Contributors
J. Michael Schmidt, PhD, Assistant Professor of Neuropsychology (in Neurology) Neurological Intensive Care Unit, Columbia University College of Physicians and Surgeons,
7 Medical Management of Subarachnoid Hemorrhage
Yekaterina K. Axelrod , MD, Fellow
Department of Neurosciences Critical Care, Washington University School of Medicine, St. Louis, Missouri , USA
Michael N. Diringer, MD, Professor a and Director b
aDepartments of Neurology and Neurological Surgery, bNeurocritical Care Unit, Washington University School of Medicine , St. Louis , Missouri , USA
PRESENTATION
A sudden dramatic elevation of intracranial pressure (ICP) accounts for the classic presentation of acute aneurysm rupture—the instantaneous onset of a severe headache often described by patients as “the worst headache of my life” ( 1 ). Headache frequently presents during physical exertion and is often associated with nausea and vomiting. Syncope (mainly due to increased ICP) is seen in about half of cases and is usually followed by a gradual improve-ment in the level of consciousness. Focal neurologic signs appear in only 10% to 15% of cases and typically represent mass effect from a giant aneurysm (e.g., an enlarging posterior communicating artery aneurysm pressing upon the third cranial nerve), intraparenchymal hemorrhage, subdural hematoma, or a large localized subarachnoid clot. There may be a report of a seizure, but it is often unclear as to whether these episodes are true epileptic events or abnormal posturing. Up to 20% of patients with subarachnoid hemorrhage (SAH) may develop intraocular hemorrhages ( 2 ).
EVALUATION AND INITIAL MANAGEMENT
The initial steps in the assessment of a patient with suspected SAH should always focus on the level of neurologic function and the ability to protect the airway. The Hunt and Hess Scale and the more recently developed World Federation of Neurological Surgeons Scale are the most widely used measures of the patient’s clinical condition ( Tables 1 and 2 ).
Diagnostic Evaluation
The best initial diagnostic test is a noncontrasted computed tomography (CT) scan, which is able to detect more than 95% of subarachnoid bleeds on the day of onset ( 3,4 ). Blood typically appears as a high-density signal in the perimesencephalic, interpeduncular, and basal cisterns, Sylvian fi ssure, and sulci of the brain ( Fig. 1 ). CT scans may be falsely negative under certain circumstances: if hemorrhage occurred several days prior (sensitivity drops to 50% in a week) ( 5 ), if the amount of extravasated blood is small, or if hematocrit is below 27% ( 6 ). The amount of subarachnoid blood is graded by Fisher Scale ( Table 3 ), which is also used to predict the risk of vasospasm. At this stage, the size of the third ventricle and the temporal horns of the lateral ventricles should be evaluated to detect and promptly treat early hydrocephalus.
Lumbar puncture should only be performed in cases of high clinical suspicion of SAH in the context of negative CT. It may be diffi cult to distinguish a traumatic lumbar puncture from a true SAH. The presence of xanthochromia, best determined by spectrophotometry ( 4 ), refl ects high levels of either cerebrospinal fl uid (CSF) protein (>200dL mg/dL) or bilirubin, the end product of hemoglobin conversion (a complex process that takes up to 12 hr). Determining if cell count declines from the fi rst to last collected tube of CSF does not reliably distinguish between traumatic tap and SAH.
With the current emphasis on early surgical or endovascular treatment of aneurysms, conventional angiography should be performed as soon as possible ( Fig. 2 ). As multiple aneu-rysms are present in approximately 20% of cases ( 7,8 ), selective four-vessel angiography is
necessary. Magnetic resonance (MR) and CT angiography are not presently suffi ciently sensitive to replace conventional angiography, but the techniques are rapidly advancing ( 9,10 ). CT and MR angiograms may serve as an additional tool with which to plan surgery ( 11 ).
In 15% to 20% of patients, the cause of nontraumatic SAH remains undetermined even after angiography. In a few instances, this may be due to vasospasm or inadequate views incapable of detecting smaller aneurysms. Repeat angiography is therefore often recommended in a few days to two weeks. In patients whose bleeding is confi ned to the perimesencephalic and ambient cisterns, a source of bleeding might not be evident despite high-quality angiograms. Such hemorrhages are referred to as “perimesencephalic SAH”
and have been attributed to venous rather than arterial bleeding ( 12 ). However, detecting the typical perimesencephalic pattern of SAH on CT should not prevent one from pursuing angiography, because this pattern may be seen with aneurysmal rupture ( 13,14 ), especially in the posterior circulation ( 15 ).
Initial Stabilization
Patient’s clinical condition must be carefully assessed and stabilized prior to diagnostic test-ing. If the patient is lethargic or agitated, management of the airway should be addressed fi rst.
During angiography, sedation is frequently used, which could lead to airway obstruction in lethargic patients. To facilitate a safe and rapid study for agitated patients, elective endotracheal intubation should be considered.
Generalized sympathetic activation with high catecholamine levels, as well as pain and anxiety, generally cause elevated blood pressure after SAH. Because hypertension is associ-ated with aneurysmal rerupture, it requires prompt treatment; however, headache should be addressed fi rst. Nimodipine, used routinely for prevention of vasospasm, and analgesics may be suffi cient for blood pressure control in some patients, whereas others may require the administration of additional antihypertensive medications. The most widely employed agents are combined α-1 and β-blocker labetalol or other general β-blockers, none of which raises ICP.
Figure 1 Computed tomography scan of a patient with a sub-arachnoid hemorrhage.
Table 1 Hunt and Hess Scale Grade Symptoms I Asymptomatic or mild headache II Moderate to severe headache,
nuchal rigidity, with or without cranial nerve defi cits
III Confusion, lethargy, or mild focal symptoms
IV Stupor and/or hemiparesis V Comatose and/or extensor posturing
Table 2 World Federation of Neurological Surgeons Grade Glasgow Coma Scale Motor defi cits
I 15 Absent
II 14–13 Absent
III 14–13 Present
IV 12–7 Present or absent
V 6–3 Present or absent
In the presence of a mass lesion or elevated ICP, the use of vasodilators engenders concern, especially venous dilators, such as sodium nitroprusside, which can raise ICP to a signifi cant degree ( 16 ). Debate is ongoing regarding the appropriateness of administering agents that are pure arterial dilators, e.g., hydralazine, nicardipine, and angiotensin-converting enzyme inhibitors; yet many centers routinely use them ( 17 ). One trial demonstrated an improved out-come after SAH with the use of high doses of β-blockers ( 18 ). A notable exception to vigorous management of hypertension is made when patients develop hydrocephalus with potentially higher ICP, in which case, elevated blood pressure helps to maintain adequate cerebral perfu-sion. Only after the treatment of hydrocephalus should hypertension be addressed.
Cardiac abnormalities are common in the fi rst 24 to 48 hr following SAH, and they are almost always completely reversible. Electrocardiographic alterations, including tall-peaked T-waves (“cerebral T-waves”), QT segment prolongation, and ST segment elevation/depres-sion, are frequent; they have been linked to excessive levels of circulating catecholamines ( 19 ).
Cardiac enzymes may be elevated in up to one-third of patients and are variably associated with echocardiographic abnormalities. These disturbances may occur in the absence of coronary artery disease. Myocardial lesions (mainly contraction band necrosis) that have been reported in cases of SAH are pathologically distinct from ischemic lesions ( 20 ). Arrhythmias are frequently seen; however, life-threatening arrhythmias such as ventricular tachycardia have been only rarely noted ( 21 ). Cardiac wall motion abnormalities have also been detected in up to one-third of patients ( 22 ). In rare cases, a picture of a “stunned myocardium” develops, with severe pump
Table 3 Fisher Scale (Based on Initial CT Appearance and Quantifi cation of Subarachnoid Blood)
Grade CT Description
1 No subarachnoid blood on CT
2 Broad diffusion of subarachnoid blood, no clots, and no layers of blood greater than 1 mm 3 Either localized blood clots in the subarachnoid
space or layers of blood greater than 1 mm 4 Intraventricular and intracerebral blood present,
in the absence of signifi cant subarachnoid blood Abbreviation: CT, computed tomography.
Figure 2 Angiographic image of an aneurysm of the MCA. Abbreviation: MCA, middle cerebral artery.
failure, poor cardiac output, pulmonary edema, and hypotension. The most important predic-tors of cardiac dysfunction are those that refl ect the severity of the hemorrhage.
EARLY CRITICAL CARE MANAGEMENT Routine Care and Monitoring
The routine monitoring of all patients with acute SAH should include serial neurologic examinations, continuous CARDIAC monitoring, and frequent determinations of blood pressure, electrolytes, body weight, fl uid balance, and, in many centers, cerebral blood fl ow velocity by transcranial Doppler (TCD). Adequate hydration should be maintained with iso-tonic saline to avoid volume contraction. Monitoring of daily fl uid balance, body weight, and hematocrit is important to ascertain stable intravascular volume.
Anticonvulsants
The risks and implications of seizures associated with SAH are not well defi ned, and the need for and effi cacy of routinely administered anticonvulsants following SAH for seizure pro-phylaxis are yet to be established. A large number of seizure-like episodes are associated with aneurysmal rupture. It is unclear, however, whether these episodes are truly epileptic in origin. In a large prospective study, early seizures were reported to occur in 6% of patients ( 23 ). A retrospective review found that the majority of early seizures occurred before medical presentation; in-hospital seizures were rare in patients who were administered prophylactic anticonvulsants ( 24 ).
The routine use of prophylactic anticonvulsants during the perioperative period has been evaluated in several studies. Nonrandomized studies of patients who underwent craniotomy indicated a benefi t of prophylactic anticonvulsants ( 25–27 ); however, the number of patients with SAH in these reports was too small to address the issue. A study of patients who underwent coil embolization of an aneurysm (rather than surgery) reported no periprocedural seizures and a delayed de novo seizure rate of 1.7% ( 28 ). Risk factors for seizures after SAH include middle cerebral artery aneurysms ( 29,30 ), intraparenchymal hematoma ( 25,29,31 ), infarcts ( 32 ), and a history of hypertension ( 33 ). Noteworthily, a recent study found that phenytoin routinely administered to patients following SAH for seizure prophylaxis was associated with worse neurologic and cognitive outcome ( 34 ).
Steroids
Dexamethasone is widely used to reduce meningeal irritation and intra- and postoperative edema, but no convincing evidence that documents its effi cacy exists. A prospective, randomized, controlled trial of tirilazad mesylate, a nonglucocorticoid 21-aminosteroid, failed to show any benefi t ( 35,36 ).
Rebleeding
The risk of rebleeding is highest immediately following hemorrhage (4–6% over the fi rst 24 hrs) and declines over the next few days. At two weeks, the cumulative risk approaches 20%. Rebleed-ing rates are highest in women and in those with poor medical condition and with elevated systolic blood pressure. Almost a half of the patients who rebleed do not survive.
In the past, antifi brinolytic agents, such as epsilon-aminocaproic or tranexamic acid, were routinely administered to prevent rebleeding. It is clear that these agents do reduce the inci-dence of rebleeding; however, this benefi t was offset by an increase in hydrocephalus and more ischemic infarctions from vasospasm ( 37 ). A meta-analysis of several trials revealed no overall effect on outcome ( 38 ). With the advent of early surgical and endovascular management, the use of these agents has declined dramatically. Still, short-term use of antifi brinolytics has been suggested for patients awaiting surgery or endovascular treatment. The data are mixed as to whether short courses are also associated with more vasospasm.
Anecdotally, rerupture has been associated with systemic hypertension and sudden drops or elevations of ICP, the latter caused by coughing, sneezing, straining, and Valsalva maneu-vers. Hence, initial management focuses on avoiding these factors. Measures should be taken to minimize coughing and straining. In intubated patients, verifying the position of the endo-tracheal tube and administering antitussives and local anesthetics may be necessary if patients
cough excessively. Stool softeners are given routinely to prevent straining. Slow CSF drainage during lumbar puncture or ventriculostomy is recommended.
Excessive stimulation of patients has traditionally been avoided to prevent fl uctuations in blood pressure. Although adverse effects of such stimulation have never been established, it seems prudent to medicate agitated or combative patients. Ideally, they should be sedated to the point of drowsiness but should remain responsive to stimulation. Care must be taken to prevent oversedation, so that clinical deterioration can be easily recognized. Opiates provide not only sedation but also analgesia for treating headache; long-acting sedative agents, such as barbiturates, should be avoided.
Defi nitive prevention of rebleeding is accomplished by aneurysm repair. The old notion that surgery is more diffi cult and results in a worse outcome when performed early (within three days of hemorrhage) has not been supported by careful analysis. Outcome of patients with Hunt and Hess grades II and III is improved with early surgery. Additionally, repair of aneurysms has the further advantage of permitting safe elevation of blood pressure to treat vasospasm.
A multicenter, randomized trial recently compared one-year outcomes in acute SAH patients who were randomized to have their aneurysm repaired either by surgical or endovas-cular means. Of the almost 10,000 patients screened, only approximately 2000 met the inclusion criteria, which required that the treating physicians agree that the aneurysm could be successfully repaired by either means. At one-year, outcome was somewhat better in patients treated by endovascular coiling, and long-term follow-up is underway to assess rebleeding rates ( 39 ).
Hydrocephalus
Acute hydrocephalus (ventricular enlargement within 72 hr) is reported to occur in about 20%
to 30% of patients ( 40–43 ). The ventricular enlargement is often accompanied by intraventricu-lar blood ( 44,45 ), whereas hydrocephalus without intraventricuintraventricu-lar hemorrhage is associated with the amount and distribution of cisternal blood ( 46,47 ). Acute hydrocephalus is more frequent in patients with poor clinical grade and a higher Fischer grade ( 40–43 ).
The clinical signifi cance of acute ventriculomegaly after SAH is uncertain, because many patients are apparently asymptomatic and do not deteriorate ( 48 ). Yet, in patients with diminished level of consciousness, 40% to 80% show some degree of improvement after ventriculostomy ( 45,48,49 ). Based on two small series, the placement of a ventriculostomy may ( 50 ) or may not ( 51 ) be associated with rebleeding.
Delayed hydrocephalus requiring permanent shunting procedures is reported at rates of 18% to 26% of surviving patients ( 42,52,53 ). The need for permanent CSF diversion is associ-ated with older age, early ventriculomegaly, presence of intraventricular hemorrhage, poor clinical condition on presentation, and female gender ( 54–58 ). Two single center series sug-gest that routine fenestration of the lamina terminalis during microsurgical aneurysm repair reduces the incidence of chronic hydrocephalus ( 59,60 ). On the other hand, rates are no dif-ferent in patients who undergo clipping or endovascular treatment of their aneurysms ( 52,53 ).
Ventriculoatrial, ventriculoperitoneal, or lumboperitoneal shunts may improve clinical status in this group of patients ( 61,62 ).
Stunned Myocardium
In rare cases, myocardial contractility may be impaired after SAH, leading to a fall in cardiac out-put and blood pressure, with subsequent pulmonary edema. Even though this condition has been referred to as “stunned myocardium,” an element of neurogenic pulmonary edema may also be seen.
The management of this condition is similar to treatment of acute pump failure of other etiologies and includes administration of inotropic agents, diuretics, high concentrations of oxygen, and posi-tive end-expiratory pressure. This state is surprisingly transient and is usually completely reversed in an average of four days.
Pulmonary Complications
Pulmonary complications are frequent in patients with SAH and represent a significant cause of morbidity and mortality. Among the most common are nosocomial and aspiration pneumonia, pulmonary edema (neurogenic and cardiogenic), and pulmonary embolism ( 21,63 ). Older patients with worse Hunt and Hess grade and lower Glasgow Coma Scale score on admission are at higher risk for developing pulmonary complications ( 63 ).
Patients with pulmonary complications were shown to have a greater incidence of symp-tomatic vasospasm.
Most cases of pneumonia are seen in patients who require mechanical ventilation. Extu-bation should be performed as early as clinically possible. Care must be taken to prevent acci-dental extubation. Nasally placed endotracheal tubes should be avoided to diminish risk of sinusitis. All mechanically ventilated patients should be maintained in semirecumbent position to prevent aspiration. Adequate enteral nutrition is crucial; however, large gastric volumes should be avoided ( 64 ).
Neurogenic pulmonary edema may be seen acutely, within hours of the initial ictus, or a few days after the onset of aneurysmal rupture. Clinical features are nonspecifi c and similar to those of cardiogenic pulmonary edema with respiratory distress, tachycardia, and hypotension.
Treatment is generally supportive and includes the administration of both supplemental oxy-gen, to maintain adequate tissue oxygenation, and diuretics; patients with severe hypoxemia require ventilatory support with high positive end-expiratory pressure. Insertion of a pulmo-nary artery catheter may be warranted, especially if hemodynamic therapy for vasospasm is required. Neurogenic pulmonary edema is often reversible within 48 to 72 hrs.
LATE COMPLICATIONS Venous Thromboembolism
Multiple risk factors are associated with venous thromboembolism in this population, includ-ing craniotomy, lower limb paralysis, advanced age, and prolonged ICU stay with indwellinclud-ing venous catheters. Methods of mechanical prophylaxis (graduated compression stockings and pneumatic compression devices) have been proven to be safe and effective. Adjunct of low-molecular-weight heparins is a more effi cacious option at the expense of an additional mild risk of intracranial hemorrhagic complications ( 65–67 ).
Glucose Management
Hyperglycemia has been associated with poor functional outcome in patients with SAH ( 68,69 ).
An aggressive approach to achieve normoglycemia should be instituted early. A large randomized study of intensive insulin therapy in surgical critically ill patients demonstrated a signifi cant reduction in morbidity and mortality ( 70 ).
Fever
Fever is commonly seen in patients with SAH even without infections; its detrimental effects on neurologic outcome, mortality, and vasospasm have been recognized ( 71,72 ). Both novel phys-ical (intravascular catheters) ( 73 ) and conventional pharmacologic means of lowering body temperature are useful; however, many patients after SAH remain refractory to any efforts to achieve normothermia. A multicenter study of intraoperative hypothermia to 33°C that did not follow postoperative fever management failed to show benefi t in neurologic outcome ( 74 ).
Hyponatremia and Intravascular Volume Contraction
Hyponatremia occurs in up to one-third of patients following SAH. Although originally attrib-uted to the syndrome of inappropriate secretion of antidiuretic hormone, the picture is more complex. Disturbances of humoral and neural regulation of sodium, intravascular volume, and water in SAH lead to intravascular volume contraction and hyponatremia, sometimes referred to as cerebral salt wasting. When administered the usual 2 to 3 L of fl uids per day, up to half of patients develop intravascular volume contraction. The important clinical consequences include a higher rate of symptoms in patients with angiographic vasospasm and an increase in the number of delayed ischemic defi cits that progress to infarction.
Levels of several circulating natriuretic factors are elevated following SAH, yet antidi-uretic hormone levels remain elevated during hyponatremia. Administration of large volumes of isotonic fl uids can prevent volume contraction and appears to limit the severity of the hypo-natremia. Hyponatremia can frequently be managed with restriction of all free water by giving only isotonic or hypertonic intravenous fl uids (1.25–3.0% saline), minimizing oral liquids, and using concentrated enteral feedings.
Two randomized, controlled trials have evaluated the ability of the mineralocorticoid fl udrocortisone to correct hyponatremia and fl uid balance. One ( 75 ) found that it helped
correct the negative sodium balance but not volume contraction or hyponatremia; the second ( 76 ) reported a reduced need for fl uids and improved sodium levels with fl udrocortisone.
Vasospasm
In the context of SAH, the term “vasospasm” refers to a condition that is much more complex than simply constriction of blood vessels. Additional contributing factors include pathologic changes in arterial walls, with narrowing of the lumen and impairing vascular relaxation, altered vascular reactivity, and hypovolemia. The term “delayed ischemic defi cit” (DID) more aptly describes the situation in which these multiple factors conspire to result in cerebral ischemia.
Vasospasm may be defi ned on the basis of angiographic, TCD, or clinical criteria. Angio-graphic segmental or diffuse narrowing of vessels occurs in up to 70% of patients. These changes are usually seen 5 to 14 days after the hemorrhage but may occur from as early as day 2 and as late as 3 weeks following the bleed. TCD criteria defi ne vasospasm, using absolute linear blood fl ow velocity (LBFV), as mild (>120 cm/sec), moderate (>160 cm/sec), or severe (>200 cm/sec). The rate of rise in the LBFV may be a more sensitive indicator of vasospasm. When compared to angiography, TCD has a sensitivity of about 80%, refl ecting that it samples only a small segment of the vasculature.
DIDs develop in approximately one-third of SAH patients, primarily in those with large amounts of subarachnoid blood. The syndrome presents as either a gradual decline in level of consciousness or the appearance of new focal neurologic defi cits. These fi ndings may fl uctuate and can be exacerbated by hypovolemia or hypotension.
The disparity between the incidence of angiographic, TCD, and clinical vasospasm com-plicates management. All agree that when clinical symptoms develop, they should be treated aggressively; however, there is disagreement about how to respond to elevated TCD velocities or angiographic vasospasm in the absence of symptoms.
The management of vasospasm involves both the routine “prophylactic” measures used for all patients and the more aggressive intervention reserved for situations with signs of active vasospasm. Routine measures include the administration of the centrally acting calcium chan-nel blocker nimodipine, avoidance of intravascular volume contraction, and mechanical means to remove subarachnoid blood at the time of surgery.
Nimodipine has been shown to reduce the incidence of ischemic infarctions and improve outcome in several prospective controlled studies; it is routinely administered to all patients ( 77 ). Hypotension is infrequent with the usual doses (60 mg every 4 hr), especially if patients are well hydrated. In those treated with vasopressors for symptomatic vasospasm, dips in blood pressure following nimodipine administration could be more of a problem; therefore, changing the dose to 30 mg every 2 hr might be helpful.
It is routine practice to remove subarachnoid blood at surgery. Subarachnoid admin-istration of recombinant tissue plasminogen activator completely prevented vasospasm in experimental models; however, in a randomized controlled trial of SAH patients, it did not alter neurologic outcome. Preliminary reports suggest a benefi cial effect of a variation on this approach, using head shaking combined with cisternal irrigation with urokinase ( 78 ). Another recent study found improved outcome with reintroduction of the old practice of draining large volumes of CSF from the lumbar space to help clear subarachnoid blood ( 79 ).
Prophylactic Hypervolemia
A number of recent studies have addressed the use of “prophylactic” hypervolemia. In one pro-spective, randomized trial, routine administration of albumin to keep central venous pressure greater than 7 mmHg did not improve cerebral blood fl ow or outcome ( 80 ). Another report suggested that outcome worsened after the discontinuation of the routine use of albumin ( 81 ).
Due to concern about hyperchloremic metabolic acidosis developing after administration of large amounts of saline, some clinicians partially replace chloride with acetate.
Other Prophylactic Measures
The implantation of prolonged release implants at the time of surgical repair of the aneurysm might help reduce the incidence of vasospasm. Preliminary work with implants impregnated with papaverine ( 82 ) and nicardipine ( 83 ) is promising. A small trial with enoxaparin suggested a reduction in DID ( 84 ). A more aggressive approach, the prophylactic use of transluminal balloon angioplasty in patients with Fisher Grade 3 SAH, was reported to markedly reduce the rate of vasospasm in this high-risk group of patients ( 85 ).
Hemodynamic Augmentation
Currently the mainstay of treatment of symptomatic vasospasm in patients with repaired aneu-rysms is hemodynamic augmentation (HA). HA consists of some combination of large volumes of fl uids, vasopressors, and inotropic agents; however, the optimal combination is unknown.
This intervention has never been tested in a randomized, controlled setting and is based on a number of case series that focused on different aspects of HA.
Once the threshold for aggressive treatment has been crossed, any possible hypovolemia should be corrected rapidly. The use of colloids is controversial, with no clear data to support their effi cacy. Although it appears that the correction of hypovolemia is important, no data clearly indicate benefi ts of hypervolemia over euvolemia. Some degree of acute volume expansion may be helpful in improving cardiac output and blood pressure, but this effect in one study plateaued at pulmonary capillary wedge pressure of 14 mmHg ( 86 ). If fl uid administration produces no immediate response, vasoactive agents should be employed, either inotropic agents (dobuta-mine) to improve cardiac output or vasopressors (phenylephrine) or combined agents (dopamine and norepinephrine) to augment systemic blood pressure.
At issue is the hemodynamic parameter that is best to augment. One approach is to place a pulmonary artery catheter to determine which parameter is most amenable to augmenta-tion. Use of a Swan-Ganz catheter is also helpful in patients with cardiac disease to help guide therapy and prevent fl uid overload or congestive heart failure. Although preliminary reports indicated a high complication rate with this therapy, subsequent studies have found that, with close monitoring, complications were low ( 87 ).
The major focus of all of these therapeutic maneuvers is to improve neurologic function.
If a hemodynamic goal is reached, but no neurologic improvement is observed, the therapeutic maneuver should be reassessed. It should be emphasized that the use of hypertensive therapy is usually not recommended in patients with unclipped aneurysms. The use of HA appears safe in patients with recently coiled aneurysms ( 88 ). Weaning hemodynamic augmentation should be performed gradually, usually over several days, the rate of withdrawal being guided, again, by the neurologic status.
Endovascular Treament of Delayed Ischemic Defi cits
Endovascular approaches to treatment of vasospasm are continuously evolving. It is now a routine practice to perform angioplasty on the proximal segments of the vasospastic cerebral vessels. The angiographic changes following angioplasty are impressive and appear to be long lasting ( Fig. 3 ).
However, clinical effi cacy has been diffi cult to establish, as this procedure is utilized in conjunction
Figure 3 Vasospasm before and after angioplasty. (A) Angiogram with vasospasm in the middle cerebral artery territory ( thin arrow ). (B) Angiogram after angioplasty with improvement in vasospasm ( thick arrow ).