CHAPTER 364 Cerebral Vasospasm
Vasospasm in Conditions Other Than Aneurysm Rupture
Vasospasm has been observed in head-injured patients both with and without subarachnoid blood clots in the basal subarachnoid cisterns, although the diagnosis has sometimes relied on transcranial Doppler (TCD) blood velocity measurements, which are difficult to interpret in the setting of brain injury.1,2 The clinical significance of posttraumatic vasospasm is debated, and there is not yet any proven value of either monitoring cerebral arterial diameters or routine treatment with calcium channel blockers in the setting of traumatic brain injury associated with SAH.3,4
Epidemiology of Vasospasm
Angiographic vasospasm is common after rupture of an aneurysm, with an overall incidence of 50% to 90%.5 Still useful but rough estimates are that at least moderate vasospasm in at least one cerebral artery will develop in two thirds of patients with ruptured aneurysms, half of these patients will become symptomatic as a result of ischemia, and a cerebral infarct will develop in about half of these patients. An analysis of 2741 patients entered into SAH trials in the 1990s found that cerebral infarction (all causes) had developed in 26% at 6 weeks, which correlated strongly with poor outcome.6 Cerebral infarction was significantly associated with increasing patient age, worse neurological grade on admission, history of hypertension or diabetes mellitus, larger aneurysm, induced hypertension, fever, and a diagnosis of symptomatic vasospasm. With modern SAH management routines, the combined risk for death and permanent disability from vasospasm alone has shrunk to less than 10%, but it still remains one of the leading causes of preventable poor outcome after rupture of an aneurysm (Figs. 364-1 and 364-2).7–9
Prediction of Vasospasm
A large volume of persistent subarachnoid clot is the most important risk factor predictive of vasospasm after SAH.10,11 The original Fisher grading scale in which clot volume and distribution on admission computed tomography (CT) are related to risk for vasospasm has been modified and in one single-center study found to have greater predictive value for delayed ischemia and prognosis.12 This modified Fisher scale (Table 364-1) scores hemorrhage noted on CT from 0 to 4: 0, no SAH or intraventricular hemorrhage (IVH) (very low risk for vasospasm); 1, focal or diffuse thin layer of SAH, no IVH (low risk for vasospasm); 2, focal or diffuse thin layer of SAH, IVH present (moderate risk for vasospasm); 3, focal or diffuse thick layer of SAH, no IVH (high risk for vasospasm); and 4, focal or diffuse thick layer of SAH, IVH present (highest risk for vasospasm). A slower rate of subarachnoid clot clearance has also been shown to be an independent predictor of vasospasm, although this is not an easy measurement in clinical practice.11
0 | No SAH or IVH: very low risk |
1 | Focal or diffuse thin layer of SAH, no IVH: low risk |
2 | Focal or diffuse thin layer of SAH, IVH present: moderate risk |
3 | Focal or diffuse thick layer of SAH, no IVH: high risk |
4 | Focal or diffuse thick layer of SAH, IVH present: very high risk |
IVH, intraventricular hemorrhage; SAH, subarachnoid hemorrhage.
Other risk factors for the development of vasospasm have been identified, including poor neurological grade or loss of consciousness on admission, cigarette smoking, and preexisting hypertension (Table 364-2).13 Factors studied but found to have either an unclear or unlikely relationship with risk for vasospasm include gender, patient age, and aneurysm location. There has been a suggestion that Japanese people may be more susceptible to vasospasm.14 In addition, a small study suggested that cocaine use is an independent risk factor for vasospasm.15
Spontaneous perimesencephalic or prepontine SAH (or both) unassociated with aneurysm rupture is typically a low-volume hemorrhage that clears quickly with a low risk for the development of vasospasm.16 Thick and persistent nonaneurysmal SAH, however, is still associated with a risk for vasospasm.
There is some evidence that endovascular coiling, as opposed to microsurgical repair of ruptured aneurysms, is associated with a lower risk for the subsequent development of vasospasm,17 although the difference was not large and a rigorous comparison has not yet been made.
Pathogenesis
Smooth Muscle Contraction
Vasospasm is prolonged cerebral arterial constriction caused by vascular smooth muscle contraction. The hemoglobin released from subarachnoid blood clots triggers the entry and release of calcium and subsequent activation of calcium/calmodulin-dependent myosin light-chain kinase, which in turn leads to phosphorylation of the myosin light chain and induces actin and myosin cross-linkage and mechanical shortening (smooth muscle contraction). Such contraction requires adenosine triphosphate and calcium, and vascular smooth muscle relies more on extracellular than intracellular calcium stores, which enter through voltage-gated and receptor-operated calcium channels. Although myofilament activation depends on calcium and high-energy phosphates, chronic vasospasm, which ensues days later and lasts up to several weeks, does not. The contractile proteins protein kinase C, Rho kinase, and protein tyrosine kinase and their corresponding signal transduction pathways have been implicated in vasospasm models when their activation shifts the contractile mechanism toward increased shortening in the absence of high intracellular calcium levels.18,19 This contiguous and second-phase “chronic” vasospasm is less reversible with pharmacologic vasodilators both in animal models20 and in humans undergoing intra-arterial, pharmacologic vasodilation treatment.
Sustained vasoconstriction is associated not only with functional impairment of the vessel but also with ultrastructural damage to the vascular wall layers, including vacuolization of endothelial cells and loss of tight junctions, breakage of the internal elastic lamina, and patchy myonecrosis in the tunica media.21
Endothelial Injury, Nitric Oxide, and Endothelin-1
Auto-oxidation of the oxyhemoglobin contained in blood clots encasing cerebral arteries produces methemoglobin and superoxide anion radical, which in turn lead to lipid peroxidation.22,23 Harmful hydroxyl radicals and lipid peroxides permeate the vessel wall and injure endothelial and smooth muscle cells.24,25 Damage to the endothelium in particular is thought to play a key role in the establishment of vasospasm, either through the loss of endothelial nitric oxide (NO) synthesis, an important vasodilator and regulator of vascular tone, or through the overproduction of endothelin, a powerful vasoconstrictor peptide.26 These two endothelial-derived substances and the possible imbalance in their production after SAH are at the center of experimental vasospasm research at the present time.
Decreased availability of the simple molecule NO may contribute to the development of vasospasm in the following ways: (1) endothelial NO synthase (eNOS) dysfunction in vasospastic vessels, (2) NO scavenging by oxyhemoglobin; (3) reversal of vasospasm by NO donors, (4) disappearance of neuronal NO synthase (nNOS) activity from the adventitia of vasospastic vessels, and (5) decreased cerebrospinal fluid (CSF) nitrite levels along with increased levels of asymmetric dimethyl-L-arginine, the endogenous inhibitor of NO synthase.27–35
Endothelin-1 (ET-1) is the predominant isoform of endothelin and has the greatest role in vasoconstriction. ET-1 is a 21–amino acid cleavage product of a 212–amino acid peptide precursor, the final step mediated by endothelin-converting enzyme. It is released on the abluminal side of the tunica media, acts on neighboring vascular smooth muscle ETA receptors, and causes profound and sustained vasoconstriction.36 ET-1 levels have been found to be elevated in the CSF of patients in whom vasospasm and brain ischemia develop,37–39 and either inhibition of ET-1 production or antagonism of its effect has prevented vasospasm in animal models.36
Inflammation, Vessel Remodeling, and Vasospasm
Although it has become clear that vasospasm is not a type of vasculitis, there is evidence that inflammatory mechanisms are activated after SAH and may therefore be involved in the development of vasospasm, possibly by contributing to vasocontraction or modifying the vessel wall extracellular matrix and smooth muscle cell phenotype—a process known as vascular “remodeling.”40 Inflammatory cytokines,41,42 intercellular adhesion molecules (ICAMs),43,44 and genetic upregulation of inflammatory, proliferative, and extracellular matrix–regulating genes45–47 have been examined. In humans, elevations of plasma complement C3a and soluble ICAM-1 have been associated with poor outcome and the development of vasospasm, respectively,48,49 and increased intrathecal levels of the cytokine interleukin-6 predicted vasospasm in another small study.50
Although the precise pathogenesis of vasospasm is still the subject of investigation, a prolonged, biphasic vasoconstrictive process in which the second, chronic phase is mechanistically distinct from the first seems most consistent with what has been seen experimentally and in humans. It remains possible that additional processes, including inflammation, may contribute to the pathogenesis of this condition (Table 364-3).
Clinical Features and Investigation
Symptoms, Signs, and Differential Diagnosis
Coincident with its delayed onset, symptoms of ischemia resulting from cerebral vasospasm most commonly appear 1 week after aneurysm rupture but often occur later; accordingly, it is important to remain vigilant for this complication for at least 2 weeks after SAH. Regular and careful bedside examination remains the simplest and most effective means of detecting early ischemia in awake, examinable patients; one should concentrate on subtle findings such as diminished attention, changes in verbal output, or a slight but new pronator drift of the upper extremity. Symptomatic vasospasm usually has a gradual onset, sometimes heralded by increased headache and either agitation or somnolence—a change in patient behavior. It then follows a progressive course if untreated. A smaller group of patients will experience precipitous deterioration.51 Signs of symptomatic vasospasm are referable to the territory that has become ischemic and are most easily distinguished when they lateralize to a middle cerebral artery (MCA) territory with monoparesis or hemiparesis and, when the dominant hemisphere is affected, aphasia. Anterior cerebral artery vasospasm can be marked by leg weakness, but because it is often bilateral in distribution, confusion, drowsiness, poverty of speech, and eventually abulia are characteristic signs. Vertebrobasilar vasospasm can also cause a more generalized deterioration, with a reduced level of consciousness being an early sign.
Delayed neurological deterioration after aneurysmal SAH has a number of causes, including increased edema surrounding intracerebral hematomas, contusions, or infarcts; rebleeding of the aneurysm or aneurysmal remnant; hydrocephalus; sepsis, including meningitis and ventriculitis; hyponatremia; hypoxia; and hypotension (Table 364-4). One or more of these conditions can magnify even a focal neurological deficit and therefore easily be mistaken for vasospasm, which has a tendency to be overdiagnosed in the setting of SAH. Conversely, it is recognized that there is a subgroup of patients who suffer a delayed and often global decline in neurological status for which no single underlying cause can be identified. “Cortical spreading depression” has been one proposed mechanism at work in these patients.19
Diagnosis
Diagnosis of symptomatic vasospasm requires that the other causes of delayed worsening listed earlier be ruled out with CT and appropriate laboratory investigations. If vasospasm remains the most likely cause of deterioration and treatment by induced hypertension reverses the deficit, the diagnosis can safely be assumed without further testing. Failure to respond in this scenario, as well as evaluation of comatose patients, requires additional testing (Table 364-5).
CBF, cerebral blood flow; CT, computed tomography; MRI, magnetic resonance imaging; SPECT, single-photon emission computed tomography; TCD, transcranial Doppler.
Transcranial Doppler
TCD works on the principle that as an artery narrows, blood flow velocity within it increases. Because examinations are noninvasive, are carried out at the bedside, and can easily be performed on a daily basis, TCD is frequently used to monitor patients with SAH for increases in intracranial blood flow velocity suggestive of incipient vasospasm. There is good general correlation between TCD velocities and vasospasm, with velocities in the MCA greater than 120 cm/sec indicative of some degree of vasospasm and those greater than 200 cm/sec consistent with severe vasospasm. Because other factors can influence velocity, including blood pressure and overall cerebral blood flow (CBF), distinguishing vasospastic from hyperemic increases in blood velocity has been reported to be facilitated by measuring cervical internal carotid artery (ICA) velocity in addition to intracranial blood velocity.52 A “Lindegaard ratio” of VMCA/VICA greater than 3 is consistent with vasospasm (hyperemia is associated with increased velocity in both the MCA and ICA, so the ratio is the same). A similar velocity ratio between the basilar artery and the extracranial vertebral artery has been proposed to improve the sensitivity and specificity of detecting basilar artery vasospasm.53
The clinical utility of TCD is best when MCA values are clearly low (<120 cm/sec) or very high (>200 cm/sec), and the respective negative and positive predictive values for significant angiographic vasospasm in the MCA trunk are close to 90%.54–56 For values in the intermediate range, additional maneuvers may improve its accuracy in detecting vasospasm, such as testing for hyperemic autoregulatory responses to transient, manual carotid compression57,58 or combining the TCD information with CBF measurements.59 TCD does not reliably detect vasospasm in more peripheral branches,60 which may account for its failure to always correlate with perfusion deficits detected on CBF studies.61 Regular surveillance of SAH patients with TCD ultrasonography by skilled technicians is considered helpful in many neurosurgical units, but the results must be considered in the context of the individual patient along with all other information available.
Cerebral Blood Flow and Perfusion
Single-photon emission CT,62 quantitative stable xenon–enhanced CT,59 and positron emission tomography61 can be used to detect cerebral ischemia after SAH, but none are practical investigations that can be easily performed or repeated in critically ill patients. Magnetic resonance imaging to look for either perfusion deficits63 or ischemia on diffusion-weighted images64,65 has the same limitation.
Perfusion CT has been used in patients with SAH66,67 and with greater experience will probably find more widespread use and validation for predicting or diagnosing vasospasm.
Thermal diffusion flowmetry has been used as bedside monitoring in SAH patients to detect vasospasm causing significant reductions in CBF.68 Provided that the white matter location into which the thermal diffusion microprobe is inserted (usually the frontal lobe) is representative of the territory at risk, continuous measurement of CBF values could be particularly useful in patients with high-grade SAH who cannot be assessed neurologically.
Microdialysis Monitoring
Cerebral microdialysis catheters allow continuous bedside measurement of extracellular concentrations of glutamate, lactate, pyruvate, glucose, and glycerol in brain tissue, thereby screening for excitotoxic cell injury characterized by elevations in lactate with respect to glucose and pyruvate levels and an increase in the glycerol concentration. This somewhat demanding monitoring technique provides chemical information about a very small region of the brain, but it has been used, along with cerebral oximetry, to detect ischemia in patients with SAH.69–71
Vascular Imaging
The most practical method of imaging large and medium-sized cerebral arteries is digital subtraction angiography or high-definition CT angiography.72 Catheter-based angiography is best in patients in whom balloon angioplasty is being considered and can accompany the procedure.
Angiographic vasospasm is a concentric narrowing that can be focal, segmental, or diffuse. It is commonly graded as mild (<25%), moderate (25% to 50%), or severe (>50%) in comparison to baseline, prevasospasm imaging. “Early” vasospasm on admission angiography and within 48 hours of aneurysm rupture has been described in a small percentage of SAH patients, although without baseline angiography it is not clear how this diagnosis can be made accurately in all cases.73 Its detection has been associated with the development of cerebral infarction and poor outcome.
Prevention of Vasospasm and Cerebral Protection
General Measures: Fluid Management and Medical Treatment
Patients have a tendency toward volume contraction in the acute stage of SAH,74 and hypovolemia should be carefully avoided. It is not clear that a deliberate attempt to induce hypervolemia with volume expansion therapy is beneficial in terms of prevention of vasospasm or ischemia or even possible in patients with normal renal function.75–78 Patients should be hydrated with at least 3 L of isotonic fluids daily, and there is some evidence that additional colloid infusions (such as albumin) are beneficial.79 Some SAH patients experience excessive natriuresis and are susceptible to the development of hyponatremia during this time (usually related to elevations in brain natriuretic peptide), an electrolyte disturbance that may increase the risk for vasospasm.80 It has been reported that fludrocortisone administration (0.3 mg/day) may help avert this complication.81
Blood transfusion has been associated with a higher likelihood of both vasospasm and poor outcome in SAH patients,12,82 thus suggesting that anemia is a marker for other factors that contribute to SAH-associated morbidity. The optimal hemoglobin concentration in patients with SAH in terms of hematocrit, blood viscosity, and oxygen delivery to the brain is not known with certainty, but is generally accepted to be higher than 9 g/dL.