.7)
The occlusion of single deep perforating arteries supplying the internal capsule, basal ganglia or brainstem can lead to the development of small lacunar infarcts. These are most commonly the result of hypertensive disease, which produces localized arterial wall pathology, termed lipohyalinosis, in these small penetrating arteries, or localized microatheroma. Lacunar infarcts are less often due to embolism from a proximal source such as extracranial atherosclerosis or intracardiac thrombus.
Fig. 10.7 Diffusion-weighted imaging scan showing acute left thalamic lacunar infarct (hyperintense lesion).
Classical lacunar syndromes include pure motor hemiparesis, pure sensory stroke, sensorimotor stroke, ataxic hemiparesis and the dysarthria/clumsy hand syndrome. Clinical pointers include clinical diagnosis of one of the classical syndromes, usually exclusion of a large artery or cardiac source and ideally neuroimaging confirmation of a small, deep infarct.
Oxfordshire classification
Another commonly used stroke classification, termed the Oxfordshire classification, divides ischaemic stroke into TACI (total anterior circulation infarction), PACI (partial anterior circulation infarction), POCI (posterior circulation infarction) and LACI (lacunar infarction). This classification is also useful and the subtypes correlate with prognosis.
Cerebral haemorrhage (haemorrhagic stroke)
Cerebral haemorrhage is generally classified into intracerebral and subarachnoid haemorrhage.
Intracerebral haemorrhage (Figs 10.9–10.10)
Non-traumatic (primary) intracerebral haemorrhage is most commonly due to hypertension, which leads to rupture of deep perforating arteries in the putamen, thalamus, central white matter, brainstem and cerebellum. The precise mechanism of this vascular rupture is uncertain, but may be related to the development of small Charcot–Bouchard microaneurysms on the vessel walls of these end-arteries, or direct rupture of vessels affected by lipohyalinosis.
Fig. 10.9 CT scan showing hypertensive putaminal haemorrhage.
Fig. 10.10 CT scan showing frontal and occipital lobar haemorrhages in a patient with amyloid angiopathy.
‘Lobar’ haemorrhage refers to superficial vascular rupture within the cerebral lobes, outside these deep arterial territories. It is sometimes due to an underlying structural lesion, such as an arteriovenous malformation, cerebral aneurysm, tumour, vasculopathies or coagulation disorders. Amyloid angiopathy is an important cause of lobar cerebral haemorrhage in elderly patients and is due to amyloid deposition in the walls of the cerebral arteries. These haemorrhages may be multiple and typically occur in patients who are normotensive and may show features of Alzheimer’s disease.
Clinical clues to ICH include the features of a sudden rise in intracranial pressure with depressed conscious state, headache and vomiting. The mortality is much higher than in ischaemic stroke. However, patients with ICH may be surprisingly alert and well looking. Conversely, patients with ischaemic stroke may have early depression of conscious state. Hence, neuroimaging, usually with CT, is mandatory in all cases to rapidly diagnose ICH.
Subarachnoid haemorrhage is classified according to pathological cause and site. The two most important identifiable causes include rupture of a berry aneurysm and arteriovenous malformation, but in up to 15% of cases no bleeding can be identified at angiography. Some of these idiopathic bleeds are due to perimesencephalic haemorrhage.
Modern principles of acute stroke management
Key principles of acute stroke management include:
1 Urgent recognition of stroke symptoms by the patient or their carer.
2 Urgent ambulance transport to a hospital with adequate diagnostic facilities and organized stroke care.
3 Urgent triage and investigation in the emergency department including CT brain scanning.
4 Assessment for acute stroke therapy, particularly thrombolysis.
5 Admission to a specialized stroke unit.
Clinical assessment of stroke
The following questions should be considered in the management of any patient with a presumed stroke.
• Is it a stroke?
• Is it an infarct or haemorrhage?
• Is the patient eligible for thrombolytic therapy or other urgent intervention?
• What is the arterial or anatomical localization and pathogenesis?
Stroke and pseudostroke
Non-vascular pathologies (‘pseudostroke’) such as cerebral tumour, subdural haematoma, abscess, migraine, metabolic disturbances and epilepsy can mimic the stroke process. All patients with suspected stroke require an urgent CT or MR scan to exclude non-cerebrovascular disorders, as well as to differentiate between infarct and haemorrhage (Figs 10.8 and 10.9). Lumbar puncture is reserved for those cases where meningitis is considered (usually after a CT scan) or where the diagnosis of subarachnoid haemorrhage is still contemplated after a normal CT scan.
The distinction between infarct and haemorrhage
The distinction between cerebral haemorrhage and infarction is vital, as some haemorrhages are considered for surgical evacuation, while patients with ischaemic stroke may be considered for thrombolysis or anticoagulation. While there are clinical pointers (see above) this differentiation is usually based on CT scan findings. Although CT scanning remains the workhorse for acute stroke assessment, recent studies indicate that MRI is at least as sensitive for intracerebral haemorrhage as CT and far more sensitive for acute ischaemia. Patients with ischaemic stroke often have early infarct changes on CT, although these may be subtle. MRI with diffusionweighted imaging (DWI) is increasingly used as this technique allows a sensitive diagnosis of cerebral ischaemia (Figs 10.3 and 10.7).
Is the patient eligible for thrombolysis?
The thrombolytic agent tissue plasminogen activator (tPA) has now been licensed in most parts of the world as the first proven stroke drug therapy, given intravenously within 3 hours of stroke onset in selected patients with ischaemic stroke. The approval of this acute therapy followed the positive results of a two-part pivotal trial, conducted in the USA. Other European trials testing tPAup to 6 hours have shown a marked trend to benefit over risk, and meta-analysis confirms the 3-hour window for tPA. Use of tPA increases the risk of symptomatic haemorrhagic complications by three- to fourfold. Major early infarct changes on CT, for example greater than one-third of the area of the middle cerebral artery, are associated with a higher risk of haemorrhagic transformation. Up to 10% of ischaemic stroke patients can be treated in well-organized centres. In contrast to the tPAtrials, three trials evaluating the role of intravenous streptokinase produced negative results, chiefly linked to the substantial risk of intracerebral haemorrhage associated with the drug.
Direct infusion of thrombolytics via intra-arterial catheters has been shown to be effective in one trial up to 6 hours after middle cerebral artery occlusion. There is interest in experimental mechanical devices which can break up thrombi, with a much lower risk of cerebral haemorrhage.
Location and pathogenesis of infarction
Cortical infarcts (Fig. 10.3). Based on the clinical examination, a distinction should be made between an infarction in the carotid or in the vertebrobasilar territory. With regard to carotid territory infarction, the presence or absence of cortical signs—dysphasia, apraxia, anosognosia (unawareness or denial of the stroke), sensory, motor or visual agnosia (inattention), acalculia, right/left confusion, dysgraphia or cortical sensory loss (loss of two-point discrimination, astereognosis, dysgraphaesthesia)—suggests an embolic source from either the extracranial vessels or the heart, rather than lacunar infarction.
Subcortical infarcts (Fig. 10.7). As indicated earlier, lacunar infarcts (less than 1.5 cm in diameter) rely on diagnosis of one of the classical lacunar syndromes. Cortical signs are not present. Lacunar syndromes reflect small vessel occlusions in the internal capsule, thalamus and brainstem.
Haemorrhage
Intracerebral haemorrhage (Fig. 10.9). As previously discussed, the rapid onset of a stroke with early depression of conscious state favours the diagnosis of a primary intracerebral haemorrhage. Primary intracerebral haemorrhage and aneurysmal subarachnoid haemorrhage can overlap in their clinical presentations. For example, a berry aneurysm can rupture primarily into the brain parenchyma, presenting as an intracerebral haemorrhage, whereas a primary brain haemorrhage can rupture directly into the ventricular system and present with marked meningeal features due to subarachnoid blood.
Urgent assessment in the emergency department
Rapid triage of suspected stroke patients should be performed in an emergency department, with skilled medical evaluation and urgent investigations. These involve blood tests, including a blood glucose measurement to exclude hypoglycaemia (which can mimic acute stroke), and an ECG to diagnose atrial fibrillation or acute myocardial infarction, which can reveal the cause of cerebral embolism. Most importantly, a CT brain scan must be performed urgently in all patients to differentiate between cerebral infarction and cerebral haemorrhage. In addition, CT scanning is essential to exclude the other brain pathologies that can simulate stroke. As noted above, many centres are now using acute MRI with DWI and magnetic resonance angiography (MRA). Computed tomographic angiography (CTA) is another useful tool. Duplex Doppler scanning is of value in the diagnosis of large vessel atherosclerosis, while echocardiography (usually less urgent) can aid in the diagnosis of intracardiac thrombi and valvular or cardiac wall kinetic abnormalities. Transoesophageal echocardiography is superior to transthoracic echocardiography.
Acute stroke units
One of the most important developments in acute stroke management in recent years has been the proven value of stroke units, based on the results of clinical trials which have compared organized expert care in a special unit to the management of patients in general medical wards. The usual stroke unit model involves a geographical area in the hospital, incorporating a skilled multidisciplinary stroke team, led by a neurologist or other physician with expertise in stroke management. Vascular surgeons and neurosurgeons should be readily available for consultation and intervention in selected cases. In a systematic overview of the benefits of stroke units, it was shown that mortality could be reduced by about 25%, also with evidence of reduced disability. In addition, stroke units have been found to reduce bed stay and hence hospital costs.
The components of team management in the stroke unit include:
1 Acute medical, surgical and nursing care.
2 Diagnosis of the pathological mechanism of the stroke and the institution of tailored secondary prevention strategies.
3 Prevention, early detection and treatment of complications such as aspiration pneumonia, pressure sores, hyperglycaemia, seizures and sepsis.
4 Use of measurement instruments and stroke registers.
5 Early integrated neurorehabilitation with evaluation of premorbid status and therapy goals.
6 Involvement, education, and support for patient and family.
7 Early, coordinated discharge planning.
Acute medical care
Three key, evidence-based advances include the use of tPA, aspirin (see below) and stroke unit care. In the acute phase, close monitoring of vital and neurological signs is paramount, given that about a third of stroke patients deteriorate after admission to hospital. Cardiac monitoring is useful for many acute stroke patients. Initial clinical evaluation should involve assessment of the patient’s function and clinical state before the stroke, the stroke type and pathogenesis, documentation of the nature and severity of neurological deficits, and comorbid diseases. Early mobilization, range of motion exercises for hemiplegic limbs, frequent turning, fluid and nutritional maintenance, dysphasia management, avoidance of aspiration, prevention of deep venous thrombosis (DVT) and pneumonia, management of incontinence, treatment of urinary tract infections and other causes of fever, and maintenance of skin integrity are all key planks in the team treatment of the acute stroke patient.
Patients should have oximetry, but there is no evidence of the value of routine supplemental oxygen by mask or nasal prongs. Patients should be well hydrated using normal saline rather than glucose-containing solutions, because of the hazards of even mild degrees of hyperglycaemia. In general, blood pressure should not be acutely lowered, because this can compromise cerebral perfusion in acute stroke. Hypertension is usual in acute stroke and this usually settles spontaneously over 2–3 days. We routinely use compression stockings and usually low-dose heparin, or low molecular weight heparin or heparinoid, in DVT prophylaxis. Airway assessment should be an urgent priority to avoid aspiration pneumonia. We generally use nasogastric feeding after 24–48 hours if the patient’s bulbar function is compromised.
Adverse prognostic factors include advancing age, depression of conscious state, severity of neurological deficit, conjugate gaze palsy and early urinary incontinence. Causes of mortality in stroke patients are predominantly neurological (transtentorial herniation) in the 1st week, and due to secondary systemic factors in the 2nd and 3rd weeks such as pneumonia, pulmonary embolism and cardiac causes.
Progressing stroke
A deteriorating neurological deficit is seen in about one-third of stroke patients. Cerebral oedema, with progressive elevation of intracranial pressure, is an important cause of death. However, this oedema is generally cytotoxic and clinical trials have demonstrated that corticosteroids are of no value in either cerebral infarction or haemorrhage. Intravenous mannitol and glycerol are sometimes empirically used, but their value has not been proven. In highly selected young patients with severe, early brain oedema, surgical decompression with hemicraniectomy should be considered.
Heparin
Heparin has been the most widely used unproven therapy in most countries. The International Stroke Trial (IST) evaluated unmonitored, subcutaneous heparin up to 48 hours after stroke onset and showed that a slight reduction in recurrent stroke was offset by an increased risk of haemorrhagic transformation. The rate of early, recurrent embolism in patients with atrial fibrillation was much lower in recent trials than in a number of earlier clinical studies.
Systematic overview of the available evidence leads to the conclusion that heparin is not generally indicated for most patients with acute ischaemic stroke, but heparin should be considered for selected patients with nondisabling ischaemic stroke and a very high risk of recurrent embolism.
Acute aspirin
Two large trials (the International Stroke Trial and the Chinese Acute Stroke Trial) both showed that, as for acute myocardial infarction, aspirin administered within 48 hours of stroke onset produced a modest improvement in outcomes at 6 months. Hence, aspirin should be used routinely in acute ischaemic stroke, unless thrombolysis is being used.
Neuroprotective therapies
A complex cascade of biochemical changes occurs in stroke, secondary to the initial ischaemia. Ischaemic brain injury is associated with elevated levels of the excitatory neurotransmitters glutamate and aspartate. These lead to excessive stimulation of the N-methyl D-aspartate (NMDA) receptor on the cell surface. This activation is followed initially by an influx of sodium and water into the cells and secondly by a sudden rise in intracellular calcium. Arange of neuroprotective compounds has been designed to inhibit various points in the excitotoxic cascade. These include calcium channel and NMDAantagonists, glutamate release inhibitors, glycine antagonists, free radical scavengers, inhibitors of the neutrophil influx into the ischaemic region and various growth factors. To date, none of these compounds have proven effective in adequately powered, Phase III clinical trials, but there are several ongoing studies. Other approaches include the combination of thrombolysis with neuroprotective drugs.
Cerebral hemorrhage
While routine surgical evacuation of haematoma is unproven, we consider evacuation in selected patients with cerebral haemorrhage, particularly in the cerebellum, as well as younger patients with lobar haemorrhage. The general principles of acute stroke management apply equally to intracerebral haemorrhage as to infarction.
Prevention of recurrent stroke—secondary stroke prevention
Patients with symptomatic high-grade carotid stenosis should be considered for subacute carotid endarterectomy. While surgery is awaited, or if endarterectomy is inappropriate, antiplatelet therapy should be instituted. In cardioembolic stroke, there is uncertainty as to the optimal timing of anticoagulation, particularly in patients with atrial fibrillation. Our approach is to employ heparin acutely if the patient has a mild deficit and there is a high risk of recurrent embolism. Many patients with atrial fibrillation are commenced on warfarin 7–10 days after onset, without prior heparin, because of the risk of haemorrhagic transformation. Lacunar infarcts are less commonly embolic and generally have a good prognosis.
