Ideally, treatment for stroke patients should be coordinated within a stroke unit, as there is a 28% reduction in mortality and disability at 3 months compared to patients treated on general medical wards.⁶ In general only those patients with a compromised airway due to depressed level of consciousness or life-threatening cardiorespiratory disturbances require admission to medical or neurosurgical ICUs. In either case, attention to basic resuscitation, involving stabilisation of airway, breathing and circulation, is self-evident.

AIRWAY AND BREATHING

Patients with Glasgow Coma Scores (GCS) of 8 or less or those with absent gag will require intubation to preserve their airway and prevent aspiration. Where this requirement is likely to be prolonged, early tracheostomy should be considered. Adequate oxygenation and ventilation should be confirmed by arterial blood gas analysis, and supplemental oxygen prescribed if there is any evidence of hypoxia. If hypercarbia occurs then ventilatory support to achieve normocarbia is necessary to prevent exacerbation of cerebral oedema. Ventilatory support in such cases has been shown to result in improved prognosis.⁷

A large number of stroke patients will have raised blood pressure on admission, presumably as an attempt by the vasomotor centre to improve cerebral perfusion. Hypertensive patients may have impaired autoregulation and regional cerebral perfusion may be very dependent on blood pressure.⁸ The patient’s clinical condition and neurological status should determine treatment rather than an arbitrary level of blood pressure. Control of even very high blood pressure (220/120 mmHg; 29.3/16.0 kPa) is not without risk and may result in progression of ischaemic stroke, so reduction should be monitored closely.⁹ It would seem reasonable on physiological grounds to avoid drugs that cause cerebral vasodilatation in that they may aggravate cerebral oedema, although there is no hard evidence for this. Animal experiments have suggested that haemodilution could improve blood flow by reducing whole blood viscosity but a recent multicentre study has failed to identify any clinical benefit.¹⁰ Cardiac output should be maintained and any underlying cardiac pathology such as failure, infarction and atrial fibrillation treated appropriately.

METABOLIC SUPPORT

Both hypo- and hyperglycaemia have been shown to worsen prognosis after acute stroke, therefore blood sugar levels should be maintained in the normal range.¹¹ In the long term, nutritional support must not be neglected and early enteral feeding instituted by nasogastric intubation. In the longer term, particularly where bulbar function is reduced, percutaneous endoscopic gastrostomy is necessary.

ANTICOAGULATION

In theory, the use of anticoagulation reduces the propagation of thrombus and should prevent further embolism. In practice, the reduction in risk of further thromboembolic stroke is offset by a similar number of patients dying from cerebral or systemic haemorrhage as a result of anticoagulation.^12,¹³ Anticoagulation can only be recommended in individuals where there is a high risk of recurrence, such as in those patients with prosthetic heart valves, atrial fibrillation with thrombus or those with thrombophilic disorders. A CT scan must be obtained prior to commencing therapy to exclude haemorrhage, and careful monitoring used. In patients with large infarcts there is always the risk of haemorrhage (haemorrhagic conversion) into the infarct and early heparinisation is best avoided.

THROMBOLYSIS

Systemic thrombolysis with streptokinase carries a very high risk of cerebral haemorrhage and should not be used.¹⁴ There is some evidence that intravenous recombinant tissue plasminogen activator (alteplase) has a better safety/efficacy profile. The National Institute of Neurological Disorders and Stroke (NINDS) trial showed a significant improvement in those patients given alteplase rather than placebo within 3 hours of acute stroke.¹⁵ This benefit was not found to be statistically significant in two European trials^16,¹⁷ and the ATLANTIS trial showed no significant benefit at 90 days in the alteplase group together with an increased risk of intracranial haemorrhage.¹⁸ The 2006 Cochrane Library database states that thrombolysis appears to result in a significant net reduction in the proportion of patients dead or dependent in activities of daily living. However, there appears to be an increase in deaths within the first 7–10 days, symptomatic intracranial haemorrhage and deaths at follow-up at 3–6 months. The data from trials using intravenous recombinant tissue plasminogen activator, from which there is the most evidence on thrombolytic therapy so far, suggest that it may be associated with less hazard and more benefit. There is heterogeneity between the published trials for some outcomes so that the optimum criteria to identify patients most likely to benefit and least likely to be harmed, the latest time window, the agent, dose and route of administration are all unclear. Nevertheless, the data are promising and may justify the use of thrombolytic therapy with intravenous recombinant tissue plasminogen activator in experienced centres in highly selected patients where a licence exists. However, the data do not support the widespread use of thrombolytic therapy in routine clinical practice at this time, but suggest that further trials are needed to identify which patients are most likely to benefit from treatment and the environment in which it may best be given.¹⁹

CEREBRAL PROTECTION

Various agents such as free-radical scavengers, calcium antagonists, magnesium, amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) antagonists, glutamate antagonists and γ-aminobutyric acid antagonists have been used in attempts to limit the deleterious effects of the biochemical changes that occur intracellularly following ischaemia. No agent has been shown to be effective in placebo-controlled phase III trials.

DECOMPRESSIVE CRANIOTOMY

This is an option in young patients with large middle cerebral artery territory infarcts in the non-dominant hemisphere. Untreated, these normally have a mortality of 80% and it is suggested that this procedure can reduce mortality to around 30%, but with residual neurological deficit. This procedure is limited to specialist centres. Other forms of surgical intervention proven to be effective in making more intracranial space and reducing intracranial pressure are drainage of secondary hydrocephalus by extraventricular drain (EVD) insertion or evacuation of haemorrhage into infarcted areas, resulting in new compressive symptoms. This is especially useful in the posterior fossa where the room for expansion of mass lesions is limited by its anatomy.

COMPLICATIONS

Local complications include cerebral oedema, haemorrhage into infarcted areas or secondary hydrocephalus. General complications include bronchopneumonia, aspiration pneumonia, deep-vein thrombosis, urinary tract infections, pressure sores, contractures and depression. A team approach of specialist nursing, physiotherapists, occupational and speech and language therapists is best able to avoid these complications.

SPONTANEOUS INTRACRANIAL HAEMORRHAGE

Spontaneous intracranial haemorrhage producing stroke may occur from either intracerebral haemorrhage (10%) or SAH (5%).

INTRACEREBRAL HAEMORRHAGE

The incidence of intracerebral haemorrhage is about 9/100 000 of the population, mostly in the age range of 40–70 years, with an equal incidence in males and females.

AETIOLOGY AND PATHOLOGY

The commonest cause is the effect of chronic systemic hypertension. This results in degeneration of the walls of vessels or microaneurysms, by the process of lipohyalinosis, and these microaneurysms then suddenly rupture. This may also occur in:

• malignant tumour neovasculature

• malignant hypertension

• primary haemorrhagic disorders

• overanticoagulation

Occasionally, cerebral aneurysms or AVMs may cause intracerebral haemorrhage without SAH. Where intracerebral haemorrhage occurs in young patients the most likely cause is an underlying vascular abnormality. In some areas this is also associated with the abuse of drugs with sympathomimetic activity such as cocaine. The rupture of microaneurysms tends to occur at the bifurcation of small perforating arteries. Common sites of haemorrhage are the putamen (55%), cerebral cortex (15%), thalamus (10%), pons (10%) and cerebellum (10%). Haemorrhage is usually due to rupture of a single vessel, and the size of the haemorrhage is influenced by the anatomical resistance of the site into which it occurs. The effect of the haemorrhage is determined by the area of brain tissue that it destroys. Cortical haemorrhages tend to be larger than pontine bleeds but the latter are much more destructive owing to the anatomical density of neural tracts and nuclei.

CLINICAL PRESENTATION

Usually, there are no prodromal symptoms and a sudden onset of focal neurology or depressed level of consciousness occurs. Headache and neck stiffness will occur in conscious patients if there is subarachnoid extension by haemorrhage into the ventricles. Where intraventricular extension occurs there may be a progressive fall in GCS as secondary hydrocephalus occurs, and this may be accompanied by ocular palsies, resulting in ‘sunset eyes’. As with ischaemic stroke, focal neurology is determined by which area of the brain is involved. The only way to differentiate absolutely between ischaemic, intracerebral or SAH is by appropriate imaging. The symptoms relate to tissue destruction, compression and raised intracranial pressure which, if progressive, will result in brainstem ischaemia and death.

INVESTIGATIONS

The general investigations are essentially those listed previously for ischaemic stroke, since it is difficult to distinguish between the two in the early stages. In addition, CT and/or MRI should be performed at the earliest opportunity. Lumbar puncture may be performed to exclude infection if mycotic aneurysm is suspected, but only after CT has excluded raised intracranial pressure or non-communicating hydrocephalus. Digital subtraction angiography will enable localisation of the source of the haemorrhage and will facilitate any surgical intervention for aneurysm or AVM (see Figures 42.7, 44.1 and 44.2).

Figure 44.1 Left frontal intraparenchmal bleed with local oedema.

Figure 44.2 Magnetic resonance imaging: thalamic bleed.

MANAGEMENT

The general management principles are identical to those for ischaemic stroke. There is, of course, no place for anticoagulation or thrombolysis, and reversal of any coagulation defect either primary or secondary to therapeutic anticoagulation must be undertaken as a matter of urgency. A full coagulation screen must be performed and the administration of vitamin K, fresh frozen plasma, cryoprecipitate, etc. directed by the results. Where intraventricular extension has occurred the insertion of an EVD may increase conscious level, particularly in the presence of secondary hydrocephalus. The EVD level should be set such that the cerebrospinal fluid (CSF) drains at around 10 mmHg. The normal production of CSF should produce an hourly output and a sudden fall in output to zero should alert staff to the possibility that the drain has blocked. This is particularly likely if the CSF is heavily blood-stained. The meniscus of the CSF within the drain tubing should be examined for transmitted vascular pulsation or the level of the drain temporally lowered by a few centimetres to see if drainage occurs. If the drain is blocked, secondary hydrocephalus will recur. Because of the risk of introducing infection and causing a ventriculitis, the drain must be unblocked in a sterile manner by the neurosurgeons. Operative decompression of the haematoma should only be undertaken in neurosurgical centres and safe transfer must be assured if this is considered. This is a controversial treatment as the most recently published trial shows that patients with spontaneous supratentorial intracerebral haemorrhage in neurosurgical units show no overall benefit from early surgery when compared with initial conservative treatment.²⁰ Subgroup analysis showed that those patients with intracerebral haemorrhage less than 1 cm from the cortical surface benefited from early surgery. Patients presenting with a GCS of less than 8/15 had an almost universally poor outcome. Not all intracerebral haematomata are amenable to surgery, and the CT scans should be reviewed by the Neurosurgical Unit, preferably prior to transfer by digital image link. Ideally clot evacuation within 6 hours maximum should be the goal.

The management of hypertension following spontaneous intracerebral haemorrhage may be difficult as too high a blood pressure may provoke further bleeding whereas too low a blood pressure may result in ischaemia. The Stroke Council of the American Heart Association has recommended that in patients with chronic hypotension the mean arterial pressure should not exceed 130 mmHg and that, if systolic blood pressure falls to 90 mmHg, then vasopressors should be used.²¹ The administration of mannitol prior to transfer should be discussed with the Neurosurgical Unit. There is no place for steroids, and hyperventilation to PaCO₂ of 30 mmHg (4 kPa) or less to control raised intracranial pressure will have detrimental effects on cerebral blood flow in other areas of the brain.

SUBARACHNOID HAEMORRHAGE

SAH refers to bleeding which occurs principally into the subarachnoid space and not into the brain parenchyma. The incidence of SAH is around 6/100 000: the apparent decrease, compared with earlier studies, is due to more frequent use of CT scanning which allows exclusion of other types of haemorrhage. Risk factors are the same as for stroke, but SAH patients are usually younger, peaking in the sixth decade, with a female-to-male ratio of 1.6:1. Black people have twice the risk for SAH as whites. Between 5 and 20% of patients with SAH have a positive family history, with first-degree relatives having a three- to sevenfold risk, whereas second-degree relatives have the same degree of risk as the general population. Specific inheritable disorders are rare and account for only a minority of all patients with SAH.²² The only modifiable risk factors for SAH are smoking, heavy drinking and hypertension, which increase the risk odds ratio by 2 or 3.²³ Overall mortality is 50%, of which 15% die before reaching hospital, with up to 30% of survivors having residual deficit producing dependency.

AETIOLOGY AND PATHOLOGY

The majority of cases of SAH are caused by ruptured saccular (berry) aneurysms (85%), the remainder being caused by non-aneurysmal perimesencephalic haemorrhage (10%) and rarer causes such as arterial dissection, cerebral or dural AVMs, mycotic aneurysm, pituitary apoplexy, vascular lesions at the top of the spinal cord and cocaine abuse. Saccular aneurysms are not congenital, almost never occur in neonates and young children and develop during later life. It is not known why some adults develop aneurysms at arterial bifurcations in the circle of Willis and some do not. It was thought that there was a congenital weakness in the tunica media, but gaps in the arterial muscle wall are equally as common in patients with or without aneurysms and once the aneurysm is formed, the weakness is found in the wall of the sac and not at its neck.²⁴ The association with smoking, hypertension and heavy drinking would suggest that degenerative processes are involved. Sudden hypertension plays a role in causing rupture, as shown by SAH in patients taking crack cocaine or, rarely, with sulfenidil.

CLINICAL PRESENTATION

Classically, there is a ‘thunderclap’ headache developing in seconds, with half of patients describing its onset as instantaneous. This is followed by a period of depressed consciousness for less than an hour in 50% of patients, with focal neurology in about 30% of patients.²⁵ About a fifth of patients recall similar headaches and these may have been due to ‘warning leaks’. The degree of depression of consciousness depends upon the site and extent of the haemorrhage. Meningism – neck stiffness, photophobia, vomiting and a positive Kernig’s sign – is common in those patients with higher GCS. The clinical severity of SAH is often described by a grade, the most widely used being that described by the World Federation of Neurological Surgeons (WFNS),²⁶ which is summarised in Table 44.1.

Table 44.1 Clinical neurological classification of subarachnoid haemorrhage

Grade	Signs
I	Conscious patient with or without meningism
II	Drowsy patient with no significant neurological deficit
III	Drowsy patient with neurological deficit – probably intracerebral clot
IV	Deteriorating patient with major neurological deficit (because of large intracerebral clot)
V	Moribund patient with extensor rigidity and failing vital centres

WFNS grade	GCS	Motor deficit
I	15	Absent
II	14–13	Absent
III	14–13	Present
IV	12–7	Present or absent
V	3–6	Present or absent

WFNS, World Federation of Neurological Surgeons; GCS, Glasgow Coma Score.

This grading, together with the extent of the haemorrhage and the age of the patient, gives some indication of the prognosis, in that the worse the grade the bigger the bleed, and the older the patient, the less likely a good prognosis.

COMPLICATIONS

The clinical status of the patient may be complicated by factors other than the physical effect of initial bleed, and factors such as acute hydrocephalus, early rebleeding, cerebral vasospasm, parenchymal haematoma, seizures and medical complications must be considered.

ACUTE HYDROCEPHALUS

This may occur within the first 24 hours postictus and is often characterised by a one-point drop in the GCS, sluggish pupillary responses and bilateral downward deviation of the eyes (‘sunset eyes’). If these signs occur, CT scan should be repeated and, if hydrocephalus is confirmed, or there is a large amount of intraventricular blood, then a ventricular drain may be inserted. This is not without risk as it may provoke more bleeding and introduce infection. Only observational studies exist and these are contradictory so no firm recommendation can be made (Figure 44.3).

Figure 44.3 Spontaneous subarachnoid haemorrhage and secondary hydrocephalus.

REBLEEDING

This may occur within the first few hours after admission and 15% of patients may deteriorate from their admission status.²⁶ They may require urgent intubation and resuscitation but not all rebleeds are unsurvivable and such deterioration should be treated. The chance of rebleeding is dependent on the site of the aneurysm, presence of clot, degree of vasospasm, age and sex of the patient. A preliminary report of the Co-operative Aneurysm Study gave the rebleeding rate in the first 2 weeks after the initial SAH was reported as 19% cumulative and then approximately 1.5% per day for the next 13 days in 1983.^27,²⁸

CEREBRAL VASOSPASM

This is the term used to describe narrowing of the cerebral blood vessels in response to SAH seen on angiography. It occurs in up to 70% of patients, but not all of these patients will have symptoms.²⁹ Use of transcranial Doppler (TCD) to estimate middle cerebral artery blood velocity has shown that a velocity of more than 120 cm/s correlates with angiographic evidence of vasospasm. This technology allows diagnosis in the ICU and provides a means of monitoring the success of treatment to reduce vasospasm, which is undertaken to reduce the severity of delayed neurological deficit secondary to vasospasm. The problem is that not all patients who have angiographic vasospasm or high Doppler velocities have symptoms. If there is evidence of a depressed level of consciousness in the absence of rebleeding, hydrocephalus or metabolic disturbances, but there is evidence of vasospasm on TCD or angiogram, then it would seem appropriate to initiate treatment to reduce the vasospasm. If vasopasm occurs at the time of angiography or coiling, then intravascular vasodilators such as papaverine have been used.

PARENCHYMAL HAEMATOMA

This may occur in up to 30% of SAH following aneurysm rupture and has a much worse prognosis than SAH alone.²⁹ If there is mass effect with compressive symptoms then evacuation of haematoma and simultaneous clipping of the aneurysm may improve outcome.

MEDICAL COMPLICATIONS

During the placebo-controlled study of nicardipine in WFNS grade I and II SAH patients, there was a 40% incidence of at least one life-threatening medical complication in the placebo group.³⁰ The mortality due to medical complications was almost the same as that due to the combined effects of the initial bleed, rebleeds and vasospasm. The types of medical complication seen are shown in Table 44.2.

Table 44.2 Types of medical complication seen in patients with subarachnoid haemorrhage

Medical complication	Incidence
Arrhythmias	35%
Liver dysfunction	24%
Neurogenic pulmonary oedema	23%
Pneumonia	22%
ARDS and atelectasis	20%
Renal dysfunction	5%

ARDS, acute respiratory distress syndrome.

INVESTIGATIONS

The general investigations for stroke should be performed and early CT imaging is mandatory. Blood appears characteristically hyperdense on CT and the pattern of haemorrhage may enable localisation of the arterial territory involved. Very rarely, a false-positive diagnosis may be made if there is severe generalised oedema resulting in venous congestion in the subarachnoid space. Small amounts of blood may not be detected and the incidence of false-negative reports is around 2%.³¹ It may be difficult to distinguish between post-traumatic SAH and primary aneurysmal SAH, which precipitates a fall in the level of consciousness that provokes an accident or fall.³² MR scanning is particularly effective for localising the bleed after 48 hours when extravasated blood is denatured, and provides a good signal on MRI.³³

Lumbar puncture is still necessary in those patients where the suspicion of SAH is high despite a negative CT, or there is a need to exclude infection. There must be no raised intracranial pressure and at least 6 hours should have passed to give time for the blood in the CSF to lyse, enabling xanthochromia to develop.

Angiography via arterial catheterisation is still the most commonly used investigation for localising the aneurysm or other vascular abnormality prior to surgery. It is generally performed on patients who remain, or become, conscious after SAH. It is not without risk and aneurysms may rupture during the procedure and a meta-analysis has shown a complication rate of 1.8%.³⁴ Other methods under investigation include CT angiography and MR angiography, and it is likely that these techniques may replace catheter angiography for diagnosis, although intra-arterial catheterisation would still be needed for endovascular therapy.^35,³⁶

Intracranial pressure monitoring is of limited use in SAH patients except in those where hydrocephalus or parenchymal haematoma is present and early detection of pressure increases may be the trigger for drainage or decompressive surgery.

Transcranial Doppler studies may be useful in detecting vasospasm or those patients in whom autoregulation is impaired.³⁷ The technique is dependent on there being a ‘window’ of thin temporal bone allowing isonation of the Doppler signal along the middle cerebral artery. It is very user-dependent and 15% of patients do not have an adequate bone window.

MANAGEMENT

GENERAL CARE

The initial management of SAH is influenced by the grading, medical comorbidity or complications, and the timing or need for surgery. Patients with decreased GCS may need early intubation and ventilation, simply for airway protection, whereas those with less severe symptoms require regular neurological observation, analgesia for headache and bedrest prior to investigation and surgery. Other management options are:

• stress ulcer prophylaxis

• deep-vein thrombosis prophylaxis using compression stockings or boots

• seizures control with phenytoin or barbiturates

If the patent is sedated and ventilated, the use of an analysing cerebral function monitor should be considered to detect subclinical seizure activity.

Hyponatraemia is a common finding and adequate fluid therapy with normal saline is required with electrolyte levels maintained in the normal range. Occasionally, as in other types of brain injury, excessive natriuresis occurs and may result in hyponatraemic dehydration – cerebral salt-wasting syndrome (CSWS). Its aetiology is not known but some suggest increased levels of atrial natriuretic peptide. It usually occurs within the first week after insult and resolves spontaneously in 2–4 weeks. Failure to distinguish CSWS from the syndrome of inappropriate secretion of antidiuretic hormone (SIADH) could lead to inappropriate treatment by fluid restriction, which would have adverse effects on cerebral perfusion. Urine sodium concentrations are usually elevated in both SIADH and CSWS (> 40mmol/l) but urinary sodium excretion, urine sodium concentration [Na mmol/l] × urine volume [l/24 hours] is high in CSWS and normal in SIADH. If CSWS does not respond to fluid replacement with saline or is not self-limiting then fludricortisone therapy may be useful.

BLOOD PRESSURE CONTROL

Elevation of blood pressure is commonly seen after SAH and there are no precise data on what constitutes an unacceptably high pressure that is likely to cause rebleeding. Equally, there are no precise data on a minimum level of pressure below which infarction is likely to occur, since this will depend on the patient’s normal pressure, degree of cerebral oedema and the presence or absence of intact autoregulation. One observational study has demonstrated reduced rebleeding but higher rates of infarction in newly treated compared with untreated post-SAH hypertensive patients.³⁸ If blood pressure exceeds 200 mmHg (26.6 kPa) systolic or 100 mmHg (13.3 kPa) diastolic, then empirically it would seem wise to reduce this pressure in SAH patients who have unclipped aneurysms. β-adrenergic blockers or calcium antagonists are the most widely used agents, since drugs producing cerebral vasodilation may increase intracranial pressure.

VASOSPASM

Angiographic demonstration of vasospasm may be seen in about 70% of SAH patients but only about 30% develop cerebral symptoms related to vasospasm. Transcranial Doppler-derived flow velocities in the middle cerebral arteries of more than 120 cm/s are accurate in predicting ischaemia.³⁷ Symptoms tend to occur between 4 and 14 days postbleed, which is the period when cerebral blood flow is decreased after SAH.

One method of pre-empting vasospasm is the prescription of oral nimodipine at 60 mg given 4-hourly for 21 days, which has been shown to achieve a reduction in the risk of ischaemic stroke of 34%.³⁹ Intravenous nimodipine should be used in the patients who are not absorbing, but it must be titrated against blood pressure to avoid hypotension. Other calcium antagonists, notably nicardipine and the experimental drug AT877, reduce vasospasm but do not improve outcome.

Low cerebral blood flow is known to worsen outcome and this resulted in the development of prophylactic hypertensive hypervolaemic haemodilution – so-called triple-H therapy.⁴⁰ As originally described, the therapy involved the use of fluid loading to achieve haemodilution and vasopressor therapy to increase cerebral blood flow, and was combined with surgery within 24 hours if possible. The therapy was continued for 21 days and all patients remained neurologically stable or improved as a result, apparently, of the absence of vasospasm. Very few centres use the strict protocol as originally described, but fluid loading rather than fluid restriction is the norm and inotropes or vasopressors are used subsequently if neurological function decreases. Despite its widespread use, there remains no prospective randomised trial that demonstrates its utility.

The use of intravascular catheters to deliver papaverine into vasospastic vessels remains experimental. Where symptoms develop it is important to exclude other causes such as rebleeding, hydrocephalus or metabolic disorder.

SURGERY

Clipping of the aneurysm is the surgical treatment of choice, with wrapping, proximal ligation or bypass grafting being used if the aneurysm is inaccessible to Yasargil clipping. The timing of surgery remains debatable. Early surgery, within 3 days of the bleed, has the advantage of fewer deaths occurring from rebleeding but is technically difficult due to the friability of associated tissues. Delayed surgery, around 10–12 days postbleed, gives better operating conditions but allows some patients to suffer a rebleed. There is no type 1 evidence but an observational study suggested that there is no difference in outcome after early or late operation.⁴¹

ENDOVASCULAR COILING

The development of detachable microcoils made of platinum by Gugliemi in 1990 has resulted in endovascular embolisation of aneurysms or AVMs by interventional radiology.⁴² A microcatheter is passed from the femoral artery to the cerebral aneurysm and the coils positioned sequentially in the lumen of saccular aneurysms to occlude it. Rupture of the aneurysm or adjacent vessel occlusion, causing ischaemia, is the most frequent complication.⁴³ The International Subarachnoid Aneurysm Trial (ISAT) of neurosurgical clipping versus endovascular coiling in 2143 patients with ruptured intracranial aneurysms – a randomised comparison of effects on survival, dependency, seizures, rebleeding, subgroups and aneurysm occlusion – has come down in favour of coiling over the open surgical technique at 1-year follow-up. In patients with ruptured intracranial aneurysms suitable for both treatments, endovascular coiling is more likely to result in independent survival at 1 year than neurosurgical clipping; the survival benefit continues for at least 7 years. The risk of late rebleeding is low, but is more common after endovascular coiling than after neurosurgical clipping.⁴⁴ Additional complications of coiling include rupture during catheter placement in the aneurysm, coil embolisation and vasopasm. Not all aneurysms, particularly those with wide necks, multiple filling vessels or giant aneurysms, are suitable for coiling.