Intracranial aneurysms and arteriovenous malformations. (includes subarachnoid haemorrhage) - technical article.
Intracranial or 'berry' aneurysms are usually saccular but may occasionally be fusiform; they arise as an outpouching from the vessel wall, usually at a branching point. The saccular form comprises a neck and fundus. The neck is usually narrower than the fundus, which can be simple or lobulated and measure from a few millimetres up to several centimetres in diameter. Lobulated aneurysms can be complex and involve nearby branching vessels. Aneurysms are more commonly found on the major vessels of the anterior cerebral circulation (arteries of the circle of Willis), with the region of the anterior communicating artery, the middle cerebral bifurcation, and the internal carotid at the origin of the posterior communicating artery being the most common sites. The posterior circulation (vertebral or basilar arteries) typically accounts for 10 to 15 per cent of cerebral aneurysms that rupture.
Intracranial arteriovenous malformations
Arteriovenous malformations in the central nervous system can be considered as vascular hamartomas; they are not neoplastic. The vascular mesenchyme is split by the development of the brain and so arteriovenous malformations can affect all layers of the brain, skull, and scalp. Several groups of these malformations are recognized, including the more common intracererbral or brain arteriovenous malformation. Other types include cavernous angiomas (cavernoma), venous malformations or capillary telangiectasis; these seldom bleed.
Brain arteriovenous malformations
Brain arteriovenous malformations consist of abnormal communications between arteries and veins and sometimes direct arteriovenous fistulas; the abnormal communications are 50 to 250 mm, with a nidus fed by one or more enlarged feeding arteries and draining veins. The nidus has a sponge-like appearance. They vary very widely in size from tiny lesions of less than 1 cm to large lesions of 6 cm or more. Berry aneurysms can form on feeding arteries to brain arteriovenous malformations; they have an increased risk of rupture and greater complexity of treatment. Brain arteriovenous malformations may present with either haemorrhage, epilepsy or neurological deficit, or they may be found incidentally. Haemorrhage can be into the subarachnoid layer, intraparenchymal, or intraventricular. Ischaemia can be due to a 'steal' phenomenon whereby the malformation acts as an arteriovenous fistula and blood is diverted from nearby cerebral tissue, causing ischaemia. Compression and gliosis around these malformations can cause epilepsy.
Subarachnoid haemorrhage occurs when an intracranial aneurysm ruptures, with haemorrhage into the subarachnoid layer around the brain. However, aneurysms can present with pure subdural, intraparenchymal (intracerebral), or intraventricular haemorrhage. Subarachnoid haemorrhage may be caused by other intracranial lesions such as in head injury, haemorrhage stroke, and brain arteriovenous malformations.
Cerebral aneurysm and subarachoid haemorrhage
Intracranial aneurysms can be found at any age, but more commonly in adults, with 1 to 3 per cent of adults found to be affected at autopsy. The yearly incidence of rupture is low. Recent evidence suggests that the risk of bleeding from a previously unruptured aneurysm (less than 10 mm in diameter) is less than 0.1 per cent per annum. However, rupture of an intracranial aneurysm (subarachnoid haemorrhage) is a devastating event with a 30 per cent immediate mortality. The 70 per cent who survive the initial ictus may suffer neurological damage ranging from minimal to severe. The risk for further rupture of a previously ruptured aneurysm is substantially higher in the first 2 weeks (approximately 20 per cent), with a 60 per cent mortality, than after 6 months (3 per cent per annum). The frequency of aneurysmal subarachnoid haemorrhage is estimated to be between 6 to 12 per 100 000 per annum in most Western populations.
The cause of intracranial aneurysm remains unknown. Extensive genetic studies have failed to find a common molecular biological basis. Some inherited conditions increase the risk of intracranial aneurysm formation, most notably adult polycystic kidney disease. However, the vast majority as yet do not have a known genetic basis, apart from epidemiological evidence that consanguineous first-degree relatives have a fourfold increase in suffering a subarachnoid haemorrhage. We do not as yet understand enough about the natural history to justify the risks of investigation and treatment following screening; at present, screening for intracranial aneurysm is not justified.
Intracranial aneurysms may cause neuronal damage by compression or rupture. Compression may cause pain, dysfunction, or epilepsy. Pain is usually due to meningeal irritation, but vascular stretching may also be a component. Dysfunction can result from the compression of cranial nerves, nearby vasculature, or brain. Brain compression may produce focal neurological changes or, more rarely, epilepsy. Rupture usually results in subarachnoid haemorrhage, but may bleed directly into the brain (intracerebral) or occasional into other intracranial layers such as subdural or intraventricular.
Subarachnoid haemorrhage leads to a chemically induced meningitis, vasospasm of the cerebral arteries, and acute obstructive or subacute communicating hydrocephalus. The cause of the cerebral vasospasm is unknown; it is the most common cause of further neurological deterioration in the presence of adequate treatment of the aneurysm and hydrocephalus. Cerebral vasospasm occurs in 60 per cent of survivors and causes delayed neurological deterioration in 25 to 35 per cent, of whom some will slowly recover.
Typically, patients present with a sudden (within a fraction of a second) and unusual (never experienced before) headache, which is frequently described as 'like being hit on the back of the head' and many sufferers accuse people nearby of assault. Migraine sufferers struck down by subarachnoid haemorrhage can usually distinguish between their familiar migrainous pain and the ictal headache of the haemorrhage.
The level of consciousness may be reduced and there may be focal neurological signs associated with neck stiffness from the chemical meningitis. Vomiting usually accompanies the headache. Thirty per cent of patients will die at the ictus. The clinical status of the patients should be record according to the World Federation of Neurosurgeons (WFNS) scale, in which the Glasgow coma scale (GCS) is used to measure consciousness: grade 1 = headache + GCS 15; grade II = GCS 13–14, no major neurological deficit; grade III = GCS 13–14 with focal deficit; grade IV = GCS 7–12; grade V = GCS 6–3.
Rebleeding (approximately 4 per cent in the first 24 hours and 16 per cent in next 2 weeks) is the most dangerous sequel of aneurysmal subarachnoid haemorrhage. It is associated with a mortality of the order of 60 per cent and significant higher risk of poor outcome. Other causes of deterioration include hydrocephalus (obstructive or communicating) and delayed ischaemic neurological deficits or vasospasm may cause secondary deterioration.
Clinical management is aimed at maintaining the patient's airway, blood pressure, and pain control, since poor ventilation, myocardial complications, and severe head pain all cause further deterioration. The next step is to establish the diagnosis by appropriately timed investigations. Any patient with sudden, unusual headache must be investigated to exclude subarachnoid haemorrhage as an emergency.
The timing of appropriate tests is important in preventing overinvestigation of headaches not due to subarachnoid haemorrhage. Computed tomography (CT) is the technique of choice in establishing the diagnosis. A CT scan within the first 24 h of the ictus carries a 95 per cent sensitivity, falling to 50 per cent at 3 days; urgent access to CT scanning is therefore essential. If the CT is normal or equivocal, lumbar puncture is mandatory to look for xanthochromia (preferably after 3 hours of ictus to allow bilirubin to accumulate). The sample should be spun (3000 rev/min for 10 min) soon after collection to separate out the red blood cells, which otherwise may continue to lyse, supply further oxyhaemoglobin, and allow bilirubin to form in vitro. The presentation of acute bacterial meningitis can simulate that of subarachnoid haemorrhage, so microbiological examination of the cerebrospinal fluid should also be requested.
Cerebral angiography remains the investigation of choice for establishing the cause of the subarachnoid haemorrhage. Magnetic resonance or CT angiography do not yet have equivalent resolution, although recent developments with contrast agents may improve the spatial resolution. The most common cause of subarachnoid haemorrhage is a 'berry' aneurysm in patients over 40 years of age, with arteriovenous malformation more common in patients under 40 years.
Negative cerebral angiography presents a challenge to further management. False-negative angiography (CT scan positive) has been described systematically. Some patterns of haemorrhage in the region of the pons, known as perimesencephalic haemorrhage, are consistently associated with negative angiography. However, some negative angiograms are due to temporary thrombosis of a 'berry' aneurysm or vasospasm of feeding vessels, with subsequent risk of repeated haemorrhage and death. Angiography should be repeated where clinical doubt remains, but must be balanced against the risks of the investigation (less than 0.1 per cent risk of neurological complications when done by an experienced neuroradiologist).
Treatment of patients with aneurysmal subarachnoid haemorrhage
Medical management includes bed rest, adequate intravenous fluids (3 litres normal saline/24 h), pain relief, and monitoring of blood pressure, tissue oxygen saturation and neurological status.
Prevention of rebleeding
The greatest risk to a patient with subarachnoid haemorrhage from an established berry aneurysm is rebleeding, which carries about a 60 per cent chance of fatality. Intervention is aimed at stopping the rebleeding; this can be achieved by the surgical application of a clip around the neck of the aneurysm, and has been the standard treatment for many years. Recently, endovascular techniques using detachable platinum coils have been introduced to occlude the aneurysm and prevent rebleeding, thus avoiding the risks associated with craniotomy.
Surgical treatment of aneurysms
Having decided that open intervention via a craniotomy is the appropriate treatment, the next controversial decision revolves around timing. Early surgery (within 3 days) reduces the incidence of early rebleeding, avoids operating with vasospastic arteries, but is technically more demanding as the surgeon is operating around swollen, friable brain. Delayed surgery (after 10 days) allows the brain and arteries to settle after the initial ictus, but the patient must accept the additional 10 per cent mortality from rebleeding. Various neurosurgeons have investigated outcomes from either approach but there has been no rigorous, randomized prospective trial to clarify any difference between early and late surgical intervention.
A multitude of different approaches, both traditional and 'micro', are described for aneurysms of both the anterior and posterior circulations. Standard approaches to aneurysms of the anterior circulation include a frontal incision with the raising of a pterional free or vascularized bone flap and exposure of the sylvian fissure.
The main principles governing the intradural component of the operation are to secure proximal control of the feeding artery(s), remove cerebrospinal fluid either directly or via a lumbar/ventricular drain, and dissect microscopically the arachnoid layers to expose the aneurysm neck.
Clipping the neck
Applying the clip (magnetic resonance imaging-compatible where possible) to the neck of the aneurysm requires knowledge of the feeding artery and its subsequent branches. Incomplete exclusion of the aneurysm for the circulation may lead to rebleeding or reformation. Incorporating branches of the parent vessel may be lead to ischaemic injury and death from brain swelling. It can be extremely difficult to exclude some aneurysms from the circulation, and a compromise between cure and neurovascular injury must be accepted. Some giant aneurysms have thick, calcified walls and external clipping may not be possible; resection of these lesions requires significant periods of operative ischaemia. Brain cooling and other methods of cerebral protection have been tried with varying success.
Incomplete exclusion of the aneurysm can be supported by various preparations such as skeletal muscle (temporalis), acrylic cement, and cottonoid wool. These additional methods have not be shown to alter significantly the incidence of rebleeding and can produce comorbidity from granulation/fibrotic reactions affecting nearby cranial nerves.
This requires rigorous attention to fluid balance, blood pressure, and neurological status. Neurological deterioration should be investigated, with suitable blood tests to exclude hyponatraemia, estimation of ventilatory status (blood gases), and CT scanning to exclude postoperative haematoma. Transcranial Doppler may be helpful following the development of vasospasm, but can be difficult to interpret in the postoperative period. Postoperative cerebral angiography can help in demonstrating occlusion of the aneurysm and the presence of vasospasm.
Patients can be mobilized when the risk of vasospasm (usually after 5 days) has subsided. Nimodipine, 60 mg 4-hourly, where started preoperatively, should be continued for 21 days. Neurological outcome can be assessed between 6 months and 1 year postoperatively.
Endovascular treatment of intracranial aneurysm
Description of technique
The endovascular treatment of intracranial aneurysms is achieved by the use of detachable platinum coils. These are introduced via a series of catheters from the femoral artery, with placement of a guiding catheter in the internal carotid or vertebral artery, under digital fluoroscopy using a modern digital angiographic system. A microcatheter (1.8–2 F or 1.5 mm) and fine guidewire (0.010–0.014) are introduced into the cerebral circulation and navigated under fluoroscopic guidance into the aneurysm. A series of platinum coils is then placed, commencing with one approximately to the size of the aneurysm. These coils come in a helical configuration with a diameter ranging from 2 to 20 mm. When it has been satisfactorily positioned, the coil it is then detached from the introducing wire by electrolysis of the junction (Guglielmi detachable coil). When a 'basket' has been formed in the centre it is packed with small coils until dense occlusion of the aneurysm is achieved, producing physical obstruction of blood flow into it.
Angiography is used during and at the end of the procedure to confirm the occlusion. Further follow-up angiography is performed, usually about 6 months, later to confirm that the occlusion is stable.
Selection of patients for treatment
Aneurysms in the posterior fossa (basilar derived) are best treated with intravascular methods using detachable platinum coils if possible because of the relatively higher surgical risks in this location. Intervention via craniotomy or the intravascular route for intracranial aneurysms has been studied in the the International Subarachnoid Aneurysm Trial (ISAT).
The International Subarachnoid Aneurysm Trial (ISAT) was a large multicentre, prospective randomised clinical medical trial, comparing the safety and efficacy of endovascular coil treatment and surgical clipping for the treatment of brain aneurysms. The study began in 1994. The first results were published in The Lancet in 2002, and the 10 year data were published again in The Lancet in early September 2005. 2,143 study participants were mostly drawn from U.K. hospitals with the rest drawn from North American and European hospitals. The study found better results with endovascular coil treatment compared to surgical clipping, however subsequent studies have questioned this conclusion. The study was criticised by many clinicians and not well accepted by surgeons. Primary criticisms were related to the study's patient population's generalisability to the wider population, and the long term prognosis of coil embolisation.
Communicating hydrocephalus can be treated by repeated lumbar puncture and usually resolves over 10 days from the ictus. Obstructive hydrocephalus is best treated by a ventricular drain or ventriculoperitoneal (or atrial) shunt.
Haematoma associated with midline shift and reduce consciousness, particularly when outside the parenchyma, can be considered for craniotomy and evacuation; this can either be accompanied by clipping of the aneurysm or preceded by intravascular occlusion.
Management of vasospasm
Cerebral vasospasm following subarachnoid haemorrhage is an important cause of additional and potentially avoidable neurological deterioration. Increasing the amount of blood in the subarachnoid space increases the risk of vasospasm. Overall 25 to 35 per cent of patients with subarachnoid haemorrhage will suffer delayed ischaemic neurological deficits due to vasospasm. With current medical management the majority of these patients will suffer permanent neurological deficits as a result.
Medical management of cerebral vasospasm in subarachnoid haemorrhage relies on maintaining adequate blood pressure with intravenous fluids (mixture of crystalloid and colloids, 3.5 l/day), since cerebral perfusion is related to mean arterial pressure and intracranial pressure. The haematocrit can be allowed to fall and inotropes used to push the mean arterial pressure to supraphysiological levels—triple H therapy. A number of other complex metabolic responses also occur during this period (usually day 3 to day 10 post ictus), such as natriuresis, hyponatraemia, and a variety of cardiac arrhythmias.
The prognosis after subarachnoid haemorrhage is largely determined by the initial clinical grade of the patient (WFNS), the amount of intracranial blood (Fisher grade), age, and comorbidity such as hypertension. Patients below the age of 70 years with an initial clinical grade II or better should have a 60 to 80 per cent chance of good or excellent outcome, such as return to work. At the other end of the clinical grades, 70 per cent mortality can be expected in grade V patients, with a few per cent achieving excellent outcomes.
The additional factors that cause deteriorating outcome include rebleeding, cerebral vasospasm, and hydrocephalus; these are potentially avoidable and are best managed in specialist units by those with an interest in neurovascular conditions.
Despite considerable improvements in early diagnosis, prompt intervention to stop rebleeding, and intensive medical management the outcome from subarachnoid haemorrhage, when input grade is compared to outcome, is still worse than for other types of cerebrovascular accident.
Treatment of brain arteriovenous malformations
The risks of rebleeding from brain arteriovenous malformations are much lower than from an aneurysm; thus there is less urgency about treatment. The rebleeding rate is about 2 to 4 per cent per annum. Treatment is aimed at preventing rebleeding or, occasionally, improving focal neurological deficits. The surgical resection of extensive arteriovenous malformations in eloquent areas of the brain is impossible, but these can be treated by a combination of intravascular embolization and stereotactic radiosurgery using a precisely focused radiotherapy or g-ray beam. Some arteriovenous malformations are untreatable by any technique.
The management of patients with brain arteriovenous malformations is complex and depends on a multitude of factors: age, presentation, location and size of the malformations, and particularly whether it is in an eloquent area of the brain. If a decision is made to treat, then the options are as follows:
- Surgery, usually for accessible and small arteriovenous malformations.
- Embolization using microcatheters in feeding arteries and tissue adhesive to obliterate the residue of the malformation. Only about 15 per cent of arteriovenous malformations can be completely obliterated by this method but their size can often be reduced significantly, which makes the lesion more accessible to stereotactic radiation or surgical removal.
- Stereotactic radiosurgery using a gamma knife unit; this consists of about 200 separate cobalt sources and provides highly controlled, focused radiation treatment of brain lesions. In small lesions (less than 1 cm), radiosurgery is over 90 per cent successful but takes 2 years for full effect; for larger lesions the success rate falls. This type of radiation can also be performed by stereotactically adapted linear accelerators, although data on whether the technique is as effective are lacking.
- A combination of the above treatments is frequently used.
Bromberg JE, Rinkel GJ, Algra A, Limburg M, van Gijn J. Outcome in familial subarachnoid hemorrhage. Stroke 1995; 26(6): 961–3.
Caplan LR. Should intracranial aneurysms be treated before they rupture. New England Journal of Medicine 1998; 339: 1774–5.
Fisher CM, Kistler JP, Davis JM. Relation of cerebral vasospasm to subarachnoid haemorrhage visualised by computerised tomographic scanning. Neurosurgery 1980; 6: 1–9.
International Study of Unruptured Intracranial Aneurysms. Unruptured intracranial aneurysms—risk of rupture and risks of surgical intervention. New England Journal of Medicine 1998; 339: 1725–33.
Shaw MDM. Chs 11 and 12. In: Miller JD, ed. Northfield's surgery of the central nervous system, 2nd edn, p. 378.