munir.tv
munirelias.com
spine.surgery
humanneurophysiology.com
neurosurgery.tv
craniopharyngiomas.com
pituitaryadenomas.com
meningiomas.org

 
Most of the site will reflect the ongoing surgical activity of Prof. Munir Elias MD., PhD. with brief slides and weekly activity. For reference to the academic and theoretical part, you are welcome to visit  www.munirelias.com

Since its recognition over 60 years ago, cerebral vasospasm has remained an enigma to neurosurgeons. Laboratory and clinical research efforts have lent considerable insight into the etiology, pathophysiology, and potential therapeutic strategies for this disorder, yet the basic mechanisms of arterial narrowing after subarachnoid hemorrhage remain largely undiscovered and current treatments are mostly palliative. Even the term cerebral vasospasm may be imprecise, as controversy exists regarding the role of active vasoconstriction as a cause of decreased vessel caliber. Furthermore, the relationship between angiographic arterial narrowing and potentially reversible neurologic deterioration is poorly understood, further complicating the development of effective therapeutic modalities. In its broadest sense, cerebral vasospasm can be defined as the insidious onset of delayed focal or diffuse narrowing of large capacitance arteries at the base of the brain following hemorrhage into the subarachnoid space. Aneurysmal subarachnoid hemorrhage (SAH) is the most common etiology for vasospasm, although it is seen after hemorrhage from arteriovenous malformations, tumors, or head trauma. The angiographic narrowing, which follows a relatively consistent temporal course, may be entirely asymptomatic or disguised by other complications related to the SAH. In about one-half of cases, however, vasospasm manifests by the occurrence of a delayed neurological ischemic deficit, which may resolve or progress to permanent cerebral infarction. Despite advances in diagnosis and treatment, cerebral vasospasm remains the greatest treatable cause of morbidity and mortality in patients who survive the ictus of SAH. In most contemporary series, up to 15 percent of such patients suffer stroke or death from vasospasm despite maximal therapy.

The presumed mechanism of neurological deterioration in cerebral vasospasm is diminished regional or global cerebral blood flow (CBF) through narrowed segments of major capacitance arteries at the circle of Willis. Some authors have suggested that neurologic deficits in vasospasm are due to direct effects of blood on the brain, disorders of the cerebral microcirculation, or distal emboli. However, positron emission tomography (PET) has documented decreased CBF, increased blood volume, and increased oxygen extraction in the distribution of affected large arteries with vasospasm consistent with regional ischemia. In addition, computed tomography (CT) or autopsy evidence of cerebral infarction after vasospasm is usually (but not invariably) consistent with ischemia in the distribution of major arteries. The gradual onset and non occlusive stenosis of cerebral arteries in vasospasm likely produces wide areas of marginal CBF analogous to the penumbra which occurs after vascular occlusion. Neurons in this marginally perfused region may be nonfunctional at these CBF levels (10 to 20 ml/100 g per min), yet permanent damage may not occur for hours or days. This unique aspect of ischemia from cerebral vasospasm may explain several important features of the disorder, including its evanescent course and the effectiveness of therapeutic strategies such as hypervolemia and hem dilution and calcium channel antagonists. With prompt recognition and institution of therapy, cerebral vasospasm is potentially the most treatable cause of cerebral ischemia.

Clinical and Angiographic Features

The angiographic correlate of cerebral vasospasm is narrowing of intracranial segments of major cerebral arteries at the base of the brain, usually in contrast to a prior angiogram documenting normal caliber of the involved vessels. The arterial narrowing can be focal or diffuse, and is often associated with radiographic evidence of diminished flow in the distal territory of the affected artery. Angiographic vasospasm has a typical temporal course, with onset at 3 to 5 days after the hemorrhage, maximal narrowing at 5 to 14 days, and gradual resolution over 2 to 4 weeks.

Clinical features of vasospasm are less predictable, and depend on many variables including severity and location of the arterial narrowing, age and clinical condition of the patient, the presence of complicating factors (e.g., elevated intracranial pressure or hypotension), and the extent of collateral circulation to the ischemic brain. For these and other indeterminate reasons, one-half of patients with angiographic vasospasm remain asymptomatic. In symptomatic patients, clinical manifestations of vasospasm vary in their presentation, severity, and duration for similar reasons. The delayed ischemic neurological deficit associated with symptomatic vasospasm usually presents shortly after the onset of angiographic vasospasm with the acute or sub acute development of focal or generalized symptoms and signs. For lesions affecting the anterior circulation, these signs include hemiparesis, hemi sensory deficits, visual disturbance, dysphasia, or change in level of consciousness. For vasospasm of the posterior circulation, manifestations may include dysarthria, diplopia, vertigo, ataxia, or altered sensorium. Frequently the onset of neurological signs is preceded by fever, increased meningismus, and nonspecific light-headedness. Once manifest, symptoms and signs may progress in severity or fluctuate, and are often associated with changes in intravascular volume status or blood pressure. Progression to permanent cerebral infarction occurs in approximately 50 percent of symptomatic cases; recovery without deficit in the remaining individuals may occur despite the persistence of angiographic vasospasm.

Incidence

Analysis of the incidence of cerebral vasospasm is complicated by the lack of consistent diagnostic criteria among reported studies. Nevertheless, the incidence of cerebral vasospasm following aneurysmal SAH has probably declined somewhat in recent years. The Cooperative Aneurysm Study in 1987 reported the incidence of angiographic vasospasm at over 50 percent, with symptomatic vasospasm in 32 percent of patients. These values have remained consistent in numerous retrospective reviews. Most patients in recent reports have been treated with regimens of volume expansion and hem dilution (see below). In recent major prospective trials for oral nimodipine, angiographic (50 to 66 percent) and symptomatic (30 to 40 percent) vasospasm were relatively consistent among both placebo and treatment groups. Recent uncontrolled retrospective studies utilizing intravenous calcium-channel antagonists have reported the incidence of symptomatic vasospasm lower than 10 percent.

Vasospasm following head injury is less often recognized because of the declining use of angiography and the coexistence of significant neurological deficits which make diagnosis more difficult, In a retrospective study, Wilkins and Odom described angiographic vasospasm in 19 percent of 350 patients with moderate or severe head injury. The advent of the noninvasive technique of transcranial Doppler (TCD) sonography has renewed interest in vasospasm as a complication of head injury, and may better define its incidence in this setting. Cerebral vasospasm following hemorrhage from arteriovenous malformations is considerably less, frequent than that of aneurysmal subarachnoid hemorrhage, probably because of the absence of extensive subarachnoid blood collections in the former disorder.

Predisposing Factors

Conditions associated with a higher incidence of cerebral vasospasm after SAH are listed in Table. The initial observations of Takemae et al. regarding the relationship between the volume of subarachnoid blood on CT and the subsequent development of vasospasm have been subsequently confirmed in numerous reports, Fisher et al. described a scaling system for the volume of subarachnoid hemorrhage which correlated highly with the risk of subsequent vasospasm. Clinical grade of the patient at the initial presentation also correlates with vasospasm risk, probably because poor-grade patients tend to have more extensive subarachnoid bleeding, The presence of intraventricular blood and coexistent hydrocephalus have been associated with an increased incidence of vasospasm. Elevated peripheral white blood cell counts often predict an increased risk of vasospasm, although this parameter usually manifests at the onset of symptoms. Hyponatremia has been suggested as a predictor of vasospasm, probably due to the associated hypovolemia. Antifibrinolytic agents such as epsilon-aminocaproic acid increase the risk of vasospasm. Several studies have not shown that increased risk of vasospasm is related to sex or aneurysm location; a single retrospective study suggested an increased risk for women with middle cerebral artery aneurysms. Angiographic vasospasm may be more common in younger individuals, although it is less likely to be symptomatic in this age group.

Tab-1:Conditions Associated with Increased Risk of Cerebral Vasospasm

Increased volume of subarachnoid blood on CT
Worse clinical grade
Intraventricular blood and hydrocephalus
Fever/peripheral leukocytosis
Hyponatremia (hypovolemia)
Antifibrinolytic agents
Female with MCA aneurysm

Diagnosis

Effective therapy for cerebral vasospasm depends on early recognition of clinical manifestations: Cognizance of predisposing risk factors for vasospasm is essential. In this regard, the CT scan within 48 h after SAH is a good predictor of vasospasm risk. The definitive diagnosis of vasospasm is made by angiography; however, the invasive nature of this procedure necessitates its use primarily to confirm a diagnosis which is based upon clinical and noninvasive measures. Comprehensive monitoring of patients after SAH, especially those at high risk for vasospasm, will increase the diagnostic accuracy.

Noninvasive Diagnosis

The recent advent of TCD has greatly facilitated the diagnosis of vasospasm. This modality uses range-gated pulsed Doppler insonation through thinner parts of the skull to determine blood velocity in large arteries at the cranial base. With SAH, elevated cerebral arterial blood velocity correlates highly with angiographic vasospasm. Normative values for major cerebral arteries have been established; the ratio of middle cerebral to cervical carotid artery velocity (carotid index) can help differentiate vasospasm from increased cerebral blood flow due to a hyperdynamic state.  A prominent increase in TCD velocity during the first week after SAH is highly characteristic of vasospasm and often precedes the onset of clinical symptoms by several hours. Similarly, restitution of normal TCD velocities usually signals the remission of vasospasm, and can aid in determining therapy duration. Additional noninvasive adjuncts to the diagnosis of vasospasm include methods to assess regional cerebral blood flow, including 133-Xe, xenon-CT, single photon emission tomography (SPECT) and positron emission tomography (PET).

Clinical Diagnosis

The diagnosis of cerebral vasospasm is largely based on careful sequential neurologic examinations by personnel familiar with its manifestations. A high index of suspicion can facilitate diagnosis, thereby accommodating estimated risk factors and the temporal course of the disease. Fever and a slight leukocytosis in peripheral blood frequently herald the onset of sYlJ1ptoms.71.88 A sharp in­crease in TCD velocity (e.g., middle cerebral artery velocity > 120 cm/s) should alert physicians to impending symptoms. At this stage, any change in neurological condition mandates a thorough evaluation to exclude other causes of deterioration, including hydrocephalus, seizures, cerebral edema, electrolyte abnormalities, drug reactions, respiratory insufficiency, or new intracranial hemorrhage. Cerebral ischemia from other causes (e.g., emboli from the aneurysm) must also be considered. Appropriate tests include CT scanning, serum electrolyte and blood gas determinations, and monitoring of intracranial pressure. If alternative causes are excluded, delayed neurological deterioration in this setting is almost certainly due to ischemic consequences of vasospasm. Documentation of focal or global alterations in CBF by various noninvasive measures, as mentioned earlier, further substantiates the diagnosis.

Therapy

Several treatment strategies have emerged due to increased understanding of the pathophysiology of cerebral vasospasm. Nevertheless, no single therapy is a panacea for this disorder, and stroke or death related to vasospasm accounts for a significant proportion of poor outcome related to SAH in patients treated with maximal therapy. Therapy is often initiated prophylactically for some treatments (e.g., hypertension and hypervolemia, thrombolytic agents, and calcium antagonists), or at the onset of symptoms for other therapies (e.g., transluminal angioplasty). Prompt initiation of therapy often produces rapid improvement, again emphasizing the importance of prompt diagnosis of vasospasm.

Poiseuille's law describes theoretical flow through a blood vessel according to the formula:

Flow = ∆рπr4/8LN,

where ∆p = pressure gradient, r = radius, L = length, and N =viscosity. From a simplistic standpoint, current therapies for vasospasm address flow through stenotic segments of cerebral arteries by either augmenting those variables comprising the numerator (i.e., increasing pressure gradient and radius), or by reducing components of the denominator (i.e., decreasing length and viscosity). Of these alternatives, changes in radius are clearly the most important variable in determining flow, yet they have proved to be the most resistant in the treatment of vasospasm.

Vasodilating Agents

Much of the early experimental work in vasospasm focused on developing agents to reduce the narrowing in cerebral arteries, which was presumed to be caused by focal vasoconstriction. In 1986, Wilkins summarized three decades of research into prevention and treatment of intracranial arterial spasm; these methods included the use of vasodilators and antagonists to vasoconstrictors, as well as other pharmacologic means of inhibiting smooth muscle cell contraction. Despite occasional promising reports in experimental models, these agents were uniformly unsuccessful in reversing vasospasm in clinical trials. Chronic exposure to perivascular blood renders cerebral vessels relatively insensitive to both vasoconstricting and vasodilating agents. Calcium-channel antagonists, which were proposed as a therapy for vasospasm based on their ability to inhibit cerebral smooth muscle contraction, are probably effective because of mechanisms other than dilation of narrowed large vessel segments. Thus. interest in vasodilating agents as a treatment for vasospasm waned in the past decade. Recent reports have renewed interest in this mode of therapy, however. In a prospective, randomized trial, intravenous nicardipine ameliorated angiographic vasospasm but failed to improve outcome at 3 months. In uncontrolled trials, high dose intra-arterial papavarine (administered at the site of arterial narrowing) effectively reduced both angiographic and symptomic vasospasm in patients refractory to other therapies. Perhaps ongoing trials will determine the efficacy, durability, and safety of these new treatment modalities.

Hypertension/Hypervolemia/Hemodilution

The normal brain maintains CBF at relatively constant levels over a wide range of blood pressure (autoregulation) by intrinsic mechanisms controlling vascular tone in small arterioles. In ischemic brain, these arterioles are maximally dilated, so that CBF varies more directly with blood pressure (or cardiac output) in a passive manner. Additionally, augmentation of cardiac output by volume expansion with colloid or crystalloid typically lowers blood viscosity by hem dilution. In this manner "triple-H" therapy (hem dilution/hypervolemia/hypertension) potentially might improve CBF in vasospasm by affecting several variables of the Poiseuille equation.

Following the initial report of Kosnik and Hunt in 1976, several reports described resolution of deficits from vasospasm following elevation of blood pressure and/or volume expansion. In these cases, distinct changes in neurological deficits were occasionally observed with fluctuated depending on blood pressure. Outcome related to vasospasm in these uncontrolled series was considerably better than historical controls, leading to widespread application of triple-H therapy. Subsequent reports suggested further reduction in vasospasm when therapy was initiated prior to the onset of symptoms.

Little is known regarding the specific mechanisms by which triple-H therapy affects cerebral vasospasm. Its efficacy has not been demonstrated in controlled trials, and studies of CBF after starting therapy have been equivocal. In addition, it is unclear which component of this therapy (hem dilution vs. hypervolemia vs. hypertension) is most important. Only a portion of patients with vasospasm respond to triple-H therapy, with stroke and death rates approaching 15 percent in the best series.

Initiation of triple-H therapy is associated with significant risk, including cardiac failure, electrolyte abnormalities, cerebral edema, bleeding abnormalities, and rupture of an unsecured aneurysm. Patients receiving this treatment should be monitored in an intensive care setting with Swan-Ganz catheter. arterial line and frequent serum electrolyte determinations. Most protocols utilize measurements of left ventricular end diastolic pressure (LVEDP) and cardiac output to optimize hemodynamics according to the Starling curve. Volume expansion is accomplished using either crystalloid or colloid to achieve LVEDP in the range of 12 to16 mmHg, depending on the patient's age and cardiac status.

Hem dilution with reduction of heamatocrit to less than 35 percent is usually coincident with volume loading. Blood pressure is most often maintained at physiologic levels; augmentation to supranormal values with dopamine or dobutamine may be reserved for neurologic deterioration refractory to hem dilution and hypervolemia. Therapy may be more effective if initiated prophylactically prior to the onset of symptoms (preferably after clipping of the aneurysm, and should be continued beyond the period of risk for vasospasm or until vasospasm abates by clinical and TCD parameters.

Calcium-Channel Antagonists

Calcium-channel antagonists may affect pathologic processes in cerebral vasospasm by a number of mechanisms (Table -3). This class of drugs consists of dihydropyridines (nimodipine, nicardipine, nifedipine), diphenyl alkamines (verapami]), and benzothiazepines (diltiazam), which act to block receptor-mediated calcium channels (L-channels) on smooth muscle cells. Certain agents (diltiazam, nicardipine, and nimodipine) have affinity for cerebra] arterial smooth muscle and effectively antagonize agonist mediated vasoconstriction in vitro. In this regard, calcium channel antagonists might affect either spastic large cerebral arteries or augment collateral flow by dilatation of smaller pial or penetrating vessels. In addition, lipophilic calcium-channel antagonists readily cross the blood-brain barrier, bind to neurons, and inhibit calcium influx after stimulation of glutamate receptors during ischemia. Finally, calcium-channel antagonists affect platelet aggregation and erythrocyte membrane deformability, thus potentially augmenting capillary flow in low-flow states.

Based on the multiple potential beneficial effects for calcium channel antagonists in cerebra] vasospasm, a number of prospective, randomized trials for nimodipine were initiated in the past decade. The characteristics of these trials can be summarized as follows: (I) oral nimodipine consistently reduced poor outcome due to vasospasm in all grades of patients; (2) with the exception of one trial, the incidence of symptomatic vasospasm was not affected by nimodipine treatment; (3) vessel caliber by angiography was not affected by nimodipine therapy; and (4) complications and side effects of the drug were minimal. Combined with observations that nimodipine had no effect upon CBF in vasospasm, these data suggest that the protective effect of nimodipine may have been due to limitation of calcium influx in marginally ischemic neurons, rather than dilatation of large capacitance arteries. Several uncontrolled trials have reported even lower incidence of permanent deficit from vasospasm following the intravenous administration of nimodipine, with rates ranging from 1 to 10 percent. In a prospective, randomized dose­escalation trial, however, there was no significant difference in symptomatic vasospasm or outcome at 3 months for either 0.15 mg/kg per h or 0.3 mg/kg per h compared to controls. Of interest, angiographic vasospasm was significantly less in the nicardipine-treated group. Further studies to elevate the effectiveness of intravenous calcium-channel antagonists are ongoing.

A new class of calcium antagonists have been developed which act to sequester intracellular calcium and inhibit protein kinase C, both of which mediate contractile mechanisms in smooth muscle. One such agent, AT877, significantly reduced symptomatic and angiographic vasospasm and improved outcome at 3 months in a prospective randomized trial.

Clot Removal and Agents Affecting Fibrinolysis

Antifibrinolytic agents (epsilon-aminocaproic acid, tranexemic acid) were widely employed in the 1970s and 1980s to reduce the risk of rebleeding in patients awaiting surgery. Presumably these agents stabilize the thrombus at the site of aneurysm rupture by inhibiting plasmin-mediated thrombolysis. However, they also probably inhibit the lysis of thrombus adjacent to arteries in the subarachnoid space, thus potentially exacerbating vasospasm. These concepts were substantiated by the results of the Cooperative Aneurysm Study in 1987. In this trial, rebleeding rates among patients with delayed surgery were significantly less among those treated with antifibrinolytic agents (24 percent) as compared to non treated patients (45 percent). However, this protective effect was negated by a significant increase in delayed ischemic neurological deficits for patients receiving antifibrinolytics (42 vs. 24 percent). The increasing practice of early surgery in patients with aneurysmal SAH has obviated somewhat the necessity for preventing rebleeding with antifibrinolytic agents; in those patients with delayed surgery the beneficial effect of antifibrinolytic agents must be weighed against the risk of exacerbating vasospasm.

Considerable clinical and experimental evidence bas related the severity of cerebral vasospasm to the volume and duration of perivascular thrombus in the subarachnoid space, as described below. This concept led to aggressive clot removal at surgery, as first advocated by Suzuki et al. However, subarachnoid clot is often tenaciously adherent to the brain and pial vessels, and many neurosurgeons are hesitant to perform additional dissection and retraction during surgery. Weir's group postulated that intrathecal fibrinolytic agents [e.g., tissue plasminogen activator (tPA) and urokinase] might ameliorate vasospasm by hastening the lysis of subarachnoid clot. On the other hand, accelerated thrombolysis by these agents may also increase the risk of postoperative hemorrhage. Following encouraging results in a primate model of vasospasm, this modality has recently been applied in selected clinical cases. Intracisternal recombinant tPA is currently under investigation in a prospective, randomized trial

Transluminal Angioplasty

Zubkov and colleagues at Leningrad Nuerosurgical Institue first reported successful resolution of symptomatic cerebral vasospasm by dilatation of the narrowed segment using an intravascular balloon. This report .remained largely unnoticed until the late 1980s, when steerable balloon catheters enabled safe navigation into intracranial arteries. Since that time there have been numerous uncontrolled reports describing profound improvement in neurological deficits for patients with vasospasm refractory to other modes of therapy. The effects of transluminal angioplasty can be summarized as follows:

(1) significant improvement occurs in 60 to 80 percent of patients. often within minutes of the dilatation; (2) normal angiographic caliber is achieved in nearly all cases. which persists without recurrent vasospasm; (3) evidence of improved CBF by TCD or SPECT correlates with clinical improvement: and (4) complications (rupture of vessels or unsecured aneurysm) occur in approximately 5 percent of cases. Although controlled trials have not been done, the generally good outcome described in these reports is significant due to the ominous natural history of vasospasm in the subset of patients with symptomatic vasospasm refractory to conventional therapy. Several unresolved issues concerning transluminal angioplasty include its application in patients with asymptomatic vasospasm or those with unsecured aneurysms. its safety in widespread application, and its role in relation to intra­arterial papavarine infusion.

Timing of Surgery

Because mechanical stimulation causes transient constriction of cerebral arteries, it was assumed that surgery would exacerbate existing or developing vasospasm. In addition, inflammation and swelling of the brain were thought to coincide with vasospasm, thereby increasing surgical morbidity due to retraction injury. These concepts led to the practice of delaying surgery beyond the period of maximal therapy. Recently, there has been a trend toward early surgery. Early surgery clearly reduces the risk of rebleeding, facilitates the removal of perivascular thrombus, and enables the institution of aggressive therapies for vasospasm such as triple-H therapy, thrombolytics, and angioplasty. The role of early surgery in exacerbating vasospasm remains unresolved due to conflicting reports from uncontrolled trials. Data: from the Cooperative Aneurysm Study sug­gested that surgical intervention in the risk period for cerebral vasospasm (4 to 14 days) was associated with higher morbidity and mortality rates, However, overall morbidity and mortality was not different for early (less than 3 days) or delayed (more than 14 days) surgery, presumably because of vasospasm in the early surgery group and rebleeding in the late surgery group.

For surgery prior to the onset of vasospasm (less than 3 days after SAH). the benefits of initiating early therapy probably out­weigh the risks of worsening vasospasm. For patients considered for surgery during the period of peak vasospasm (days 4 through 10). the decision to operate may be based on a number of variables, including clinical status. assessment of risk factors for vasospasm, noninvasive indicators of vasospasm, and response to therapeutic intervention. In the setting of existing vasospasm, surgical management should include avoiding intraoperative hypotension or hypovolemia and maintaining aggressive therapy in the postoperative period. Surgery followed by immediate angioplasty has been proposed as an alternative strategy in this group of patients.

Antioxidant and Anti-Inflammatory Agents

Experimental studies have implicated both free-radical-mediated lipid peroxidation and inflammatory responses in the pathogenesis of cerebral vasospasm. Although improvement in vasospasm has been noted in animal models while using both anti­inflammatory agents (e.g., ibuprofen or methylprednisolone) and antioxidants (21-aminosteroids or deferoxamine) clinical trials demonstrating efficacy are limited to date, In a prospective, nonrandomized study, Chyatte et al  showed a reduction in cerebral vasospasm in patients treated with high-dose methylprednisolone. Current prospective, randomized trials are testing the efficacy of Tirilazad, a nonglucocorticoid 21-aminosteroid with antioxidant and iron-chelating properties.

Pathophysiology and Experimental Aspects of Vasospasm

Despite considerable advances in this field, the precise mechanism by which SAH elicits delayed arterial narrowing remains uncertain. In fact, controversy persists as to (1) whether arterial narrowing is a consequence of active vasoconstriction or passive structural changes in vessel; (2) whether large vessel narrowing is integral to the pathologic process; or (3) the specific component of blood that elicits delayed arterial narrowing and the process by which this occurs.

Experimental Models

Much of the uncertainty regarding vasospasm relates to variability among several experimental models. Early work utilizing cerebral artery preparations in vitro did not account for important factors such as chronic exposure to putative spasmogens and the maintenance of intact endothelium. Animal models vary considerably in the time course of arterial narrowing after SAH. In nonprimates, this probably relates to the rapid clearance of subsarachnoid blood following intracisternal injection. This issue has been addressed by strategies such as multiple blood injection or placement of barriers to limit thrombus degradation. It is not clear whether short­term arterial narrowing observed in single-injection small animal models corresponds to chronic vasospasm in humans. A primate model employing direct application of autologous thrombus to the middle cerebral artery probably best simulates human vasospasm.

Blood Exposure to Cerebral Arteries

Several lines of evidence have suggested that the volume and duration of exposure for blood adjacent to cerebral arteries are critical to vasospasm development. Substances released from the perivascular thrombus have ready access to the vessel wall through a porous adventitia. In humans, the volume of subarachnoid blood on CT scan is a strong predictor of vasospasm. The critical period of blood exposure appears to be approximately 72 h; in animal models, the direct removal of thrombus prior to that time eliminates vasospasm. Intracisternal administration of recombinant tPA during the same period effectively lysed subarachnoid thrombus and reduced cerebral vasospasm in a primate model. Current application of this concept to human vasospasm includes the aggressive removal of thrombus at surgery and ongoing trials for tPA. .

Potential Spasmogens in Blood

Approximately 20 agents which elicit vasoconstriction in cerebral arteries in vitro have been identified in whole blood, including catecholamines, serotonin, prostaglandin derivatives, thrombin, and various kinins. However, the majority of these substances are not present in significant concentrations at the time of delayed arterial narrowing, and therefore are not likely participants in chronic vasospasm. One exception are the constituents of erythrocytes, most notably the reduced form of hemoglobin (oxyhemoglobin). The temporal course of red cell lysis (3 to 5 days) in the subarachnoid space corresponds to the onset of clinical vasospasm, and bloody CSF at these time intervals is a potent vasoconstrictor for cerebral arteries. Extracts of lysed erythrocytes and oxyhemoglobin (but not its oxidized form methemoglobin) produce sustained contraction of cerebral arteries in vitro. Chronic in vitro exposure of cerebral arteries to washed erythrocytes, or oxyhemoglobin causes delayed arterial narrowing with angiographic and histologic features of vasospasm; exposure to leukocytes and platelet-rich plasma, erythrocyte membranes, methemoglobin, or bilirubin does not induce vasospasm in these models.

Although these data strongly implicate constituents of the erythrocyte cytosol (primarily oxyhemoglobin) in the pathogenesis of vasospasm, the specific mechanism producing delayed arterial narrowing remains uncertain. In addition to its direct vasoconstrictor effect, oxyhemoglobin generates activated oxygen species (i.e., superoxide anion radical, hydrogen peroxide, and singlet oxygen) through its autoxidation to methemoglobin. In conjunction with free iron, these free radicals propagate the peroxidation of membrane lipids by the Haber-Weiss reaction. Lipid peroxidation may stimulate smooth muscle contraction or mediate structural changes in the vessel wall by cytoxic action. Antioxidant agents reduced vasospasm in several experimental models; clinical trials for their use in humans are forthcoming,

Structural Changes in Vessel Wall

Light and electron microscopic alterations in cerebral artery structure after SAH have been consistently described in human post­mortem and intraoperative specimens, which correlated with angiographic vasospasm and cerebral infarction in the territory of the affected vessel. Smith et al. reported similar postmortem angiopathic changes in 24 of 28 patients autopsied after SAH from ruptured cerebral aneurysm. A number of animal models for SAH have shown that continuous exposure of large cerebral arteries to clotted blood over several days was associated with consistent ultrastructural changes in the vessel wall. Cerebral artery morphology was characterized by alterations in endothelial cell morphology, thickening and discontinuities of the elastica, smooth muscle vacuoles with occasional frank myonecrosis, proliferating "myointimal cells" migrating into the intima, and periadventitial inflammation with loss of perivascular axon

At 2 weeks to 6 months post-SAH, there was regression of the subintimal proliferation, increases in luminal diameter, and deposition of collagen in all three vessel layers. Structural changes in cerebral arteries associated with vasospasm may determine in part the unique physiologic abnormalities seen in this disorder. lmmunhistochemical studies showed loss of contractile protein and increases in vessel wall collagen associated with chronic arterial narrowing. Cerebral arteries exposed to subarachnoid blood for several days were less distensible than controls, and were relatively insensitive to both vasocontricting and vasodilating agents. This suggests that "fibrosis" of the vessel in its contracted state may represent one component of prolonged vasospasm. The effectiveness of angioplasty in reversing vasospasm may be based to this mechanism.

Endothelial Factors

Numerous changes have been demonstrated in cerebral endothelium after exposure to subarachnoid blood, including alterations in prostaglandin metabolism, increased permeability. and diminished secretion of endothelium dependent relaxation factor (EDRF). Endothelin (ET) is a long-lasting. potent vasoconstrictor secreted by vascular endothelial cells. Both plasma and CSF endothelin concentrations were increased in patients after SAH and intracisternal injection of ET produced prolonged narrowing of cerebral arteries in vivo. At present. however. the specific role of endothelial factors in the pathogenesis of vasospasm remains indeterminate.

Inflammation

As described above, inflammatory processes have been implicated in a number of putative mechanisms for cerebral vasospasm. Potential inflammatory mediators include eicosanoids (prostaglandins, leukotrienes), immune complexes (immunoglobin and complement), and cytokines (IL-I ). Inflammatory processes may also be linked to cytotoxic lipid peroxidation, as described above. Although effective in certain experimental models, anti-inflammatory agents have not been widely tested in clinical trials.