Vasospasm after aneurysmal subarachnoid hemorrhage

Introduction
Vasospasm after subarachnoid hemorrhage was first described in by Ecker and Riemenschneider in 1951. Subsequent work showed that it tended to have onset at post-bleed day (PBD) 3, reach its maximum at PBD 6-8, and resolve by day 12, but that these were not absolutes.

Vasospasm leads to both large and small vessel vasoconstriction, resulting in ischemia (often referred to as delayed cerebral ischemia, or DCI), and when severe, infarction.

Preclinical data
The potent vasoconstrictor endothelin-1 (ET-1) was first isolated from porcine aorta endothelial cells in 1988 and is expressed in cerebral endothelial cells. Intracisternal injection of endothelin into dogs, cats, rabbits, and monkeys induces a dose-dependent degree of large vessel vasoconstriction,      and similar results have been seen with animal and tissue mounted human specimens of cerebral vasculature bathed in endothelin. This vasoconstriction can be prolonged, lasting for 90 minutes in one study of cats and up to 3 days in a study of dogs. Endothelin's action is likely to be on the adventitial side of vessels, as intravascular injection in one study of dogs and cats did not lead to vasospasm but intracisternal injection did, but studies in rabbits showed that intravascular injection also caused vasoconstriction. Usually it is produced by the endothelium and released predominantly on the abluminal side to affect the smooth muscle cells in the arterial wall.

ET-1 is produced by enzymatic cleavage from big ET-1, via an endothelin-converting enzyme (ECE), which is present in the arterial wall itself. Big ET-1 does not have any direct activity on ET receptors. There are two subtypes of ECE, ECE-1 which is membrane-bound, and ECE-2 which is intracellular. In humans there are at least 3 and 4 different isoforms of ECE-1: ECE-1a, ECE-1b, ECE-1c, and ECE-1d. ECE-1a and ECE-1c are located on the membrane, while ECE-1b is intracellular and closely associated with the Golgi network. All three of these appear to have similar kinetics. ECE-1d is located both intracellular and extracellularly.

with evidence localizing ECE-1α and probably others to the cerebral vasculature endothelium. . However, there also appears to be ECE in the smooth muscle of the arterial wall as well, but probably to a lesser extent. . In cell culture, hemolysate may activate ECE to produce more ET-1 from big ET-1. Big ET-1 levels in the CSF may be lower after ruptured aneurysm clipping, but the significance of this finding remains unclear, as this was not clearly associated with a change in the incidence of vasospasm or delayed cerebral ischemia.

There are two types of receptors, ETA and ETB. Human pial vessels in two studies of vessels taken during neurosurgical procedures (non-subarachnoid hemorrhage patients) were only found to have ETA receptors and not ETB receptors. Similarly, cultures of cerebral vascular smooth muscle cells had mRNA coding both types of receptors, but only seemed to express ETA receptor protein. ETA is located in smooth muscle cells as well as endothelial cells and appears to mediate contraction in response to ET-1. The role of ETB remains unclear. Some have suggested that ETB receptors may be located on the parasympathetic nerves innervating the cerebral vasculature. Other data suggests significant ETB receptor presence in the brain. ETB receptors are divided into ETB1 and ETB2 receptors with variable effects. ETB1 receptors are generally found on endothelial cells and mediate vasodilation through generation of nitric oxide and/or prostacyclin. and possibly also adrenomedullin. ETB2 receptors are located on smooth muscle and mediate vasoconstriction. However, this effect is not seen in all vessels, due to variable expression of receptors.

ET-1 may also increase its own release through a positive feedback mechanism, possibly via ETB1 receptors. Therefore ET-1 binding to ETB1 may cause more ET-1 release. However, this remains unproven, as ETB receptor agonist (sarafotoxin S6c) had no effect on contractile action in one study and variable effects on relaxation in another study.

Studies suggest that ETB stimulation may initially cause vasodilation (possibly via stimulation of ETB1 receptors), but that with repeated or prolonged application this response goes away and the vasoconstrictor response (possibly through ETB2 receptors) develops.

Vasoconstriction in response to endothelin is due to multiple effects, including: Endothelin may also serve as a growth factor for smooth musclereceptors and activating protein kinase C. It also may increase angiogenesis by via ETA receptor eventually causing increased expression of VEGF. Whether this is due to associated ischemic or direct effects of endothelin itself remains un certain. In the systemic circulation, it may increase adrenergic tone itself.
 * 1) An increase in cytosolic calcium in smooth muscle cells  via influx of calcium ions through L-type (dihydropyridine) calcium channels, which is a G-protein coupled mechanism.   While its effects may not be directly inhibited by nifedipine, diltiazem, or verapamil they do appear to be affected by nicardipine, papaverine, and other calcium channel blockers acting on the L-type channels. Note that results with nicardipine have been variable.
 * 2) An increase in myosin light chain phosphorylation in response to a given calcium level.
 * 3) Activation of phospholipase C (to convert PIP2 to DAG and IP3) via a G-protein coupled mechanism.  However, it does not necessary lead to increased protein kinase C levels.
 * 4) Inhibition of adenylate cyclase
 * 5) Activation of mitogen-activated protein kinase (MAPK)
 * 6) Activation of Src tyrosine kinase
 * 7) Activation of phosphotidyinositol-3-kinase
 * 8) Activation Janus tyrosine kinase-2
 * 9) Activation of protein tyrosine kinase (PTK)
 * 10) Activation of protein kinase C (PKC)   . However this enzyme did not appear to be upregulated after SAH in a rabbit model.
 * 11) Other mechanisms: Activating Na/H exchange and phospholipase A2 (to generate arachidonic acid), and inhibiting adenylate cyclase and guanylate cyclase.
 * 12) Rho kinase may be involved but does not appear to be functionally upregulated.

Endothelin may also cause ischemia through alternative mechanisms, such as inhibition of Na/K ATPase function, which may be due to stimulation of the ETB receptor, and may causing cortical spreading depolarization. . An increase in Na/K ATPase function in response to endothelin can also be seen in tissues, and may be mediated by the ETA receptor.

Iloprost also appears to vasodilate vessels that were constricted by ET-1, but whether this serves as a direct antagonist to ET-1 or simply a vasodilating medication remains unclear. However, ET-1 may attenuate the vasodilatory response of iloprost. Similarly, nitric oxide may modulate the contractile response of ET-1.

Not all vasodilating substances work, however. Acetylcholine did not help to relax vasospastic vessels in a rabbit model.

The following substances may increased ET-1 production or release in the cerebral vasculature, possibly via PKC:
 * 5-HT
 * ET-1 itself (possibly via ETB receptors)
 * Lysophosphatidic acid (LPA)
 * Oxyhemoglobin
 * Phorbol 12-myristate 13-acetate
 * Thromboxane

The following substances may decrease ET-1 production or release in the cerebral vasculature.


 * Nitric oxide (NO)

ET-1 may also be released from activated mononuclear leukocytes in SAH patients.

In vitro, oxyhemoglobin increased the production of ET-1 by bovine endothelial cells in culture,   and this may be mediated by protein kinase C (PKC). Although it was to a much lesser extent, oxyhemoglobin also increased the production of ET-1 by rat thoracic aorta smooth muscle cells in culture, which seemed to be mediated by both PKC and cyclic AMP (cAMP) pathways. In isolated cerebral vessels from rabbits, extraluminal oxyhemoglobin potentiated the effects of intraluminal endothelin administration. This is a plausible mechanism for vasospasm, that oxyhemoglobin and its breakdown products may cause an increase in ET-1 production leading to vasospasm. Moreover, perivascular oxyhemoglobin and deoxyhemoglobin levels (obtained by microdialysis) peak at PBD 7 after SAH in monkeys and are dramatically higher than controls.

Elevated levels of ET-1 and big ET-1 are noted in cerebral vessels, plasma, and CSF of dogs and rats and coincide with the timing of vasospasm,     with the vasospastic effects partially reduced by administration of a monoclonal antibody against ET-1. In contrast to these results, Pluta et al. did not find any significant increase in ET-1 microdialysis samples in a monkey model, and did not find any significant changes in CSF or plasma levels of ET-1 in the setting of vasospasm. They did, however, find changes in ET-1 levels in ischemia, suggesting that these levels may be increased by ischemia and not as a cause of cerebral vasospasm. They also noted that ET-1 seemed to be released by astrocytes and not endothelial cells.

ET-1 appears to amplify the contractile effects of norepinephrine and serotonin in human vessels in vitro. It also seems to have synergistic responses with other vasoconstrictors, such as 5-HT and thromboxane A2.

In a monkey SAH model, there was a significant increase in levels of prepro-ET-1 mRNA in the underlying brain parenchyma, but there was no change in prepro-E T-1 or ET-1 mRNA in vasculature or ET-1 CSF concentration during vasospasm. However, vasospastic cerebral arteries showed a significant increase in expression of the ETB receptor mRNA, with a trend towards increases in numbers of the ETB receptor itself. Thus the mechanism here is likely to involve more than simple upregulation of ET-1. ETA receptors may also be upregulated during SAH, as studies have shown an increase in ETA mRNA expression in a dog SAH model. Other studies have suggested that there may be a shift in expression of ETA to ETB after SAH. In support of a change in receptors in response to SAH, vessels isolated from rats with SAH two days prior had stronger contractions in response to ET-1 than vessels isolated from control animals. Both ETA and ETB receptors may play a role, as one study in rabbits suggested better reversal of vasospasm using an antagonist to both receptors instead of an ETA specific one, and selective ETB receptor antagonists also seem to improve vasospasm.

Endothelins may also cause degenerative changes in the vascular wall, as studied in rabbits and rats that may lead to long-lasting constriction of the vessel. It also may open the blood-brain barrier via ETA receptors.

ET-1 may also increase COX-2 expression and prostaglandin E2 production by ETB more than ETA receptors in macrophages.

ET-1 levels rise after SAH until day 4 in a canine model but then declined by day 7, while prolonged enhancement of PKC was observed after. Therefore ET-1 may initiate vasospasm but is not likely to be responsible for maintaining the PKC activation at later stages.

Human data
In humans with SAH, several studies have shown peaks in CSF endothelin levels, and more specifically ET-1 levels and CSF big ET-1 (an ET-1 precursor) that correspond with presence and the timing of vasospasm. Peaks of CSF ET-3 have been more variable, with some studies showing an increase in CSF levels and others showing no increase in CSF ET-3. Other studies, however, have shown no increase in CSF concentrations of ET-1 or big-ET-1    or ET-3. Of note, endothelin concentrations in CSF may be increased in older patients relative to younger ones. In a study of 20 patients with SAH, CSF ET-1 levels correlated with the degree of angiographic vasospasm in patients without poor GCS. ET-1 levels were increased in the clinical vasospasm and low-GCS groups, moreso than with the angiographic vasospasm and no vasospasm groups. CSF levels were low prior to symptom onset, and remained low in those with angiographic but not clinical spasm. Therefore ET-1 appears to rise in the setting of neuronal damage, and not necessarily due to vasospasm. CSF ET-1 levels did not predict vasospasm in another study. However, given local brain production, there may be higher concentrations in the jugular vein, and researchers have suggested that the difference between arterial and jugular vein ET-1 levels may be predictive of vasospasm, although no definitive relationship was seen.

Results for plasma levels have been variable. Several studies have shown no changes in plasma levels with vasospasm. However, other studies have suggested variable increases in plasma levels of ET.

Importantly, one study showed elevated levels of CSF ET-3 in epilepsy patients, so these findings are nonspecific.

Relevant drugs
Actinomycin: This drug inhibits RNA synthesis, and prevented development of vasospasm in a dog SAH model. One hypothesis is that it might prevent ET-1 formation by preventing transcription of its mRNA. Low-doses of actinomycin actually exacerbated vasospasm in one dog model, which high dose prevented vasospasm.

Aminoglycosides: These appear to have an inhibitory effect on ET-1 associated contraction in vitro, possibly due to inhitibition of PKC.

BQ-123: In 1993, using a rat model of vasospasm after SAH, animals that received an intracisternal injection of BQ-123 (a potent ETA receptor antagonist) prevented a decrease in cerebral blood flow. In monkey model, animals that received intracisternal injection of BQ-123 developed less vasospasm than controls. In isolated canine basilar arteries, BA-123 prevented ET-1 contractions. Results in dog models of SAH have been more variable, with vasospasm prevented in some studies but not in another. In rats, rabbits, and piglets given intracisternal ET-1, administration of BQ-123 relaxes basilar artery constriction. However, the drug does not cross the blood-brain barrier, which limits its utility.

BQ-485: This drug is similar in terms of effects to BQ-123 but is easier to synthesize. With subcutaneous administration, it prevented vasospasm in a dog model of SAH.

BQ-610: This drug is similarly an ETA antagonist which partially blocked the effects of vasospasm in basilar arteries but not completely, suggesting that ETA receptors are not the only cause of endothelin-induced vasospasm.

BQ-788: This is a selective ETB antagonist that relieved vasospasm in a rabbit SAH model. However, it had no effect on relieving vasospasm in human cerebral vessels removed during surgery.

Bosentan (RO 47-0203): This is a competitive antagonist of both ETA and ETB receptors. A study using a monkey vasospasm model showed no benefit of this agent when injected intracisternally, in contrast to monkeys given BQ-123. It was hypothesized that this may have been caused by insufficient CSF levels of bosentan. In contrast, intravenous or oral bosentan in rabbit, dog, and cat models of SAH improved vasospasm. Bosentan also prevented vasoconstriction in human cerebral vessels exposed to ET-1 in vitro. In a small clinical trial in 1997 (n=24) published only in abstract form, bosentan decreased transcranial doppler velocities in patients after SAH-associated vasospasm.

Captopril: This is a potential inhibitor of the ET-converting enzyme that converts big ET-1 to ET-1. In rabbits, big ET-1 was injected intracisternally, and captopril pretreatment prevented vasoconstriction.

CGS 26303: This is an inhibitor of the ET-converting enzyme that converts big ET-1 to ET-1. In rabbits, intravenous or topical CGS 26303 blocked vasoconstiction in response to big ET-1 but not ET-1 itself. In rabbits with experimental SAH, intraperitoneal or intravenous administration of CGS 26303 significantly attenuated the delayed vasospastic response. It also seemed to work better with continuous infusion or oral administration. . In endothelial cell culture, it attenuated endothelial cell injury induced by hemolysate.

[d-Val22]big ET-1[16–38]: This is an inhibitor of the ET-converting enzyme that converts big ET-1 to ET-1. In rabbits, big ET-1 was injected intracisternally, and [d-Val22]big ET-1[16–38] pretreatment prevented vasoconstriction. In isolated rat basilar arteries, however, it actually lead to increased contractions in one study, for unclear reasons.

Doxorubicin: Low-doses of this drug (in comparison to the high-doses required of actinomycin) prevented vasospasm in dog and rat models of SAH-induced vasospasm.

ETant / cyclo(D-Asp-L-Pro-D-Val-L-Leu-D-Trp): This is a peptide ETA receptor antagonist that when injected intracisternally reversed cerebral vasoconstriction after ET-1 administration and after SAH in a rabbit model.

Gingko biloba: Extract of Gingko biloba attenuated the rise in plasma and CSF ET-1 in SAH rat and dog models, and decreased the associated vasospasm.

FR139317: In 1993, using a dog model of vasospasm after subarachnoid hemorrhage, animals that received an intracisternal injection of FR139317 (an ETA receptor antagonist) had much larger diameter vessels compared with vasospastic controls, suggesting that endothelin antagonism may prevent vasospasm. This also prevented vasoconstriction in human cerebral vessels exposed to ET-1 in vitro.

IRL 1620: This is a selective ETB agonist, and caused vasodilation in rat basilar arteries.

L-arginine: Injection of this amino acid in a rat SAH model attenuated the rise in plasma levels of ET-1. This is likely because it serves to produce NO, which may downregulate the production of ET-1.

LU-208075: This is a selective ETA receptor antagonist. In isolated rat basilar artery, it attenuated vasospasm in response to ET-1 or big-ET-1 administration. It even led to some relaxation, suggesting that the remaining ET-1 was now able to bind ETB receptors to cause vasodilation.

PD145065: This is an antagonist of both ETA and ETB and showed great effect on reversal of vasospasm induced in a rabbit SAH model, better than an ETA antagonist alone.

PD155080: This is an ETA selective antagonist, and it improved vasospasm in a rabbit model of SAH.

PD156707: This is an ETA selective antagonist and it reversed ET-1 established constriction in human cerebral vessels in vitro. It also prevented vasospasm when give by continuous infusion in a dog SAH model.

Phosphoramidon: This drug inhibits the ET-converting enzyme that converts big ET-1 to ET-1. Big ET-1 was injected into dogs in the presence or absence of phosphoramidon, which caused vasospasm and elevations in ET-1 levels. The dogs that received phosphoramidon had lower ET-1 levels and were protected from vasospasm. In a dog model of SAH and vasospasm, injection of phosphoramidon led to increased levels of CSF ET-1 on PBD2, but the levels fell on PBD 7. Nevertheless, the CSF levels were significantly higher even on PBD 7 than they were in controls that did not receive phosphoramidon. This was hypothesized to possibly be due to phosphoramidon inhibiting an enzyme that degrades ET-1. Regardless, vasospasm was decreased in the dogs that received phosphoramidon. However in other experiments using a dog model, animals that received intracisternal injection of phosphoramidon did not have different degrees of vasospasm compared with controls. Effects in rat SAH models have been variable. Effects in human vessels in vitro show that it inhibits contraction otherwise induced by big ET-1.

Prepro-ET-1 mRNA antisense oligoDNA: In a rat model of vasospasm by exposing the basilar artery to oxyhemoglobin, this drug inhibited the development of vasospasm. This also had effects in a dog SAH model and reduced ET-1 expression, but the effects on vasospasm were more mild.

RES-701-1: This is a selective ETB1 receptor antagonist that prevented vasospasm in a rabbit SAH model.

RO-46-2005: This is an ETA and ETB antagonist, and in a rat model of SAH, IV administration reduced vasoconstriction from vasospasm.

RO-61-1790: This is a predominantly ETA antagonist (has 1000-fold selectivity compared with ETB) that was formulated to have high water solubility. In a dog model of SAH, it prevented and reversed vasospasm in a dose dependent manner.

S-0139: This is predominantly ETA antagonist (has 1000-fold selectivity compared with ETB) and reversed the effects of ET-1 intracisternal administration in a dog model. When administered IV or intracisternally, it also reduced vasospasm in a dog SAH model.

SB 209670: This is a selective nonpeptide ETA and ETB antagonist. Intracisternal administration of SB 209670 in a dog model of SAH reduced the development of vasospasm.

Sitaxentan / TBC 11251: This is a nonpeptide selective ETA antagonist (6400-fold selectivity compared with the ETB receptor) that attenuated spasm when given in a preventive manner in a rabbit SAH model. However, at higher doses it did not attenuate vasospasm, giving it a lack of dose-dependency in this disease. This was hypothesized to be due to blockade of ETB receptors at higher doses.

TA-0201: This is a nonpeptide selective ETA antagonist. In an isolated canine basilar artery model without endothelium, IV treatment with TA-0201 inhibited ET-1 induced vasospasm. In a canine SAH model, it prevented vasospasm as well.

TAK-044: This is a selective nonpeptide ETA and ETB antagonist that prevents vasospasm after bathing dog cerebral vessels in ET-1.

It was administered in a randomized controlled trial to 420 patients with SAH in Europe within 96 hours of SAH onset. There was a trend towards benefit in the primary outcome of delayed-cerebral ischemia (29.5% in treatment group, 36.6% in plaebo group, RR 0.8, 95% CI 0.61-1.06). There were no clear differences in the incidence of delayed neurological deterioration within 10 days, or outcomes at 3 months as evaluated by GOS. The treatment groups had more hypotension, headache, and pneumonia compared with the placebo group.

Relevant drugs
Sodium nitroprusside reverses vasospasm in a rabbit model of SAH.

Clinical presentation

 * WFNS score
 * WFNS score of 4-5 on admission is associated with a higher likelihood of angiographic vasospasm. but not DCI-associated deterioration or cerebral infarction in one small study of 196 patients.

Imaging markers
==== Modified Fisher scale ====
 * The modified Fisher Score (mFS) was developed initially by Claassen et al., who specifically used the parameter of bilateral IVH (not unilateral) in the score. However, in Frontera et al. they subsequently described the utility of the scale using any IVH in the score. This scale is more ordinal than the original Fisher score (although mFS 2-3 confer similar rates of DCI), and is measurable on modern imaging. It should be used in lieu of the original Fisher score.
 * Admission mFS of 3-4 on admission is associated with angiographic vasospasm (aOR 1.440, 95% CI 1.006-2.063), DCI-caused clinical deterioration (aOR 1.813, 95% CI 1.143-2.874), and cerebral infarction (aOR 2.027, 95% DCI 1.346-3.053) in one small study of 196 patients.

Lab studies

 * Alkaline phosphatase
 * Higher alkaline phosphatase values were associated with angiographic vasospasm (aOR 1.019, 95% CI 1.002-1.036) and DCI-caused clinical deterioration (aOR 1.019, 95% CI 1.142-3.874) in one small study of 196 patients, but this was not associated with cerebral infarction. However, it was associated with unfavorable functional outcome of mRS 3-6 (aOR 1.083, 95% CI 1.041-1.127).

Monitoring
Close monitoring for vasospasm with transcranial doppler (TCD), CT angiography (CTA) or digital subtraction angiography (DSA) is important, and is a Neurocritical Care Society Clinical Performance Measure.

UK/Netherlands/Eire TAK-044 Subarachnoid Hemorrhage Phase II Trial
This trial used a nonselective nonpeptide ETA and ETB antagonist. It was administered in a randomized controlled trial to 420 patients with SAH in Europe within 96 hours of SAH onset. There was a trend towards benefit in the primary outcome of delayed-cerebral ischemia (29.5% in treatment group, 36.6% in plaebo group, RR 0.8, 95% CI 0.61-1.06). There were no clear differences in the incidence of delayed neurological deterioration within 10 days, or outcomes at 3 months as evaluated by GOS. The treatment groups had more hypotension, headache, and pneumonia compared with the placebo group. The study was not adequately powered to detect a difference in outcome.

Clazosentan
Several experimental studies suggested that

Nitric oxide therapies
One hypothesis of the pathophysiology of vasospasm involves oxyhemoglobin and its degradation products scavenging nitric oxide (NO), preventing the vasodilatory activity of NO and leading to vasospasm.

Molsidomine
The drug molsidomine is a NO donor that can be administered intravenously. Unlike drugs such as nitroprusside, it has no risk of cyanide toxicity. In a non-randomized study of 79 patients, 29 received molsidomine (20-40 mg per day, titrated from 3.5-10 mg/hr) and nimodipine, and were compared with 45 patients who received nimodipine only and standard therapy. Molsidopine was well-tolerated, mild hypotension requiring low-dose pressors in some cases, but otherwise no major issues. Few vasopasm-related infarcts developed in the molsidomine-treated group (14%) compared with the standard group who did not develop vasospasm (35%) and those who did develop vasospasm (60%). At 3-month follow-up, the median mRS was 1 (IQR 0-3) in the molsidomine group, 5 (IQR 2-6) in the standard group with vasospasm, and 4 (IQR 2-5) in the standard group without vasospasm. The improved outcomes in the molsidomine group were significant (molsidomine vs. standard with vasospasm with p=0.0011; molsidomine vs. standard without vasospasm p=0.0166 after Bonferroni correction).

Sodium nitroprusside
Others have used nitroprusside administered intraventricularly. This was first described in abstract form in 1998 in eight patients with SAH and refractory vasospasm, and observed that as sole initial treatment it showed dramatic improvement in two patients, while all the remaining patients did well, but some received additional therapies such as angioplasty which limited conclusions. Several subsequent uncontrolled case series with small numbers of patients also suggested safety and potential benefit with reversal of angiographic vasospasm in many cases. One study co-administered sodium nitroprosside with thiosulfate. Thiosulfate is used by the enzyme rhodanase to convert cyanide to thiocyanate. Thus thiosulfate was added to prevent cyanide toxicity. In one protocol, patients were pretreated with ondansetron 32 mg IV and dexamethasone 10 mg IV 15 minutes prior to treatment. The study used a combination of sodium nitroprusside (4 mg) and thiosulfate (10 mg) mixed together in 1 mL NS, with 1 mL infused over 1-2 minutes, repeated every 5 minutes to a total of 10 mL (40 mg), and preceded each time by the withdrawal of 5-10 mL of CSF.