Nature-Delayed Neurological Deterioration After Subarachnoid Haemorrhage

REVIEWS
Delayed neurological deterioration after

subarachnoid haemorrhage
R. Loch Macdonald
Abstract | Subarachnoid haemorrhage (SAH) causes early brain injury (EBI) that is mediated by effects of
transient cerebral ischaemia during bleeding plus effects of the subarachnoid blood. Secondary effects
of SAH include increased intracranial pressure, destruction of brain tissue by intracerebral haemorrhage,
brain shift, and herniation, all of which contribute to pathology. Many patients survive these phenomena,
but deteriorate days later from delayed cerebral ischaemia (DCI), which causes poor outcome or death
in up to 30% of patients with SAH. DCI is thought to be caused by the combined effects of angiographic
vasospasm, arteriolar constriction and thrombosis, cortical spreading ischaemia, and processes triggered
by EBI. Treatment for DCI includes prophylactic administration of nimodipine, and current neurointensive
care. Prompt recognition of DCI and immediate treatment by means of induced hypertension and balloon or
pharmacological angioplasty are considered important by many physicians, although the evidence to support
such approaches is limited. This Review summarizes the pathophysiology of DCI after SAH and discusses
established treatments for this condition. Novel strategies—including drugs such as statins, sodium nitrite,
albumin, dantrolene, cilostazol, and intracranial delivery of nimodipine or magnesium—are also discussed.
Macdonald, R. L. Nat. Rev. Neurol. 10, 44–58 (2014); published online 10 December 2013; corrected online 17 December 2013;
doi:10.1038/nrneurol.2013.246
Introduction
In 2008, stroke dropped from third to fourth place angiographic vasospasm and clinical outcome, has been
among the most common causes of mortality in the USA, performed.7 Among 14 studies involving 4,235 patients,
in part owing to better supportive care and, possibly, angiographic vasospasm was significantly reduced by the
lower incidence. The incidence of subarachnoid haemor- test drug, but clinical outcomes were unchanged.
rhage (SAH) from an intracranial aneurysm, however, This article reviews our current understanding of
has not declined and still affects nine per 100,000 people angiographic vasospasm and associated DCI after SAH,
per year in the USA, with about 600,000 cases annually as well as management of these conditions. Treatments
worldwide.1 SAH comprises only 5% of all strokes, but currently in development, and possible reasons for
the case fatality is high (40%), owing in part to delayed failure of past trials, are also considered.
cerebral ischaemia (DCI) and neurological deterioration
days after the haemorrhage. Case fatality from SAH has Subarachnoid haemorrhage
decreased by 50% over the past two decades, probably Pathophysiology
because of better diagnosis, immediate aneurysm repair At the time of aneurysm rupture, blood pours into the
to prevent rebleeding, improved medical management, subarachnoid space and sometimes the brain paren-
and introduction of nimodipine.2 The cost to society of chyma and ventricles. The intracranial pressure rises
SAH is greater than that of ischaemic stroke owing to the sharply and might increase enough to compromise cer-
younger average age of those affected (52 years versus ebral perfusion and cause transient global ischaemia,
70–80 years).3 Moreover, many survivors of SAH have which usually produces temporary unconsciousness.
permanent neurological, cognitive or functional deficits Brain injury results from transient global ischaemia and
and cannot return to work.4 from the effects of the intracranial blood itself.
Randomized clinical trials in North America and SAH affects the heart, lungs and other body systems.
Europe have shown that nimodipine and endovascular Increased sympathetic nervous system activity might
coiling improve outcomes of SAH. 5,6 A meta-analysis contribute to effects of SAH such as acute lung injury,
of randomized controlled trials in SAH that tested pulmonary oedema and cardiac dysfunction, including
drugs to improve outcome, and that collected data on takotsubo cardiomyopathy.8 Up to 60% of patients with
Division of SAH develop a systemic inflammatory response syn-
Neurosurgery,
St Michael’s Hospital, drome that is characterized by increased or decreased
Competing interests
30 Bond Street, body temperature, tachypnoea, tachycardia, and leuko-
Toronto, ON M5B 1W8, R. L. Macdonald declares an association with the following
Canada. company: Edge Therapeutics, Inc. See the article online for full cytosis or leukopenia.9 Elevated plasma inflammatory
[email protected] details of the relationship. cytokines might contribute to these effects.10
44 | JANUARY 2014 | VOLUME 10 www.nature.com/nrneurol

© 2014 Macmillan Publishers Limited. All rights reserved
REVIEWS
Key points circulation aneurysm (Figure 1).11 Aneurysm rebleed-

ing is also an important cause of poor outcome, but it is
■ The two main effects of subarachnoid haemorrhage (SAH) are early brain injury
less common now because many ruptured aneurysms
and delayed cerebral ischaemia (DCI)
■ The pathogenesis of DCI is hypothesized to be multifactorial and includes are repaired acutely.12 Many other factors can contribute
angiographic vasospasm, cortical spreading ischaemia, microthrombosis and to a poor outcome.13 The effects of the initial haemor-
microcirculation constriction rhage, which manifest in part through the admission
■ Early brain injury, which refers to the acute effects of subarachnoid blood and neurological condition, have been called early brain
the transient global ischaemia that may accompany aneurysm rupture, has also injury (EBI; Figure 1).14
been suggested to contribute to DCI Factors that develop later and are associated with poor
■ Risk of DCI is increased by volume, density and persistence of the
outcome include cerebral infarction, DCI, fever, and use
subarachnoid thrombus, early brain injury, factors that reduce brain oxygen and
glucose supply and, probably, pre-existing hypertension
of anticonvulsants, principally phenytoin.11 Neurological
■ Treatments that reduce the risk of DCI include nimodipine, rescue therapy deterioration after neurosurgical clipping of rup-
consisting of balloon or pharmacological angioplasty, and induced hypertension tured aneurysms was also identified in up to 43% of
■ Drugs under investigation for DCI include intrathecal dihydropyridines, patients entered into one randomized clinical trial,
magnesium, albumin, sodium nitrite, statins, dantrolene and cilostazol and this deterioration contributed significantly to poor
outcome.15 Broderick and colleagues called attention to
the major effect of the initial severity of the haemorrhage
Causes of poor outcome on outcome.16 In their study, 34 of 36 deaths (94%) were
A multivariate analysis of prognostic factors for attributed to the initial SAH or rebleeding, whereas only
outcome in 3,567 participants in randomized clinical two deaths (6%) were attributed to delayed neurological
trials of tirilazad between 1991 and 1997 showed that deterioration from angiographic vasospasm. Although
poor outcome after SAH was associated with admis- death due to DCI is infrequent,17 morbidity is high, and
sion factors such as worse neurological condition, older DCI doubles the risk of poor outcome.18 Current rescue
age, high blood pressure, a large subarachnoid blood therapies are only marginally effective and are expensive,
clot, pre-existing medical conditions, and posterior costing about US$40,000 per patient.19
Early brain injury Normal

Acute subarachnoid haemorrhage Blood–brain
CBF (microcirculatory constriction,
microthrombosis), blood–brain barrier barrier
Platelet
disruption, cerebral oedema, endothelial cell
Mechanical effects apoptosis, neuronal apoptosis, autophagy Tight junction
Brain tissue destruction by ICH, Astrocyte
mass effect, hydrocephalus endfoot
Cortical spreading ischaemia Erythrocyte
Swollen neuron Basal

Intraventricular Microthrombi lamina
haemorrhage Endothelial
Subarachnoid Pericyte cell
haemorrhage
Subarachnoid haemorrage
Activated Apoptotic
Cellular changes endothelial cell
NO, K+, oxidative stress, increased ET-1, platelet
(TNF-R1,
Aneurysm inflammation (increased cytokines, activation of
chemokines and adhesion molecules; caspase-3 and
activation of matrix metalloproteinases and caspase-8)
apoptotic pathways; platelet activation and Disrupted
aggregation; ATP depletion; thrombin activation) tight junction
■ Platelet activation
Systemic complications ■ MMP-9 degrades collagen IV, laminin, fibronectin
Intracerebral Sympathetic nervous system activation and tight junctions
haemorrhage
Aneurysm bleeding or rebleeding
Systemic complications
Neurogenic pulmonary oedema, stunned
Pulmonary myocardium, takotsubo cardiomyopathy,
Acute global ischaemia ECG changes, hyponatraemia, systemic
Causes CBF secondary to ICP, oedema
inflammatory response syndrome
CPP, acute hydrocephalus
Can lead to microcirculatory Microscopic
constriction, microthrombosis Contraction band necrosis
Cardiac
dysfunction
Figure 1 | Early pathophysiology of subarachnoid haemorrhage. Acute haemorrhage from an aneurysm can physically
damage the brain and lead to acute transient global ischaemia. Transient global ischaemia secondary to increased ICP can
also trigger sympathetic nervous system activation, leading to systemic complications. The contribution of each process to
the pathophysiology is unknown, but transient global ischaemia and subarachnoid blood result in early brain injury,
characterized by microcirculation constriction, microthrombosis, disruption of the blood–brain barrier, cytotoxic and
vasogenic cerebral oedema, and neuronal and endothelial cell death. Abbreviations: CBF, cerebral blood flow; CPP, cerebral
perfusion pressure; ECG, electrocardiographic; ET-1, endothelin-1; ICH, intracranial haemorrhage; ICP, intracranial pressure;
MMP-9, matrix metalloproteinase-9; NO, nitric oxide; TNF-R1, tumour necrosis factor receptor 1.
NATURE REVIEWS | NEUROLOGY VOLUME 10 | JANUARY 2014 | 45

REVIEWS
Disrupted Microcirculatory Angiographic vasospasm

Platelet basal constriction Haemoglobin, oxidative stress and inflammation lead
lamina Microthrombosis to endothelial cell and perivascular nerve injury, leading to
NO, endothelin, leading to membrane depolarization, open
Vasoconstriction voltage-gated calcium channels and activation of TRP channels
Microthrombus Cortical spreading ischaemia

Activated Swollen neuron Microthrombi
platelet
Erythrocyte
Platelet
extravasation
Spreading depressions, neuronal swelling, distorted

dentritic spines, neuronal depolarization, glutamate release
and depression, associated with pathological spreading
wave of vasoconstriction, caused by K, haemoglobin
Infarction and inhibition of NO synthase that are present after SAH
and possibly mediated by increased astrocyte endfeet
calcium that activates large-conductance calcium-activated
potassium channels
Waves of cortical
spreading ischaemia Delayed neurological complications
Seizures, infection, complications of aneurysm repair,
Microemboli aneurysm rebleeding, hydrocephalus, brain swelling,
intracranial haematoma, increased ICP, reduced CPP
Subarachnoid
blood Delayed systemic complications
Infection, fever, pulmonary oedema, cardiac failure,
organ failure, drug adverse effects, hyponatraemia,
Carotid artery hypercarbia, hypoglycaemia, low haemoglobin,
systemic inflammatory response syndrome
Figure 2 | Pathophysiological processes in delayed cortical ischaemia. Key processes include angiographic vasospasm,
microcirculatory constriction and formation of microthrombi, and waves of cortical spreading ischaemia, all of which can
contribute to cerebral infarction. Delayed effects of the early brain injury such as neuronal and endothelial cell apoptosis,
and systemic complications, can also occur. Abbreviations: CPP, cerebral perfusion pressure; ICP, intracranial pressure;
NO, nitric oxide; SAH, subarachnoid haemorrhage; TRP, transient receptor potential.
Delayed neurological deterioration and DCI was shown to be associated with a single nucleo-
Patients can deteriorate days after SAH because of tide polymorphism (Thr786Cys) in the gene encoding
intra cranial or systemic complications (Figure 2). endothelial nitric oxide (NO) synthase.23 Another study
Neurological worsening that cannot be attributed pri- found that polymorphisms in the plasminogen activator
marily to these causes is ascribed to DCI, which occurred inhibitor 1 gene were associated with DCI.24
in 33% of over 30,000 cases of SAH reported before 1994,
and in 29% of 23,806 cases reported between 1994 and Angiographic vasospasm
2009.20 DCI was diagnosed in 22% of 10,739 patients About two-thirds of patients with aneurysmal SAH
who received calcium channel antagonists compared develop angiographic vasospasm 3–14 days after the
with 32% of 13,267 who did not receive such drugs. initial rupture.20 Angiographic narrowing is due to vaso-
Several studies, including one in which patients with constriction induced by subarachnoid blood clot. The
SAH were randomly allocated to clipping or coiling, severity and duration of spasm is related to the thickness,
found that angiographic vasospasm and DCI were less density, location and persistence of the subarachnoid
likely to occur after endovascular aneurysm repair than blood.25,26 Angiographic vasospasm can be reproduced
after neurosurgical clipping.18 DCI and—in particular— in experimental animals through sub arachnoid appli-
infarction from DCI are among the most important cation of blood, and removal of the blood clot pre-
adverse prognostic factors for outcome after SAH.11,21 vents or reverses angiographic vasospasm in animals
and humans.27
Pathophysiology of DCI The main driver of angiographic vasospasm is release
Angiographic vasospasm is associated with DCI, but of haemoglobin and erythrocyte contents through haem-
other factors are hypothesized to contribute to DCI olysis, which initiates several processes, the key ones prob-
pathophysiology, including microcirculatory constric- ably being oxygen free radical reactions, inflammation and
tion, microthrombosis, cortical spreading ischaemia, and endothelial injury, which lead to increased expression of
delayed cell apoptosis (Figure 2). At a physiological level, endothelin-1 and reduced levels of NO.28 Support for
loss of or abnormal autoregulation by the conducting involvement of vasoconstriction and endothelin-1 comes
arteries, arterioles and microvasculature, variations in col- from evidence in humans that dihydropyridines and
lateral and anastomotic blood flow, and metabolic, genetic endothelin receptor antagonists prevent angiographic
and epigenetic variations influence whether a given degree vasospasm.7,29,30 Angiographic vasospasm always resolves,
of angiographic vasospasm leads to DCI (Figure 3). 22 although arterial fibrosis, endothelial thickening and
Increased risk of developing angiographic vasospasm reduced arterial compliance can persist.31

REVIEWS
Superior view of brain the nimodipine clinical trials, treatment did not reduce
angiographic vasospasm, leading to theories that the
Brain susceptibility:
APOE ε4 allele drug improved outcomes because—at doses that do not
cause hypotension—it slightly attenuates angiographic
■ Cardiac output
vasospasm, inhibits cortical spreading ischaemia, and
Infarct
■ Blood pressure and has fibrinolytic activity that reduces microthrombi.34,35
intracranial pressure
Brain ■ Collateral flow
metabolic ■ Collateral fibres Microthrombi
demand ■ Anastamoses
Autopsies of 29 patients with SAH found microthrombi
in brain regions that correlated with premortem angio-
graphic vasospasm and DCI.36 Microthrombi were also
especially abundant in regions of brain infarction. They
Border zone
appeared within 2 days of SAH or were delayed for
several days until DCI developed.
Circle of Willis Cerebral blood flow is reduced within hours of SAH,
especially in poor-grade patients, and again after a
delay, during DCI.37 The early blood flow reduction is
postulated to be due to microcirculatory constriction
Infarct and microthrombosis. In an experimental rat model of
SAH, platelet aggregates were seen in the cerebral pial
microvasculature as early as 10 min after endovascular
perforation, and lasted for up to 3 days.38,39
Genetic factors: Similar processes occur in the intraparenchymal micro-
■ Gene polymorphisms circulation.40 Platelets have been shown to aggregate in
(eNOS, Thr786Cys, plasminogen
activator inhibitor) parenchymal microvessels, forming microthrombi and
■ Epigenetic factors (histones), DNA methylation extravasating into the brain parenchyma within 10 min
Figure 3 | Pathophysiological causes of ischaemia. Clinical and physiological of SAH, and persisting for up to 24 h.41 This process
factors influence whether a given degree of angiographic vasospasm causes was associated with breakdown of collagen IV in the
cerebral infarction. These factors include brain susceptibility to ischaemia, which blood vessel basal lamina, possibly mediated by platelet
can be affected by genetic variations (APOE alleles, and polymorphisms in the activation—a process that releases proteases that digest
genes encoding eNOS and plasminogen activator inhibitor-1) and epigenetic collagen IV, such as matrix metalloproteinase 9. Whether
variations, DNA polymorphisms, the adequacy of the circle of Willis, brain
transient global ischaemia or the subarachnoid blood clot
metabolic demand, and systemic factors (cardiac output, cerebral perfusion
pressure, collateral flow and anastamoses). Abbreviations: APOE, apolipoprotein E; cause the microcirculatory changes after SAH is unclear,
eNOS, endothelial nitric oxide synthase. but transient global ischaemia causes similar pathology
that has been termed the ‘no-reflow’ phenomenon. 42
Clinical and experimental studies suggest that constriction
A causal relationship between angiographic vaso- of cortical and/or intraparenchymal arterioles precedes
spasm and DCI has been questioned. Angiographic thrombus formation both acutely and days after SAH.40,43
vaso spasm has been successfully reduced in some In addition to formation of microthrombi within hours
studies, but outcome was not improved.7 Only about of SAH, large-artery angiographic vasospasm injures the
50% of patients with angiographic vasospasm develop endothelium and causes platelet adherence, coagulation,
DCI, in part because only severe vasospasm is associated and formation of further microemboli.44 Transcranial
with reduced cerebral blood flow and cerebral perfusion Doppler ultrasound of cerebral arteries detects micro-
(Figures 3 and 4).25,32 Among 381 patients with aneurys- embolic signals days after SAH.45 In addition, when an
mal SAH who underwent baseline and follow-up cath- aneurysm ruptures, clotting is activated, which contrib-
eter angiography, as well as serial CT scanning, delayed utes to cessation of the haemorrhage but also potentially
cerebral infarction occurred in 3%, 10% and 46% of generates emboli. Platelet aggregates and clots in the
patients with no/mild, moderate or severe angiographic aneurysm can embolize into the distal circulation. Finally,
vasospasm, respectively. Another study reported similar rats undergoing SAH were found to be in a hypercoag-
findings.32 Development of DCI, therefore, depends on ulable state, suggesting that a systemic hypercoagulation
the extent and severity of angiographic vasospasm, as response to SAH can develop, thereby predisposing the
well as on pre-existing collateral and anastomotic blood animals to formation of microthrombi.46
flow, cerebral metabolic demand, blood pressure, and The importance of microthrombi to brain injury and
other such parameters (Figure 3). outcome in SAH has been studied indirectly in trials of
For DCI to occur, processes in addition to angiographic antiplatelet drugs that should decrease microthrombi, but
vasospasm, such as cortical spreading ischaemia, micro- these trials have been unsuccessful so far (Supplementary
circulatory constriction and/or thrombosis, are thought Tables 1 and 2 online). After experimental SAH induced
to be required. Nimodipine, an L-type calcium channel by endovascular perforation in mice, attenuation of
antagonist, is the only pharmacological agent shown to microthrombosis with a mutant thrombin-activated
improve outcomes in patients with SAH.33 In most of urokinase-type plasminogen activator reduced overall

REVIEWS
a b c d e Cortical arteriole constriction in response to endothelin-1

was also increased acutely after experimental SAH.52
Changes in microvascular reactivity might have a
crucial role in pathogenesis, as studies in which micro-
thrombi were observed after clinical and experimental
SAH suggest that constriction precedes thrombosis.40,43
f g h i
A limitation is that most of these studies examined the
microcirculation within a few days of SAH—a time point
at which DCI might not have developed.
j k Inflammation
The main classes of genes that show increased mRNA
expression after experimental SAH include genes
encoding proteins involved in inflammation (including
inflammation-related cytokines), cell proliferation and
the cell cycle, and metabolic pathways, signal transduction
and other cellular processes, as well as genes encoding
membrane receptors, extracellular matrix components,
cell adhesion molecules, and proapoptotic proteins.53 Most
studies of inflammation in SAH have focused on the role
of inflammation in angiographic vasospasm.54 Vasospastic
arteries are infiltrated by inflammatory cells and immuno-
globulins.54 White blood cells are hypothesized to cross
Figure 4 | CT scan and cerebral angiography of SAH. a | A 40-year-old woman was arterial walls and infiltrate the subarachnoid clot, where
admitted to hospital with SAH caused by a small aneurysm of the precommunicating
they increase expression of endothelin and inflammatory
segment of the left anterior cerebral artery. b | Oblique view of left internal carotid
artery angiogram. Diffuse thick SAH was evident on CT scan, World Federation of
cytokines, and trigger oxidative stress.54
Neurological Surgeons grade was 2 and the cerebral arteries were of normal calibre. Haemoglobin may mediate brain injury after SAH
c | Coronal view of CT angiography. d | Sagittal view of CT angiography. e | Axial view and, as the naturally occurring protein haptoglobin binds
of CT angiography. The aneurysm was clipped on the day of admission and the haemoglobin and potentially reduces its toxicity, studies
postoperative CT scan did not show hypodensities. Her consciousness deteriorated have examined the role of haptoglobin in SAH.55 There are
and she developed right hemiparesis, which did not improve with induced two allelic forms of the haptoglobin gene in humans—Hp1
hypertension. f,g | Catheter angiography showed severe diffuse vasospasm 4 days and Hp2. The Hp 1-1 genotype is associated with more-
(f) and 5 days (g) after SAH. Balloon and pharmacological angioplasty with milrinone
potent anti-inflammatory and vasodilatory effects than
was performed 4 and 5 days after SAH. h,I | Angiograms after treatment with balloon
and pharmacological angioplasty. j,k | Multiple cerebral infarcts developed days later
are Hp 2-1 and Hp 2-2.56 In one study, mice expressing
and the patient died. Abbreviation: SAH, subarachnoid haemorrhage. the human Hp2 allele had more vasospasm, macrophages
and neutrophils in the subarachnoid space after SAH than
did mice expressing the Hp1 allele.56 In another study,
mortality.47 The discrepancy between these experimen- patients with SAH who had the Hp 1-1 genotype possibly
tal findings and clinical trial results might be due to the had less vasospasm than patients without this genotype, as
fact that the specific mechanism targeted in the animal measured by transcranial Doppler ultrasound.57
experiment was not targeted in the human studies, or Complement activation also plays a part in angio-
because of other limitations of human trials in SAH graphic vasospasm. Decaying erythrocytes activate
(discussed below). complement, which then promotes erythrocyte lysis and
release of haemoglobin.58 Levels of activated complement
Microvascular constriction components can be increased in cerebrospinal fluid (CSF)
Blood applied to the cortical surface of rats in the absence after SAH in humans, and depletion of comp lement
of increased intracranial pressure causes pial arteri- decreases experimental SAH-induced vasospasm.59
ole constriction, microthrombi and egress of platelets Numerous inflammatory biomarkers—including inter-
into the brain, as well as disruption of the blood–brain leukins, selectins and adhesion molecules—are expressed
barrier (BBB).48 The same effects are observed hours at elevated levels in plasma after SAH, with varying pro-
to days after experimental SAH.40,49,50 The cause of pial files and associations with DCI.60 In a meta-analysis of
arteriole constriction could be altered vascular reactivity studies in patients with SAH, increased levels of CSF
due to endothelial injury. SAH created by endovascular inflammatory biomarkers—such as tumour necrosis
perforation in rats was associated with impaired corti- factor (TNF), soluble TNF receptor 1, and IL-1 receptor
cal pial arteriolar vasodilation in response to topical antagonist—were associated with poor outcome, but not
adenosine or sodium nitroprusside.51 Given that nitro- convincingly with angiographic vasospasm or DCI.61
prusside acts directly on smooth muscle cells, the results Studies in animals demonstrate that angiographic
suggest either that arteriolar smooth muscle was dys- vasospasm can be reduced by pharmacological inhibi-
functional or that subarachnoid blood prevented the tion of the proinflammatory signal transduction mol-
NO-mediated vasodilatory effect of nitroprusside. ecules c-Jun N-terminal kinase and poly(ADP-ribose)

REVIEWS
polymerase, or administration of monoclonal antibodies in an inverse haemodynamic response (possibly due

to block the function of the adhesion molecules ICAM-1 to increased astrocytic calcium levels and activation of
or CD11/CD18.62 Ibuprofen, which inhibits ICAM and large-conductance, calcium-activated potassium chan-
VCAM, was found to reduce experimental vasospasm.63 nels), which can cause waves of cortical hypoperfusion
Other anti-inflammatory strategies have been unsuccess- or cortical spreading ischaemia.77
ful both experimentally 64 and clinically (Supplementary Evidence suggests that spreading depolarizations are
Tables 1 and 2 online). 65 In one study involving 95 initiated by neuronal depolarizations associated with
patients who were randomly assigned to receive placebo sodium and calcium influx through multiple types of ion
or methylprednisolone, the active group did not show channels, leading to loss of the normally electrically neg-
decreased DCI compared with the control group, but did ative intraneuronal state.35,78 This depolarization prevents
have significantly improved outcome.66 Statins have mul- action potential formation and could propagate as a wave
tiple actions of potential benefit in SAH, including anti- across the cortex. Spreading depolarization can result in
inflammatory actions, but meta-analyses of the studies neuronal death if neurons lack adequate energy sup-
conducted to date have reached conflicting conclusions plies to re-establish transmembrane ionic gradients, but
as to whether this approach had effects on angiographic whether neuronal death occurs depends on the severity
vasospasm, DCI or outcomes in patients.67,68 of the injury that initiated the depolarization. After SAH,
Inflammation is also involved in the pathophysiol- conditions in the subarachnoid space—such as elevated
ogy of the microvascular and brain response to cerebral potassium concentrations from the subarachnoid blood
ischaemia and, therefore, contributes secondarily to clot, and reduced levels of NO, which is avidly scav-
DCI.69 An emerging concept is that biological processes enged by haemoglobin—favour development of cortical
including inflammation can have detrimental and ben- spreading ischaemia.35
eficial effects after stroke, so prolonged administration Patients with SAH undergoing craniotomy for aneur-
of nonspecific inhibitors may not be efficacious.70 This ysm clipping have been monitored with subdural strip
concept has not yet been studied in SAH. electrodes.35 Up to 70% of these patients had repetitive
cortical spreading depolarizations. In one report of 18
Blood–brain barrier disruption patients, seven patients developed DCI that was spa-
Preclinical studies demonstrate disruption of the BBB tially and temporally associated with angiographic vaso-
after SAH, which can occur acutely and contribute to spasm and cortical spreading depolarization. Four patients
EBI.71 The blood–artery wall barrier is disrupted in large who had depolarizations lasting longer than 60 min devel-
arteries several days after experimental SAH, but how oped infarction on CT or MRI. Experimentally, cortical
this affects vasospasm and DCI is not well-studied.72 spreading ischemia after SAH can be inhibited by nimodi-
Administration of antioxidants such as (±)-N,N’- pine, N-methyl-d-aspartate (NMDA) receptor antagonists,
propylenedinicotinamide (also known as anti-vasospasm and prevention of systemic volume depletion.35
substance [AVS]), or abrogation of activity of the pro-
apoptotic protein PUMA through use of small interfer- Predictors of DCI
ing RNA, reduced BBB disruption, cerebral oedema and Volume, location, persistence over time and density of
mortality and improved neurological function 24–48 h the subarachnoid blood clot are all associated with the
after experimental SAH.73,74 risk of angiographic vasospasm and DCI.26 Poor clinical
In humans, BBB disruption after SAH was inferred condition on admission and loss of consciousness at ictus
on the basis of contrast enhancement on CT scans, increase the risk of DCI. Strong evidence suggests that
and occurred in about 40% of patients within 5 days of smoking increases the risk of DCI, and moderate-strength
SAH, and in 11% of patients thereafter.75 Breakdown evidence indicates that diabetes, systemic inflammatory
of the BBB probably contributes to brain injury in SAH response syndrome, hyperglycaemia and hydrocephalus
but, similar to inflammation, might also have benefi- also increase the risk.79 Sex of the patient and history of
cial effects. For example, activation of matrix metallo- hypertension seem to have a lesser role, if any. Evidence
proteinases that acutely disrupt the BBB may facilitate regarding the influence of age on DCI risk is inconsistent.
brain repair at a later stage.76 A literature review on CSF molecules concluded that
endothelin-1, IL-6 and some markers of thrombin activa-
Cortical spreading ischaemia tion might be useful for predicting DCI following SAH.61
Spreading depolarizations represent waves of cortical Serum biomarkers—namely, TNF, IL-6, S100β, ubiqui-
neuronal swelling, dendrite spine distortion, and slow tin C-terminal hydroxylase 1, phosphorylated axonal
changes in brain electrical potential (spreading depolar- neurofilament heavy chain, matrix metalloproteinases,
ization), as well as decreased brain electric activity von Willebrand factor, endothelin-1, vascular endothelial
(spreading depression).35 growth factor, selectins and adhesion molecules—have
Cortical arterioles can respond to spreading depolar- shown an association with DCI risk, but none has been
izations with vasodilation and hyperaemia, which reproducibly validated as a marker for DCI.60
is similar to the normal hyperaemic neurovascular
coupling response to neuronal activity, and tends not Diagnosis of DCI
to result in permanent brain damage. In some cases, Delayed neurological deterioration days after SAH occurs
such as after SAH, however, cortical arterioles constrict in up to 50% of patients and can have both systemic and

REVIEWS
neurological causes (Figures 2 and 3). Such deterioration time threshold will vary between centres and scan-
emerges 3–14 days after a single SAH, with only about 5% ners, but the principle that this parameter is a sensitive
of cases occurring after day 10.80 measure of cerebral blood flow has been reproduced.89
A group of physicians who manage patients with SAH
convened to standardize definitions and terminology Transcranial Doppler ultrasound
for delayed neurological deterioration after SAH.81 The Transcranial Doppler ultrasound is still used for detec-
diagnosis of DCI is one of exclusion. Systemic and neuro- tion of angiographic vasospasm, but it requires a tech-
logical complications (Figure 2) need to be identified nician or experienced user. This approach has been
by laboratory and radiological investigations. Patients reported to have over 95% sensitivity for severe angio-
often have multiple abnormalities on these tests plus graphic vasospasm,90 although in practice this level of
some degree of angiographic vasospasm, so the cause accuracy may not be seen.79 Limitations of transcranial
of delayed neurological deterioration can be DCI, other Doppler ultrasound are that it only assesses a small
factors or a combination thereof. number of large arteries, and that it cannot be applied
Detection of DCI remains difficult, especially in poor- in some patients owing to thick temporal bones. False
grade patients, who may be sedated and ventilated.82 positives can occur when blood pressure is augmented,
Angiographic vasospasm occurs in 67% of patients with and false negatives can be seen in cases with very severe
SAH, and is moderate to severe in one-third of cases.30 artery narrowing owing to low blood flow. Pulsatility
It can be asymptomatic, however, and is not always index and abnormal autoregulation can also be assessed
associated with DCI. DCI can also be asymptomatic: in by this technique.91
patients who were not diagnosed clinically with DCI,
CT scanning and MRI revealed cerebral infarcts in EEG and other technologies
10–20% and 23%, respectively, of patients.83,84 Cerebral Other methods to detect angiographic vasospasm and
infarction has a strong relationship with poor outcome, DCI include continuous EEG, monitoring of brain tissue
and the two most common causes of infarction are the oxygen levels, jugular-bulb oximetry, microdialysis,
aneurysm-securing procedure and DCI.85 near-infrared spectroscopy, and focal cerebral blood
Predicting who will develop DCI has been difficult, flow measurement by thermal diffusion.92 Continuous
and a reduction in the level of monitoring in patients EEG has the advantage of being able to monitor broad
at low risk would be helpful to avoid adverse effects of regions of brain and to detect epileptiform discharges,
monitor ing and to reduce health-care costs. A study which were observed in 23% of 108 patients with SAH in
of 307 patients with SAH found that patients over one series.93 Among 32 patients with SAH, reduced rela-
68 years of age who were World Federation of Neurologic tive (to other waves) variability in alpha-wave activity was
Surgeons grades 1–3 on admission with no or thin SAH 100% sensitive and 50% specific for vasospasm (defined
on CT scans had low risk of DCI, and might, therefore, as increased transcranial Doppler flow velocities or angio-
be candidates for less-frequent monitoring.86 graphic arterial narrowing).94 Overall, studies of continu-
ous EEG are limited, and were judged to be of moderate
Angiography to low methodological quality and likely to be biased.95
Potential contributors to DCI such as cortical spreading
depolarization and microthrombi are not easily detect- Invasive brain-tissue monitoring
able clinically, so angiographic vasospasm has been a key Brain tissue oxygen monitoring measures tissue oxygen
marker for DCI. Catheter angiography remains the gold in a small region around the inserted probe, and carries
standard for detection of angiographic vasospasm, but is risks associated with any invasive monitor that is inserted
being replaced by CT angiography. Several studies show into the brain. This approach has been reported to
good correlation between catheter and CT angiography, predict onset of DCI after SAH and to correlate with
with accuracy of 87.0–97.5% and negative predictive outcome, although the data are limited and were judged
value of 95.0–99.5%.87 CT angiography can overestimate of moderate to low quality.95
the degree of vessel narrowing. Moreover, determin- Brain microdialysis is another focal, invasive moni-
ing the degree of vasospasm can be difficult owing to toring method that might have utility for identification
‘beam-hardening’ artefacts from clips and coils, but this and monitoring of DCI. Cerebral ischaemia is associ-
limitation can be at least partly overcome by windowing ated with elevated glutamate and glycerol levels and
and levelling of the CT, as well as by adjusting the plane an increased lactate:pyruvate ratio. These compounds,
of image reconstruction. measured by microdialysis, correlate well with ischaemia
A retrospective study of 27 patients with SAH com- if the microdialysis probe is inserted near to the site of
pared diagnosis with CT angiography, CT perfusion, ischaemia. Abnormal values were detected before onset
catheter angiography and transcranial Doppler ultra- of DCI in six of 11 patients in one series, suggesting that
sound.88 The most accurate combination of tests to diag- the sensitivity of the method could be low.96
nose angiographic vasospasm on catheter angiography, Cerebral blood flow can be monitored focally by inva-
with 93% accuracy, was qualitative assessment of artery sive thermal diffusion flowmetry. In 14 patients with
narrowing on CT angiography combined with CT perfu- SAH, reduced cerebral blood flow measured by this
sion mean transit time—the mean time it takes for blood method was a more sensitive and specific marker of DCI
to perfuse a region of tissue—of >6.4 s. The mean transit than was transcranial Doppler ultrasound, and preceded

REVIEWS
Table 1 | Drugs in development for treatment of SAH

Drug class Putative mechanisms of action Status
Statins Inhibit HMG-CoA reductase, thereby reducing synthesis of Six randomized clinical trials of statins in
cholesterol, geranylgeranylpyrophosphate and patients with SAH68
farnesylpyrophosphate122 Systematic review of these studies found no
Preserve endothelial function via increased synthesis of nitric effect of statin treatment on poor outcome;
oxide, and decreased synthesis of endothelin-1 and RhoA mortality was 10% in statin group versus 21%
Anti-inflammatory effects in controls (RR 0.62, 95% CI 0.36–1.06); DCI
was significantly reduced in statin group
Antioxidant effect via decreased production of oxygen free
radicals and peroxynitrite, and decreased expression of Quality of the studies overall was low to
angiotensin receptors and NADPH oxidase moderate
Antithrombotic actions At least two ongoing clinical trials123,125
Vascular protection through decreased expression of matrix Current recommendation: only administer
metalloproteinases statins if the patient was already receiving
them at time of SAH
Neuroprotective and neurorestorative action (increased
synaptogenesis, increased neurogenesis)
Magnesium Vasodilation via antagonism of calcium channels on vascular Studied in seven randomized clinical trials127,154
smooth muscle and increased endothelial cell prostacyclin126 Meta-analysis reported no effect of
Endothelial protection through inhibition of platelet aggregation magnesium on poor outcome
and decreased production of angiotensin-converting enzyme Therapeutic intravenous infusions of
Protect the blood–brain barrier magnesium are not recommended for patients
Reduce cerebral oedema via decreased aquaporin-4 expression with SAH131
Anticonvulsant (N-methyl-D-aspartate receptor antagonism)

Anti-inflammatory Prevent vasospasm, inhibit IL-2 production, and prevent T-cell Literature review concluded that NSAIDs have
drugs dysfunction (using cyclosporin A)132,155 little measurable benefit for treatment of SAH134
Inhibit complement activation and inflammation (using FUT-175 Glucocorticoid steroids studied in three SAH
[nafamostat mesilate])133 trials, but too few patients studied (256) to
Acetylsalicylic acid and thromboxane synthase inhibitors have draw conclusions about efficacy65
anti-inflammatory and antiplatelet activity33,112 Randomized controlled trial found no effect of
Prevent upregulation of endothelial cell adhesion molecules methylprednisolone on angiographic
that allow binding of macrophages and neutrophils (which vasospasm or DCI after SAH, but significantly
infiltrate the subarachnoid space, die and degranulate, improved outcome at 1 year compared with
releasing endothelin-1 and oxygen free radicals)54 controls66
Corticosteroids have multiple anti-inflammatory actions, mostly

on chronic inflammation
NSAIDs inhibit cyclo-oxygenase, which decreases prostaglandin
synthesis; ibuprofen inhibits expression of endothelial
adhesion molecules and reduces subarachnoid inflammation
Abbreviations: DCI, delayed cerebral ischaemia; HMG-CoA, 3-hydroxy-3-methylglutaryl coenzyme A; SAH, subarachnoid haemorrhage.
onset of DCI.97 Finally, near-infrared spectroscopy is SAH (Supplementary Table 1 online). Some drugs—
a noninvasive method that measures cerebral blood including tirilazad, magnesium and clazosentan—have
oxygen saturation, and has been used, with variable been studied in thousands of patients, and meta-analysis
results, to detect DCI in patients with SAH.95 of these trials has shown that only nimodipine and pos-
sibly fasudil are effective (Supplementary Table 2 online).
Treatment of DCI A common theme is that treatment reduces angiographic
The principles of treatment of DCI are to avoid factors vasospasm and/or DCI, but does not improve outcome.7
that reduce supply of (or increase demand for) oxygen- One explanation for the lack of observed improvement
rich and glucose-rich blood to the brain, and to maintain could be that the sample size in some trials was too small.
physiological variables as close to normal as possible. However, this reason is unlikely to apply to drugs that
Little of this approach is based on high-quality medical were studied in thousands of patients, unless we are over-
evidence. Most trials have targeted angiographic vaso- looking a major cognitive improvement or flaws in study
spasm, although drugs with multiple known actions, design. Another possibility is that the outcome measure,
including neuroprotection, are increasingly being tested which is generally a dichotomous (‘good’ or ‘bad’) assess-
(Table 1 and Supplementary Tables 1 and 2 online). ment of the modified Rankin or Glasgow Outcome Scale
score, was not sensitive to clinical improvement.98
Targeting angiographic vasospasm Some have suggested that angiographic vasospasm
Lack of translation to improved outcome does not cause poor outcome—a hypothesis that is dif-
Many randomized clinical trials with at least single ficult to dispute until a treatment that specifically inhibits
blinding and not directed at aneurysm repair or preven- angiographic vasospasm is shown to improve outcome.
tion of rebleeding have been conducted in patients with Drug adverse effects such as hypotension and pulmonary

REVIEWS
complications could counteract the benefit of the drugs in with reduced cerebral blood-flow velocities as measured
reducing angiographic vasospasm. Other mechanisms that by transcranial Doppler ultrasound.106
cause brain injury in addition to angiographic vasospasm Other primarily vasodilatory treatments under
may not be targeted by some drugs. Moreover, practice investigation include NO donors. Sodium nitroprusside
misalignment in clinical trials—whereby modification of and nitroglycerin have limited efficacy owing to dose-
treatment according to the features of an individual patient limiting hypotension, development of tolerance, and
is not possible owing to the trial protocol—can mask ben- cyanide toxicity. Sodium nitrite (NaNO2) is a vasodilator
efits of drug treatment that only apply to a subgroup of that also possesses cytoprotective and antiapoptotic
patients. Finally, use of rescue therapy is more common actions.107 This chemical is abundant in green vegetables
in placebo groups than in drug-treated groups in some and can be formed in the bowel by conversion of nitrate
studies, and could improve outcome such that benefit of to nitrite. Intravenous NaNO2 decreased angiographic
the drug cannot be demonstrated. vasospasm in a nonhuman primate model of SAH, and
did not induce hypotension when infused into humans
Clinical trials for up to 48 h at doses associated with plasma NO con-
The only drug approved for SAH in the USA and Europe centrations that would be expected to be therapeutic.107
is the dihydropyridine, L-type calcium channel antago- A study of infusion of NaNO2 for 14 days in 18 patients
nist nimodipine. Meta-analysis of clinical trials found with SAH found no toxicity or systemic hypotension at
that oral nimodipine reduced the risk of DCI and poor doses that were comparable to the efficacious doses in
outcome.33 Possible mechanisms of action include reduc- nonhuman primates.108
tion of angiographic vasospasm, increase in fibrinolytic
activity, neuroprotection, and inhibition of cortical Local versus systemic delivery
spreading ischaemia. Lack of major effect of nimodi- Drugs that target angiographic or microcirculatory
pine on angiographic vasospasm in these trials argues vasconstriction will, theoretically, dilate cerebral and
for these other actions. systemic blood vessels to the same extent unless the
Experimental evidence that endothelins mediate drug acts on receptors or mechanisms that are present
angiographic vasospasm supported the conduct of multi- only in cerebral blood vessels. Drugs such as calcium-
ple clinical trials of the endothelin A receptor antagonist channel and endothelin-receptor antagonists and
clazosentan. 30,99,100 Meta-analysis of trial results con- nitric oxide donors do not have high cerebrovascular
firmed that clazosentan reduced angiographic vasospasm selectivity, so systemic administration is limited by
without a significant effect on outcome.101 Drug adverse systemic hypotension.
effects included hypotension and pulmonary compli- One potential solution to this problem is local,
cations, which could have counteracted the beneficial intrathecal drug delivery. This approach has been used
effects of the drug. for delivery of fibrinolytic drugs to promote clearance
Fasudil, a rho-kinase inhibitor that also inhibits of the SAH, for delivery of vasodilator drugs to prevent
other kinases, reduces smooth muscle contraction and angiographic vasospasm, and for administration of
inhibits TNF-induced IL-6 release from C6 glioma drugs to prevent processes that are thought to contrib-
cells.102 Other effects in experimental animals include ute to vasospasm, such as inflammation.27,109 Fibrinolytic
suppression of endothelial cell tissue factor expression, agents—principally urokinase and recombinant tissue
stimulation of endogenous neural stem cells, increased plasminogen activator—are the most studied in humans.
growth factor release, and decreased intracellular Intrathecal thrombolytics were studied in five random-
calcium-mediated signalling.103 ized trials that included 465 patients.27 The rationale for
Fasudil was studied in eight randomized clinical this therapy is based on the hypothesis that subarachnoid
trials that included 843 patients (386 exposed to the blood causes angiographic vasospasm and possibly
drug).104 Treatment significantly reduced the incidence other delayed complications (namely, cortical spreading
of angiographic vasospasm and cerebral infarction, and ischaemia and microthrombosis), and that clearance
improved the odds ratio for good recovery compared of the clot will rapidly reduce these complications and
with placebo or nimodipine and other drugs (OR 1.58, improve outcome. Meta-analysis of these studies found
95% CI 1.12–2.23). Limitations of the studies included that thrombolysis was associated with significant reduc-
outcome assessment very early after SAH, small sample tions in angiographic vasospasm, delayed neurological
sizes, and concomitant use of other medications such as deficits, hydrocephalus and poor outcome.
steroids and ozagrel. Hypotension is the main adverse Many other drugs have been injected intracisternally
effect of fasudil. The drug is approved for use in patients in animal models of SAH but, given the delayed onset
with SAH in Japan and China, but has not been pursued of DCI, prolonged drug infusions or sustained-release
for development in Europe and the USA. formulations might be necessary. Sustained-release form-
Dantrolene inhibits ryanodine receptors and reduces ulations of papaverine, dihydropyridine calcium channel
intracellular calcium release in smooth muscle. The drug antagonists, ibuprofen, NO donors, calcitonin gene-
may be neuroprotective.105 A dose-escalation study that related peptide and fasudil have been reported to reduce
included 10 patients with SAH found that 2.5 mg/kg angiographic vasospasm after experimental SAH.109 In
dantrolene administered over 60 min reduced cerebral humans, experience is limited to nicardipine in poly-
blood pressure in some cases, and was also associated (d,l-lactide-co-glycolide) pellets.110 A randomized trial

REVIEWS
in 32 patients with SAH found that pellets containing trials are assessing 40 mg per day simvastatin (includ-
up to 40 mg nicardipine, administered into the basal cis- ing the Simvastatin in Aneurysmal Subarachnoid
ternae at the time of craniotomy for aneurysm repair, Haemorrhage [STASH] trial123), and other studies are
significantly reduced angiographic vasospasm, cerebral comparing 40 mg per day with 80 mg per day simva-
infarction and mortality. 110 Approximately 236 addi- statin.124,125 At present, no compelling evidence supports
tional patients have been treated in open-label studies administration of statins to patients with SAH. The
in Japan.111 general consensus is that statins should be continued in
patients who are already receiving such drugs when they
Targeting microthromboembolism present with SAH.
Meta-analysis of seven randomized clinical trials of Magnesium has multiple potentially beneficial effects
antiplatelet drugs in a total of 1,385 patients found in SAH, such as cerebrovascular dilation and neuro-
trends towards reduction in poor outcome (RR 0.79, protection, but these properties have not resulted in
95% CI 0.62–1.01) but also towards increased intra- improved outcomes in humans with SAH (Table 1
cranial haemor rhage (RR 1.36, 95% CI 0.59–3.12).112 and Supplementary Tables 1 and 2 online).126,127 High
Further trials are needed to determine whether these plasma magnesium concentrations can lead to adverse
drugs are efficacious in this indication. One question effects such as hypotension, arrhythmia, cardiac toxi-
is whether these drugs could be beneficial in patients city, and decreased level of consciousness.128 Questions
who are undergoing endovascular aneurysm repair, as about the studies of magnesium include whether treat-
this procedure can be associated with thromboembolic ment was started sufficiently early after SAH, and
complications.113 However, a retrospective assessment of whether it was continued for too long after the SAH,
the results of the International Subarachnoid Aneurysm thereby reducing neuronal recovery by blocking NMDA
Trial found no difference in outcomes between centres receptor-mediated neuronal recovery.129 Intravenous
that did and did not administer antiplatelet drugs during magnesium also only minimally elevates CSF magnesium
or after endovascular coiling.114 concentrations—possibly not enough to effectively antag-
onize cerebrovascular constriction or be neuroprotec-
Neuroprotection through targeting of DCI tive—and increases in CSF magnesium levels only occur
Preclinical data suggested that albumin is neuroprotec- several hours after beginning infusion.130 One alternative
tive, and observational studies in humans found that it approach has been to infuse magnesium sulphate into
reduced DCI and improved outcome.115 In an open-label the basal cisternae.131 Infusion of 15 mmol/l magnesium
dose-escalation trial of intravenous human albumin, sulphate at 20 ml/h increased CSF magnesium concentra-
patients with SAH tolerated up to 1.25 g/kg per day tions and dilated cerebral arteries in patients with SAH.131
of albumin and, at this dose, showed a trend towards At this time, therapeutic intravenous infusions of mag-
improved outcome.116 nesium are not recommended in patients with SAH, but
Erythropoietin was neuroprotective in animal models hypomagnesaemia should be avoided.127
of ischaemic stroke, prevented loss of autoregulation Anti-inflammatory drugs have been studied in
after SAH in rats, 117 and reduced angiographic vaso- limited numbers of patients with SAH (Table 1).66,132–134
spasm and delayed neurological deterioration, and Acetylsalicylic acid and thromboxane synthase inhibi-
improved outcome after SAH in rabbits.118 The com- tors, which have anti-inflammatory as well as antiplatelet
pound inhibits apoptosis and stimulates neurogenesis activity, have also been studied, but did not show clini-
and angio genesis. 119,120 Two randomized controlled cal benefit.112 Trials of glucocorticoid steroids have had
trials of erythropoietin for treatment after SAH have mixed results, and sample sizes were too small to enable
been conducted.120,121 One was terminated early owing conclusions to be drawn regarding efficacy.66 At present,
to slow enrolment and did not show a difference in out- anti-inflammatory drugs are not indicated in SAH.
comes between active and control groups.121 The second A retrospective analysis of lumbar drains after SAH
trial, however, found that patients who received erythro- found that patients who underwent this intervention
poietin had fewer cerebral infarcts, shorter duration of had a reduced incidence of DCI and cerebral infarction
autoregulatory dysfunction and better clinical outcome and better outcomes compared with patients who did
compared with patients who did not receive this drug.120 not undergo the intervention.135 A randomized clini-
cal trial that included 210 patients found that lumbar
A multitarget approach drainage reduced the incidence of DCI, but at 6 months
Statins are small-molecule derivatives of mevastatin only trends towards reduction in cerebral infarction
that inhibit 3-hydroxy-3-methylglutaryl coenzyme A and improved outcome on the modified Rankin Scale
(HMG-CoA) reductase. The various statins differ in were observed.136
terms of water solubility and BBB permeation,122 and Cilostazol is a 2-oxo-quinoline derivative that inhib-
have multiple biochemical effects that could reduce its phosphodiesterase 3 and has antithrombotic, vaso-
angiographic vasospasm and DCI and improve outcome dilatory, anti-smooth muscle proliferation and cardiac
after SAH (Table 1). inotropic and chronotropic effects.137 A randomized,
An advantage of statins is that they have minimal single-blind study of 109 patients undergoing clipping
adverse effects, only occasionally causing elevation of of ruptured aneurysms found that cilostazol significantly
liver enzymes, and muscle aches. At least two ongoing reduced angiographic vasospasm, DCI and cerebral

REVIEWS
Initial treatment Avoid

■ Early aneurysm repair ■ Hypotension ■ Hyperglycaemia ■ Anticonvulsants unless
■ CT/CTA/CTP on admission and on ■ Increased ■ Hypoglycaemia documented seizures
days 3–5, days 7–10 ICP ■ Hyponatraemia ■ Low-volume centre
■ Oral nimodipine ■ Low CPP ■ Hypomagnesaemia without access to
■ Systolic blood pressure <160 mmHg ■ Fever endovascular specialists
■ Hypocarbia and multidisciplinary
before aneurysm repair ■ Hypovolaemia ■ Hypoxia neurocritical care
■ Euvolaemia
■ Haemoglobin >8 g/dl
Low risk High risk High risk

WFNS 1, 2, no or thin SAH, WFNS 1–3, no or thin SAH with IVH Inaccessible due to sedation
no IVH, older age or thick SAH without IVH or WFNS 3–5, thick SAH with IVH
Neurological assessment every hour, fluid

Neurological assessment every 4 h, Neurological assessment every balance every 6 h, consider fludrocortisone,
fluid balance every 2 h, 2 h, fluid balance every 12 h, daily weight, TCD daily, brain tissue oxygen
TCD every second day TCD every second day and blood flow monitor, arterial catheter
No change neurologically, normal TCD, normal CTP
Low risk High risk

Consider transfer to lower level Consider transfer to lower level Elevated TCD velocities
monitoring 5–10 days after SAH monitoring 7–10 days after SAH (>120 cm/s or Lindegaard No change
ratio >3), abnormal CTP, neurologically,
abnormal brain tissue normal TCD,
Elevated TCD velocities (>120 cm/s Elevated TCD velocities (>120 cm/s oxygen or cerebral CTP, brain tissue
or Lindegaard ratio >3), increased or Lindegaard ratio >3), abnormal blood flow, neurological oxygen, cerebral
vigilance for DCI CTP, neurological deterioration deterioration blood flow
Detect causes of neurological deterioration Consider transfer to lower level

monitoring 14 days after SAH
Correct identified causes (for example, seizures,

increased ICP, hydrocephalus, No major identified causes for deterioration: CTA/CTP or catheter angiogram
hyponatraemia, infections) Consider empiric induced hypertension if aneurysm has been repaired
No major elevation of MTT or angiographic CTP elevated MTT or angiographic Deficit persists >60 min or patient
vasospasm; or deficit resolves vasospasm, persisting deficit cannot tolerate induced hypertension
Induced hypertension, fluid boluses Consider catheter angiography and balloon

Continue monitoring as necessary, vigilant monitoring and/or pharmacological angioplasty
Figure 5 | A management scheme—our approach to SAH and DCI. Abbreviations: CPP, cerebral perfusion pressure; CTA,
CT angiography; CTP, CT perfusion; DCI, delayed cerebral ischaemia; ICP, intracranial pressure; IVH, intraventricular
haemorrhage; MTT, mean transit time; SAH, subarachnoid haemorrhage; TCD, transcranial Doppler; WFNS, World
Federation of Neurological Societies grade.
infarction but had no effect on outcome. However, the evaluate brain oxygen delivery in patients with DCI
sample size in this study was small.137 who were treated with fluid augmentation, induced
hypertension or transfusion of erythrocytes.141 All three
Rescue therapies manoeuvres improved brain oxygen delivery to regions
When a patient develops DCI, rescue therapies such as where it was low, although in these noncontemporan-
haemodynamic therapy or endovascular angioplasty are eously treated cohorts, induced hypertension was most
typically employed (Figures 4 and 5). Current recom- effective. Transfusion was effective mainly in patients
mendations in the first few days after SAH and aneur- with initial haemoglobin levels of <9 g/dl.
ysm repair are to maintain euvolaemia, and to refrain In addition to considering haemoglobin concentra-
from using prophylactic hypervolaemia, induced hyper- tion, haemodynamic therapy—which includes manipu-
tension or haemodilution.138 The optimal haemoglobin lation of intravascular volume and blood pressure—is a
concentration is unknown, and guidelines recommend current rescue therapy approach. Randomized trials of
that decisions should be made on a case-by-case basis.139 this approach have been conducted, but only one used
Some physicians administer blood transfusions when current management protocols.142
haemoglobin levels are <7 g/dl in patients without DCI, A systematic review 143 found that the limited number
or <10 g/dl in patients with DCI. As blood transfusion of randomized trials precluded assessment of the effects of
has been associated with worse outcome, prevention haemodynamic therapy on outcome, so the effects of this
of anaemia (which would require a blood transfusion) approach on cerebral blood flow were assessed instead.
is recommended. The most consistent increases in cerebral blood flow were
If the patient deteriorates owing to DCI, induced seen with induced hypertension. Two randomized trials
hypertension is recommended. 140 PET was used to of prophylactic induced hypervolaemia showed no effect

REVIEWS
on outcome, but did increase the incidence of complica- management has failed or is considered to be risky owing
tions, mainly pulmonary oedema. Observational studies to cardiac or other medical problems.
also suggest that brain tissue oxygen can be elevated by In one clinical trial, 71 patients with aneurysmal SAH
induced hypertension, but not by hypervolaemia.144,145 who had their aneurysms repaired by neurosurgical clip-
An international multidisciplinary consensus confer- ping were randomly assigned to a group receiving intra-
ence reviewed many aspects of the medical management aortic balloon counterpulsation or to a control group.153
of SAH, and reported conclusions that were based on Cerebral blood flow (measured daily with 133Xe), inci-
the grades of recommendation assessment, development dence of DCI and outcome on the Glasgow Outcome
and evaluation (GRADE) system.82 The participants rec- Score did not differ between groups at 6 months.
ommended that particular attention should be paid to Moreover, one morbid complication occurred in the
blood-volume status, blood pressure and haemoglobin treated group. The results suggest that nonspecific use
concentration in patients with SAH. Hypovolaemia is of this invasive treatment is unlikely to be useful.
common in patients with SAH and is associated with
DCI.146 Maintenance of euvolaemia was recommended; Conclusions
use of fludrocortisone was only weakly recommended. Angiographic vasospasm and DCI contribute to poor
Balloon angioplasty was first reported as a treatment outcome after SAH. Early brain injury is also a major
for SAH in 1984.147 One randomized trial compared 85 determinant of clinical outcome that increases the risk of
patients undergoing prophylactic angioplasty within DCI. Outcome from SAH has improved, probably owing
96 h of SAH with controls who did not undergo the pro- to nimodipine treatment, early repair of the ruptured
cedure.148 Early angioplasty was associated with a reduced aneurysm, and intensive care management that is directed
need for therapeutic angioplasty several days later and at preventing adverse consequences of DCI. However,
with trends towards improved outcome, which were no treatment specifically for EBI is available, and treat-
counterbalanced by three deaths due to arterial rupture in ment of DCI is suboptimal given that many patients
the treated patients. Balloon angioplasty or superselective still develop this complication despite risky, marginally
intra-arterial infusions of vasodilator drugs are frequently effective and expensive rescue therapies. Although no
used therapeutically in clinical practice but have not been drugs have been approved for SAH in the past 25 years,
studied in randomized trials.149 promising new strategies include intrathecal drug deliv-
Vasodilator drugs that have been used to treat SAH ery, neuroprotection with drugs such as albumin, and
include papaverine, nicardipine, nimodipine, verapamil systemic administration of drugs such as NaNO2, statins,
and milrinone. Verapamil has a more negative inotropic cilostazol, dantrolene and others with pleiotrophic effects.
effect and a greater inhibitory action on the sinoatrial The design of clinical trials to test new treatments needs
node than do dihydropyridines.150 The limitations of to address other hypothesized limitations of prior studies,
intra-arterial drug infusions are that their effects are such as appropriate outcome measures, sample sizes,
short-lived, and that high doses cause hypotension. accounting for rescue therapy, and management of drug
Continuous infusion of verapamil into cerebral arteries adverse effects and practice misalignment.
for up to 21 h was performed in 12 patients with medi-
cally refractory neurological deterioration associated
Review criteria
with angiographic vasospasm.151 The investigators con-
sidered the treatment to be efficacious and did not report I searched PubMed for English-language full-text articles
any adverse effects. Other investigators abandoned this published until June 2013 using the search term
method, however, owing to high complication rates.152 “subarachnoid haemorrhage”, which returned 22,161
articles. Key words related to each section of the
A neurocritical care consensus conference reviewed the
Review where then used to identify relevant papers for
literature on endovascular treatments for DCI, and found inclusion, and the reference lists of these publications
that balloon-mediated or pharmacological angioplasty were searched for additional publications. When multiple
were reported to be effective in small, uncontrolled, papers addressed the same issue or the same authors
mostly retrospective studies, with no obvious superiority published multiple papers on a similar topic, the most
of one method or drug over another (with the exception recent publication was used. Clinicaltrials.gov also was
that papaverine may be associated with neurotoxicity).149 searched for completed and ongoing clinical trials on
subarachnoid haemorrhage.
These treatments could be considered when medical
1. Feigin, V. L., Lawes, C. M., Bennett, D. A., 4. Al-Khindi, T., Macdonald, R. L. & Schweizer, T. A. a randomised trial. Lancet 360, 1267–1274
Barker-Collo, S. L. & Parag, V. Worldwide stroke Cognitive and functional outcome after (2002).
incidence and early case fatality reported in 56 aneurysmal subarachnoid hemorrhage. Stroke 7. Etminan, N., Vergouwen, M. D., Ilodigwe, D.
population-based studies: a systematic review. 41, e519–e536 (2010). & Macdonald, R. L. Effect of pharmaceutical
Lancet Neurol. 8, 355–369 (2009). 5. Pickard, J. D. et al. Effect of oral nimodipine treatment on vasospasm, delayed cerebral
2. Lovelock, C. E., Rinkel, G. J. & Rothwell, P. M. on cerebral infarction and outcome after ischemia, and clinical outcome in patients with
Time trends in outcome of subarachnoid subarachnoid haemorrhage: British aneurysm aneurysmal subarachnoid hemorrhage:
hemorrhage: population-based study and nimodipine trial. BMJ 298, 636–642 (1989). a systematic review and meta-analysis. J. Cereb.
systematic review. Neurology 74, 1494–1501 6. Molyneux, A. et al. International Subarachnoid Blood Flow Metab. 31, 1443–1451 (2011).
(2010). Aneurysm Trial (ISAT) of neurosurgical clipping 8. Hinson, H. E. & Sheth, K. N. Manifestations
3. Taylor, T. N. et al. Lifetime cost of stroke in the versus endovascular coiling in 2143 patients of the hyperadrenergic state after acute brain
United States. Stroke 27, 1459–1466 (1996). with ruptured intracranial aneurysms: injury. Curr. Opin. Crit. Care 18, 139–145 (2012).

REVIEWS
9. Tam, A. K. et al. Impact of systemic inflammatory 25. Fisher, C. M., Kistler, J. P. & Davis, J. M. 42. Hossmann, K. A. Reperfusion of the brain after
response syndrome on vasospasm, cerebral Relation of cerebral vasospasm to subarachnoid global ischemia: hemodynamic disturbances.
infarction, and outcome after subarachnoid hemorrhage visualized by computerized Shock 8, 95–101 (1997).
hemorrhage: exploratory analysis of tomographic scanning. Neurosurgery 6, 1–9 43. Stoltenburg-Didinger, G. & Schwarz, K. in Stroke
CONSCIOUS-1 database. Neurocrit. Care 13, (1980). and Microcirculation (eds Cervos-Navarro, J. &
182–189 (2010). 26. Reilly, C., Amidei, C., Tolentino, J., Jahromi, B. S. Ferszt, R.) 471–480 (Raven Press, 1987).
10. Gao, C. et al. Relationship between sympathetic & Macdonald, R. L. Clot volume and clearance 44. Suzuki, S., Suzuki, M., Iwabuchi, T. & Kamata, Y.
nervous activity and inflammatory response rate as independent predictors of vasospasm Role of multiple cerebral microthrombosis in
after subarachnoid hemorrhage in a perforating after aneurysmal subarachnoid hemorrhage. symptomatic cerebral vasospasm: with a case
canine model. Auton. Neurosci. 147, 70–74 J. Neurosurg. 101, 255–261 (2004). report. Neurosurgery 13, 199–203 (1983).
(2009). 27. Kramer, A. H. & Fletcher, J. J. Locally- 45. Romano, J. G. et al. Microemboli in aneurysmal
11. Rosengart, A. J., Schultheiss, K. E., Tolentino, J. & administered intrathecal thrombolytics following subarachnoid hemorrhage. J. Neuroimaging 18,
Macdonald, R. L. Prognostic factors for outcome aneurysmal subarachnoid hemorrhage: 396–401 (2008).
in patients with aneurysmal subarachnoid a systematic review and meta-analysis. 46. Larsen, C. C., Hansen-Schwartz, J., Nielsen, J. D.
hemorrhage. Stroke 38, 2315–2321 (2007). Neurocrit. Care 14, 489–499 (2011). & Astrup, J. Blood coagulation and fibrinolysis
12. Starke, R. M. & Connolly, E. S. Jr. Rebleeding 28. Pluta, R. M. et al. Cerebral vasospasm following after experimental subarachnoid hemorrhage.
after aneurysmal subarachnoid hemorrhage. subarachnoid hemorrhage: time for a new world Acta Neurochir. (Wien) 152, 1577–1581 (2010).
Neurocrit. Care 15, 241–246 (2011). of thought. Neurol. Res. 31, 151–158 (2009). 47. Pisapia, J. M. et al. Microthrombosis after
13. Lo, B. W., Macdonald, R. L., Baker, A. & 29. Haley, E. C. Jr, Kassell, N. F. & Torner, J. C. experimental subarachnoid hemorrhage: time
Levine, M. A. Clinical outcome prediction in A randomized trial of nicardipine in subarachnoid course and effect of red blood cell-bound
aneurysmal subarachnoid hemorrhage using hemorrhage: angiographic and transcranial thrombin-activated pro-urokinase and
bayesian neural networks with fuzzy logic Doppler ultrasound results. A report of the clazosentan. Exp. Neurol. 233, 357–363
inferences. Comput. Math. Methods Med. 2013, Cooperative Aneurysm Study. J. Neurosurg. 78, (2012).
904860 (2013). 548–553 (1993). 48. Doczi, T., Joo, F., Adam, G., Bozoky, B. &
14. Sehba, F. A., Hou, J., Pluta, R. M. & Zhang, J. H. 30. Macdonald, R. L. et al. Clazosentan to overcome Szerdahelyi, P. Blood–brain barrier damage
The importance of early brain injury after neurological ischemia and infarction occurring during the acute stage of subarachnoid
subarachnoid hemorrhage. Prog. Neurobiol. 97, after subarachnoid hemorrhage (CONSCIOUS-1): hemorrhage, as exemplified by a new animal
14–37 (2012). randomized, double-blind, placebo-controlled model. Neurosurgery 18, 733–739 (1986).
15. Mahaney, K. B., Todd, M. M., Bayman, E. O. phase 2 dose-finding trial. Stroke 39, 49. Germano, A., d’Avella, D., Imperatore, C.,
& Torner, J. C. Acute postoperative neurological 3015–3021 (2008). Caruso, G. & Tomasello, F. Time-course of blood–
deterioration associated with surgery for 31. Vorkapic, P., Bevan, J. A. & Bevan, R. D. brain barrier permeability changes after
ruptured intracranial aneurysm: incidence, Longitudinal in vivo and in vitro time-course study experimental subarachnoid haemorrhage.
predictors, and outcomes. J. Neurosurg. 116, of chronic cerebrovasospasm in the rabbit basilar Acta Neurochir. (Wien) 142, 575–580 (2000).
1267–1278 (2012). artery. Neurosurg. Rev. 14, 215–219 (1991). 50. Friedrich, B., Muller, F., Feiler, S., Scholler, K.
16. Broderick, J. P., Brott, T. G., Duldner, J. E., 32. Crowley, R. W. et al. Angiographic vasospasm is & Plesnila, N. Experimental subarachnoid
Tomsick, T. & Leach, A. Initial and recurrent strongly correlated with cerebral infarction after hemorrhage causes early and long-lasting
bleeding are the major causes of death following subarachnoid hemorrhage. Stroke 42, 919–923 microarterial constriction and microthrombosis:
subarachnoid hemorrhage. Stroke 25, (2011). an in-vivo microscopy study. J. Cereb. Blood Flow
1342–1347 (1994). 33. Dorhout Mees, S. M. et al. Calcium antagonists Metab. 32, 447–455 (2012).
17. Macdonald, R. L., Rosengart, A., Huo, D. for aneurysmal subarachnoid hemorrhage. 51. Britz, G. W. et al. Time-dependent alterations
& Karrison, T. Factors associated with the Cochrane Database of Systematic Reviews, in functional and pharmacological arteriolar
development of vasospasm after planned Issue 3. Art. No.: CD000277. http://dx.doi.org/ reactivity after subarachnoid hemorrhage. Stroke
surgical treatment of aneurysmal subarachnoid 10.1002/14651858.CD000277.pub3. 38, 1329–1335 (2007).
hemorrhage. J. Neurosurg. 99, 644–652 34. Vergouwen, M. D., Vermeulen, M., Coert, B. A., 52. Park, K. W., Metais, C., Dai, H. B.,
(2003). Stroes, E. S. & Roos, Y. B. Microthrombosis Comunale, M. E. & Sellke, F. W. Microvascular
18. Dorhout Mees, S. M., Kerr, R. S., Rinkel, G. J., after aneurysmal subarachnoid hemorrhage: endothelial dysfunction and its mechanism
Algra, A. & Molyneux, A. J. Occurrence and an additional explanation for delayed cerebral in a rat model of subarachnoid hemorrhage.
impact of delayed cerebral ischemia after coiling ischemia. J. Cereb. Blood Flow Metab. 28, Anesth. Analg. 92, 990–996 (2001).
and after clipping in the International 1761–1770 (2008). 53. Rothoerl, R. D. & Ringel, F. Molecular
Subarachnoid Aneurysm Trial (ISAT). J. Neurol. 35. Dreier, J. P. The role of spreading depression, mechanisms of cerebral vasospasm following
259, 679–683 (2012). spreading depolarization and spreading aneurysmal SAH. Neurol. Res. 29, 636–642
19. Chou, C. H. et al. Costs of vasospasm in ischemia in neurological disease. Nat. Med. 17, (2007).
patients with aneurysmal subarachnoid 439–447 (2011). 54. Pradilla, G., Chaichana, K. L., Hoang, S.,
hemorrhage. Neurosurgery 67, 345–352 (2010). 36. Stein, S. C., Browne, K. D., Chen, X. H., Huang, J. & Tamargo, R. J. Inflammation and
20. Dorsch, N. A clinical review of cerebral Smith, D. H. & Graham, D. I. Thromboembolism cerebral vasospasm after subarachnoid
vasospasm and delayed ischaemia following and delayed cerebral ischemia after hemorrhage. Neurosurg. Clin. N. Am. 21,
aneurysm rupture. Acta Neurochir. Suppl. 110, subarachnoid hemorrhage: an autopsy study. 365–379 (2010).
5–6 (2011). 59, 781–787 (2006). 55. Macdonald, R. L. & Weir, B. K. A review of
21. Rosengart, A. J. et al. Outcome in patients with 37. Schubert, G. A., Seiz, M., Hegewald, A. A., hemoglobin and the pathogenesis of cerebral
subarachnoid hemorrhage treated with Manville, J. & Thome, C. Acute hypoperfusion vasospasm. Stroke 22, 971–982 (1991).
antiepileptic drugs. J. Neurosurg. 107, 253–260 immediately after subarachnoid hemorrhage: 56. Chaichana, K. L., Levy, A. P., Miller-Lotan, R.,
(2007). a xenon contrast-enhanced CT study. Shakur, S. & Tamargo, R. J. Haptoglobin 2–2
22. Ducruet, A. F. et al. Genetic determinants of J. Neurotrauma 26, 2225–2231 (2009). genotype determines chronic vasospasm after
cerebral vasospasm, delayed cerebral ischemia, 38. Sehba, F. A., Mostafa, G., Friedrich, V. Jr & experimental subarachnoid hemorrhage. Stroke
and outcome after aneurysmal subarachnoid Bederson, J. B. Acute microvascular platelet 38, 3266–3271 (2007).
hemorrhage. J. Cereb. Blood Flow Metab. 30, aggregation after subarachnoid hemorrhage. 57. Borsody, M., Burke, A., Coplin, W., Miller-
676–688 (2010). J. Neurosurg. 102, 1094–1100 (2005). Lotan, R. & Levy, A. Haptoglobin and the
23. Khurana, V. G. et al. Endothelial nitric oxide 39. Friedrich, V., Flores, R., Muller, A. & Sehba, F. A. development of cerebral artery vasospasm after
synthase gene polymorphisms predict Luminal platelet aggregates in functional deficits subarachnoid hemorrhage. Neurology 66,
susceptibility to aneurysmal subarachnoid in parenchymal vessels after subarachnoid 634–640 (2006).
hemorrhage and cerebral vasospasm. J. Cereb. hemorrhage. Brain Res. 1354, 179–187 (2010). 58. Peterson, J. W., Candia, G., Spanos, A. J.
Blood Flow Metab. 24, 291–297 (2004). 40. Sabri, M. et al. Mechanisms of microthrombi & Zervas, N. T. The calmodulin antagonist
24. Vergouwen, M. D. et al. Plasminogen activator formation after experimental subarachnoid trifluoperazine provides mild prophylactic
inhibitor-1 4G allele in the 4G/5G promoter hemorrhage. Neuroscience 224, 26–37 (2012). protection against cerebral vasospasm after
polymorphism increases the occurrence of 41. Friedrich, V., Flores, R., Muller, A. & Sehba, F. A. subarachnoid hemorrhage, but no therapeutic
cerebral ischemia after aneurysmal Escape of intraluminal platelets into brain value. Neurosurgery 25, 917–922 (1989).
subarachnoid hemorrhage. Stroke 35, parenchyma after subarachnoid hemorrhage. 59. German, J. W., Gross, C. E., Giclas, P., Watral, W.
1280–1283 (2004). Neuroscience 165, 968–975 (2010). & Bednar, M. M. Systemic complement depletion

REVIEWS
inhibits experimental cerebral vasospasm. 77. Koide, M., Bonev, A. D., Nelson, M. T. & 94. Vespa, P. M. et al. Early detection of vasospasm
Neurosurgery 39, 141–145 (1996). Wellman, G. C. Inversion of neurovascular after acute subarachnoid hemorrhage using
60. Kaura, V. & Bonner, S. Subarachnoid coupling by subarachnoid blood depends on continuous EEG ICU monitoring.
haemorrhage: early clinical indicators and large-conductance Ca2+-activated K+ (BK) Electroencephalogr. Clin. Neurophysiol. 103,
biomarkers. Trends Anaesth. Crit. Care 2, 42–47 channels. Proc. Natl Acad. Sci. USA 109, 607–615 (1997).
(2012). E1387–E1395 (2012). 95. Hanggi, D. Monitoring and detection of
61. Lad, S. P., Hegen, H., Gupta, G., Deisenhammer, F. 78. Leao, A. A. Spreading depression of activity in vasospasm II: EEG and invasive monitoring.
& Steinberg, G. K. Proteomic biomarker discovery the cerebral cortex. J. Neurophysiol. 7, 359–390 Neurocrit. Care 15, 318–323 (2011).
in cerebrospinal fluid for cerebral vasospasm (1944). 96. Unterberg, A. W., Sakowitz, O. W.,
following subarachnoid hemorrhage. J. Stroke 79. de Rooij, N. K., Rinkel, G. J., Dankbaar, J. W. Sarrafzadeh, A. S., Benndorf, G. & Lanksch, W. R.
Cerebrovasc. Dis. 21, 30–41 (2012). & Frijns, C. J. Delayed cerebral ischemia after Role of bedside microdialysis in the diagnosis of
62. Bavbek, M. et al. Monoclonal antibodies against subarachnoid hemorrhage: a systematic review cerebral vasospasm following aneurysmal
ICAM-1 and CD18 attenuate cerebral vasospasm of clinical, laboratory, and radiological predictors. subarachnoid hemorrhage. J. Neurosurg. 94,
after experimental subarachnoid hemorrhage in Stroke 44, 43–54 (2013). 740–749 (2001).
rabbits. Stroke 29, 1930–1935 (1998). 80. Petruk, K. C. et al. Nimodipine treatment in poor- 97. Vajkoczy, P., Horn, P., Thome, C., Munch, E.
63. Thai, Q. A., Oshiro, E. M. & Tamargo, R. J. grade aneurysm patients. Results of a & Schmiedek, P. Regional cerebral blood flow
Inhibition of experimental vasospasm in rats multicenter double-blind placebo-controlled trial. monitoring in the diagnosis of delayed ischemia
with the periadventitial administration of J. Neurosurg. 68, 505–517 (1988). following aneurysmal subarachnoid hemorrhage.
ibuprofen using controlled-release polymers. 81. Vergouwen, M. D. et al. Definition of delayed J. Neurosurg. 98, 1227–1234 (2003).
Stroke 30, 140–147 (1999). cerebral ischemia after aneurysmal 98. Macdonald, R. L. et al. Subarachnoid Hemorrhage
64. Tiebosch, I. A. et al. Effect of interferon-β on subarachnoid hemorrhage as an outcome event International Trialists data repository (SAHIT).
neuroinflammation, brain injury and neurological in clinical trials and observational studies: World Neurosurg. 79, 418–422 (2013).
outcome after experimental subarachnoid proposal of a multidisciplinary research group. 99. Macdonald, R. L. et al. Clazosentan, an
hemorrhage. Neurocrit. Care 18, 96–105 (2013). Stroke 41, 2391–2395 (2010). endothelin receptor antagonist, in patients
65. Feigin, V. et al. Corticosteroids for aneurysmal 82. Diringer, M. N. et al. Critical care management of with aneurysmal subarachnoid haemorrhage
subarachnoid haemorrhage and primary patients following aneurysmal subarachnoid undergoing surgical clipping: a randomised,
intracerebral haemorrhage. Cochrane Database hemorrhage: recommendations from the double-blind, placebo-controlled phase 3 trial
of Systematic Reviews, Issue 3. Art. No.: Neurocritical Care Society’s Multidisciplinary (CONSCIOUS-2). Lancet Neurol. 10, 618–625
CD004583. http://dx.doi.org/10.1002/ Consensus Conference. Neurocrit. Care 15, (2011).
14651858.CD004583.pub2. 211–240 (2011). 100. Macdonald, R. L. et al. Randomized trial of
66. Gomis, P. et al. Randomized, double-blind, 83. Rabinstein, A. A., Weigand, S., Atkinson, J. L. clazosentan in patients with aneurysmal
placebo-controlled, pilot trial of high-dose & Wijdicks, E. F. Patterns of cerebral infarction subarachnoid hemorrhage undergoing
methylprednisolone in aneurysmal subarachnoid in aneurysmal subarachnoid hemorrhage. Stroke endovascular coiling. Stroke 43, 1463–1469
hemorrhage. J. Neurosurg. 112, 681–688 36, 992–997 (2005). (2012).
(2010). 84. Schmidt, J. M. et al. Frequency and clinical 101. Vergouwen, M. D., Algra, A. & Rinkel, G. J.
67. Vergouwen, M. D., de Haan, R. J., Vermeulen, M. impact of asymptomatic cerebral infarction due Endothelin receptor antagonists for aneurysmal
& Roos, Y. B. Effect of statin treatment on to vasospasm after subarachnoid hemorrhage. subarachnoid hemorrhage: a systematic review
vasospasm, delayed cerebral ischemia, and J. Neurosurg. 109, 1052–1059 (2008). and meta-analysis update. Stroke 43,
functional outcome in patients with aneurysmal 85. Vergouwen, M. D., Ilodigwe, D. & Macdonald, R. L. 2671–2676 (2012).
subarachnoid hemorrhage: a systematic review Cerebral infarction after subarachnoid 102. Yamaguchi, S. et al. Involvement of Rho-kinase
and meta-analysis update. Stroke 41, e47–e52 hemorrhage contributes to poor outcome by in tumor necrosis factor-α-induced interleukin-6
(2010). vasospasm-dependent and -independent effects. release from C6 glioma cells. Neurochem. Int.
68. Tseng, M. Y. Summary of evidence on immediate Stroke 42, 924–929 (2011). 55, 438–445 (2009).
statins therapy following aneurysmal 86. Crobeddu, E. et al. Predicting the lack of 103. Satoh, S. et al. Amelioration of endothelial
subarachnoid hemorrhage. Neurocrit. Care 15, development of delayed cerebral ischemia after damage/dysfunction is a possible mechanism
298–301 (2011). aneurysmal subarachnoid hemorrhage. Stroke for the neuroprotective effects of Rho-kinase
69. Lo, E. H., Dalkara, T. & Moskowitz, M. A. 43, 697–701 (2012). inhibitors against ischemic brain damage.
Mechanisms, challenges and opportunities in 87. Chaudhary, S. R. et al. Prospective evaluation of Brain Res. Bull. 81, 191–195 (2010).
stroke. Nat. Rev. Neurosci. 4, 399–415 (2003). multidetector-row CT angiography for the 104. Liu, G. J. et al. Systematic assessment and
70. Iadecola, C. & Anrather, J. The immunology diagnosis of vasospasm following subarachnoid meta-analysis of the efficacy and safety of
of stroke: from mechanisms to translation. hemorrhage: a comparison with digital fasudil in the treatment of cerebral vasospasm
Nat. Med. 17, 796–808 (2011). subtraction angiography. Cerebrovasc. Dis. 25, in patients with subarachnoid hemorrhage.
71. Park, S. et al. Neurovascular protection reduces 144–150 (2008). Eur. J. Clin. Pharmacol. 68, 131–139 (2011).
early brain injury after subarachnoid 88. Wintermark, M. et al. Vasospasm after 105. Li, F. et al. The protective effect of dantrolene on
hemorrhage. Stroke 35, 2412–2417 (2004). subarachnoid hemorrhage: utility of perfusion CT ischemic neuronal cell death is associated with
72. Sasaki, T., Kassell, N. F., Yamashita, M., and CT angiography on diagnosis and reduced expression of endoplasmic reticulum
Fujiwara, S. & Zuccarello, M. Barrier disruption management. AJNR Am. J. Neuroradiol. 27, stress markers. Brain Res. 1048, 59–68 (2005).
in the major cerebral arteries following 26–34 (2006). 106. Muehlschlegel, S., Rordorf, G. & Sims, J. Effects
experimental subarachnoid hemorrhage. 89. Dankbaar, J. W. et al. Diagnosing delayed cerebral of a single dose of dantrolene in patients with
J. Neurosurg. 63, 433–440 (1985). ischemia with different CT modalities in patients cerebral vasospasm after subarachnoid
73. Yatsushige, H., Ostrowski, R. P., Tsubokawa, T., with subarachnoid hemorrhage with clinical hemorrhage: a prospective pilot study. Stroke
Colohan, A. & Zhang, J. H. Role of c-Jun deterioration. Stroke 40, 3493–3498 (2009). 42, 1301–1306 (2011).
N-terminal kinase in early brain injury after 90. Fontanella, M. et al. Vasospasm after SAH due 107. Pluta, R. M. New regulatory, signaling pathways,
subarachnoid hemorrhage. J. Neurosci. Res. 85, to aneurysm rupture of the anterior circle of and sources of nitric oxide. Acta Neurochir.
1436–1448 (2007). Willis: value of TCD monitoring. Neurol. Res. 30, Suppl. 110, 7–12 (2011).
74. Yan, J. et al. Blood–brain barrier disruption 256–261 (2008). 108. Oldfield, E. H. et al. Safety and pharmacokinetics
following subarachnoid hemorrhage may be 91. Budohoski, K. P. et al. Impairment of cerebral of sodium nitrite in patients with subarachnoid
facilitated through PUMA induction of endothelial autoregulation predicts delayed cerebral hemorrhage: a phase IIA study. J. Neurosurg.
cell apoptosis from the endoplasmic reticulum. ischemia after subarachnoid hemorrhage: 119, 634–641 (2013).
Exp. Neurol. 230, 240–247 (2011). a prospective observational study. Stroke 43, 109. Omeis, I., Jayson, N. A., Murali, R. &
75. Doczi, T. The pathogenetic and prognostic 3230–3237 (2012). Abrahams, J. M. Treatment of cerebral
significance of blood–brain barrier damage at 92. Kistka, H., Dewan, M. C. & Mocco, J. Evidence- vasospasm with biocompatible controlled-
the acute stage of aneurysmal subarachnoid based cerebral vasospasm surveillance. Neurol. release systems for intracranial drug delivery.
haemorrhage. Clinical and experimental studies. Res. Int. 2013, 256713 (2013). Neurosurgery 63, 1011–1019 (2008).
Acta Neurochir. 77, 110–132 (1985). 93. Claassen, J., Mayer, S. A. & Hirsch, L. J. 110. Barth, M. et al. Effect of nicardipine prolonged-
76. Yang, Y. & Rosenberg, G. A. Blood–brain barrier Continuous EEG monitoring in patients with release implants on cerebral vasospasm and
breakdown in acute and chronic cerebrovascular subarachnoid hemorrhage. J. Clin. Neurophysiol. clinical outcome after severe aneurysmal
disease. Stroke 42, 3323–3328 (2011). 22, 92–98 (2005). subarachnoid hemorrhage: a prospective,

REVIEWS
randomized, double-blind phase IIa study. Stroke 127. Mees, S. M. et al. Magnesium for aneurysmal 143. Dankbaar, J. W., Slooter, A. J., Rinkel, G. J.
38, 330–336 (2007). subarachnoid haemorrhage (MASH-2): & van der Schaaf, I. C. Effect of different
111. Krischek, B., Kasuya, H., Onda, H. & Hori, T. a randomised placebo-controlled trial. Lancet components of triple-H therapy on cerebral
Nicardipine prolonged-release implants for 380, 44–49 (2012). perfusion in patients with aneurysmal
preventing cerebral vasospasm after 128. Muroi, C., Terzic, A., Fortunati, M., Yonekawa, Y. subarachnoid haemorrhage: a systematic
subarachnoid hemorrhage: effect and outcome & Keller, E. Magnesium sulfate in the review. Crit. Care 14, R23 (2010).
in the first 100 patients. Neurol. Med. Chir. management of patients with aneurysmal 144. Muench, E. et al. Effects of hypervolemia and
(Tokyo) 47, 389–394 (2007). subarachnoid hemorrhage: a randomized, hypertension on regional cerebral blood flow,
112. Dorhout Mees, S. M., van den Bergh, W. M., placebo-controlled, dose-adapted trial. intracranial pressure, and brain tissue
Algra, A. & Rinkel, G. J. Antiplatelet therapy for Surg. Neurol. 69, 33–39 (2008). oxygenation after subarachnoid hemorrhage.
aneurysmal subarachnoid haemorrhage. 129. Wong, G. K. Magnesium for aneurysmal Crit. Care Med. 35, 1844–1851 (2007).
Cochrane Database of Systematic Reviews, subarachnoid haemorrhage. Lancet 380, 1381 145. Raabe, A. et al. Relative importance of
Issue 4. Art. No.: CD006184. http://dx.doi.org/ (2012). hypertension compared with hypervolemia for
10.1002/14651858.CD006184.pub2. 130. McKee, J. A. et al. Magnesium neuroprotection increasing cerebral oxygenation in patients with
113. Cronqvist, M. et al. Diffusion and perfusion MRI is limited in humans with acute brain injury. cerebral vasospasm after subarachnoid
in patients with ruptured and unruptured Neurocrit. Care 2, 342–351 (2005). hemorrhage. J. Neurosurg. 103, 974–981
intracranial aneurysms treated by endovascular 131. Mori, K. et al. Initial clinical experience of (2005).
coiling: complications, procedural results, MR vasodilatory effect of intra-cisternal infusion of 146. Hasan, D., Vermeulen, M., Wijdicks, E. F.,
findings and clinical outcome. Neuroradiology magnesium sulfate for the treatment of cerebral Hijdra, A. & van Gijn, J. Effect of fluid intake and
47, 855–873 (2005). vasospasm after aneurysmal subarachnoid antihypertensive treatment on cerebral ischemia
114. van den Bergh, W. M., Kerr, R. S., Algra, A., hemorrhage. Neurol. Med. Chir. (Tokyo) 49, after subarachnoid hemorrhage. Stroke 20,
Rinkel, G. J. & Molyneux, A. J. Effect of 139–144 (2009). 1511–1515 (1989).
antiplatelet therapy for endovascular coiling in 132. Manno, E. M., Gress, D. R., Ogilvy, C. S., 147. Zubkov, Y. N., Nikiforov, B. M. & Shustin, V. A.
aneurysmal subarachnoid hemorrhage. Stroke Stone, C. M. & Zervas, N. T. The safety and Balloon catheter technique for dilatation of
40, 1969–1972 (2009). efficacy of cyclosporine A in the prevention of constricted cerebral arteries after aneurysmal
115. Suarez, J. I. et al. Effect of human albumin vasospasm in patients with Fisher grade 3 SAH. Acta Neurochir. (Wien) 70, 65–79 (1984).
administration on clinical outcome and hospital subarachnoid hemorrhages: a pilot study. 148. Zwienenberg-Lee, M. et al. Effect of prophylactic
cost in patients with subarachnoid hemorrhage. Neurosurgery 40, 289–293 (1997). transluminal balloon angioplasty on cerebral
J. Neurosurg. 100, 585–590 (2004). 133. Yanamoto, H. et al. Therapeutic trial of cerebral vasospasm and outcome in patients with Fisher
116. Suarez, J. I. et al. The Albumin in Subarachnoid vasospasm with the serine protease inhibitor, grade III subarachnoid hemorrhage: results of a
Hemorrhage (ALISAH) multicenter pilot clinical FUT-175, administered in the acute stage after phase II multicenter, randomized, clinical trial.
trial: safety and neurologic outcomes. Stroke 43, subarachnoid hemorrhage. Neurosurgery 30, Stroke 39, 1759–1765 (2008).
683–690 (2012). 358–363 (1992). 149. Kimball, M. M., Velat, G. J. & Hoh, B. L. Critical
117. Springborg, J. B. et al. A single subcutaneous 134. Young, A. M., Karri, S. K. & Ogilvy, C. S. Non- care guidelines on the endovascular
bolus of erythropoietin normalizes cerebral steroidal anti-inflammatory drugs used as a management of cerebral vasospasm.
blood flow autoregulation after subarachnoid treatment modality in subarachnoid hemorrhage. Neurocrit. Care 15, 336–341 (2011).
haemorrhage in rats. Br. J. Pharmacol. 135, Curr. Drug Saf. 7, 197–201 (2012). 150. DeWitt, C. R. & Waksman, J. C. Pharmacology,
823–829 (2002). 135. Klimo, P. Jr, Kestle, J. R., MacDonald, J. D. & pathophysiology and management of calcium
118. Grasso, G. et al. Beneficial effects of systemic Schmidt, R. H. Marked reduction of cerebral channel blocker and beta-blocker toxicity.
administration of recombinant human vasospasm with lumbar drainage of cerebrospinal Toxicol. Rev. 23, 223–238 (2004).
erythropoietin in rabbits subjected to fluid after subarachnoid hemorrhage. 151. Albanese, E. et al. Ultrahigh-dose intraarterial
subarachnoid hemorrhage. Proc. Natl Acad. Sci. J. Neurosurg. 100, 215–224 (2004). infusion of verapamil through an indwelling
USA 99, 5627–5631 (2002). 136. Al-Tamimi, Y. Z. et al. Lumbar drainage of microcatheter for medically refractory severe
119. Tsai, P. T. et al. A critical role of erythropoietin cerebrospinal fluid after aneurysmal vasospasm: initial experience. J. Neurosurg.
receptor in neurogenesis and post-stroke subarachnoid hemorrhage: a prospective, 113, 913–922 (2010).
recovery. J. Neurosci. 26, 1269–1274 (2006). randomized, controlled trial (LUMAS). Stroke 43, 152. Volk, J. M., Culicchia, F. & Dawson, R. Use of
120. Tseng, M. Y. et al. Acute systemic erythropoietin 677–682 (2012). intra-arterial infusion of calcium-channel
therapy to reduce delayed ischemic deficits 137. Senbokuya, N. et al. Effects of cilostazol on blockers through an indwelling microcatheter:
following aneurysmal subarachnoid hemorrhage: cerebral vasospasm after aneurysmal one institution’s experience. Stroke 43, A47
a phase II randomized, double-blind, placebo- subarachnoid hemorrhage: a multicenter (2012).
controlled trial. Clinical article. J. Neurosurg. prospective, randomized, open-label blinded end 153. Bulters, D. O. et al. A randomized controlled trial
111, 171–180 (2009). point trial. J. Neurosurg. 118, 121–130 (2013). of prophylactic intra-aortic balloon
121. Springborg, J. B. et al. Erythropoietin in patients 138. Connolly, E. S. Jr et al. Guidelines for the counterpulsation in high-risk aneurysmal
with aneurysmal subarachnoid haemorrhage: management of aneurysmal subarachnoid subarachnoid hemorrhage. Stroke 44, 224–226
a double blind randomised clinical trial. hemorrhage: a guideline for healthcare (2013).
Acta Neurochir. (Wien) 149, 1089–1101 (2007). professionals from the American Heart 154. Wong, G. K., Chan, M. T., Gin, T. & Poon, W. S.
122. Liao, J. K. & Laufs, U. Pleiotropic effects of Association/American Stroke Association. Intravenous magnesium sulfate after
statins. Annu. Rev. Pharmacol. Toxicol. 45, Stroke 43, 1711–1737 (2012). aneurysmal subarachnoid hemorrhage: current
89–118 (2005). 139. Le Roux, P. D. Anemia and transfusion after status. Acta Neurochir. Suppl. 110, 169–173
123. US National Library of Medicine. ClinicalTrials.gov subarachnoid hemorrhage. Neurocrit. Care 15, (2011).
[online], http://clinicaltrials.gov/show/ 342–353 (2011). 155. Ryba, M., Pastuszko, M., Iwanska, K.,
NCT00731627 (2011). 140. Treggiari, M. M. Hemodynamic management of Bidzinski, J. & Dziewiecki, C. Cyclosporine A
124. Wong, G. K. et al. High-dose simvastatin for subarachnoid hemorrhage. Neurocrit. Care 15, prevents neurological deterioration of patients
aneurysmal subarachnoid hemorrhage: 329–335 (2011). with SAH--a preliminary report. Acta Neurochir.
a multicenter, randomized, controlled, double- 141. Dhar, R. et al. Comparison of induced 112, 25–27 (1991).
blind clinical trial protocol. Neurosurgery 72, hypertension, fluid bolus, and blood transfusion
Acknowledgements
840–844 (2013). to augment cerebral oxygen delivery after
The author receives grant support from the Physicians
125. US National Library of Medicine. ClinicalTrials.gov subarachnoid hemorrhage. J. Neurosurg. 116,
Services Incorporated Foundation, the Brain Aneurysm
[online], http://clinicaltrials.gov/show/ 648–656 (2012).
Foundation, the Canadian Stroke Network, and the
NCT01077206 (2013). 142. Lennihan, L. et al. Effect of hypervolemic therapy
Heart and Stroke Foundation of Ontario.
126. Euser, A. G. & Cipolla, M. J. Magnesium sulfate on cerebral blood flow after subarachnoid
for the treatment of eclampsia: a brief review. hemorrhage: a randomized controlled trial. Supplementary information is linked to the online
Stroke 40, 1169–1175 (2009). Stroke 31, 383–391 (2000). version of the paper at www.nature.com/nrneurol.

REVIEWS
CORRECTION
Delayed neurological deterioration after subarachnoid haemorrhage
R. Loch Macdonald
Nat. Rev. Neurol. advance online publication 10 December 2013; doi:10.1038/nrneurol.2013.246
In the version of this article initially published online, the figure legend for part e
of Figure 4 was omitted. The error has been corrected for the print, HTML and PDF
versions of the article.
NATURE REVIEWS | NEUROLOGY VOLUME 10 | JANUARY 2014 | i


Nature-Delayed Neurological Deterioration After Subarachnoid Haemorrhage

Uploaded by

Copyright:

Available Formats

Nature-Delayed Neurological Deterioration After Subarachnoid Haemorrhage

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nature-Delayed Neurological Deterioration After Subarachnoid Haemorrhage

Uploaded by

Copyright:

Available Formats

REVIEWS

Delayed neurological deterioration after

44 | JANUARY 2014 | VOLUME 10 www.nature.com/nrneurol

Key points circulation aneurysm (Figure 1).11 Aneurysm rebleed-

Early brain injury Normal

Swollen neuron Basal

NATURE REVIEWS | NEUROLOGY VOLUME 10 | JANUARY 2014 | 45

Disrupted Microcirculatory Angiographic vasospasm

Microthrombus Cortical spreading ischaemia

Spreading depressions, neuronal swelling, distorted

46 | JANUARY 2014 | VOLUME 10 www.nature.com/nrneurol

NATURE REVIEWS | NEUROLOGY VOLUME 10 | JANUARY 2014 | 47

a b c d e Cortical arteriole constriction in response to endothelin-1

48 | JANUARY 2014 | VOLUME 10 www.nature.com/nrneurol

polymerase, or administration of monoclonal antibodies in an inverse haemodynamic response (possibly due

NATURE REVIEWS | NEUROLOGY VOLUME 10 | JANUARY 2014 | 49

50 | JANUARY 2014 | VOLUME 10 www.nature.com/nrneurol

Table 1 | Drugs in development for treatment of SAH

Antithrombotic actions At least two ongoing clinical trials123,125

Anticonvulsant (N-methyl-D-aspartate receptor antagonism)

Corticosteroids have multiple anti-inflammatory actions, mostly

NATURE REVIEWS | NEUROLOGY VOLUME 10 | JANUARY 2014 | 51

52 | JANUARY 2014 | VOLUME 10 www.nature.com/nrneurol

NATURE REVIEWS | NEUROLOGY VOLUME 10 | JANUARY 2014 | 53

Initial treatment Avoid

Low risk High risk High risk

Neurological assessment every hour, fluid

No change neurologically, normal TCD, normal CTP

Low risk High risk

Detect causes of neurological deterioration Consider transfer to lower level

Correct identified causes (for example, seizures,

Induced hypertension, fluid boluses Consider catheter angiography and balloon

54 | JANUARY 2014 | VOLUME 10 www.nature.com/nrneurol

NATURE REVIEWS | NEUROLOGY VOLUME 10 | JANUARY 2014 | 55

56 | JANUARY 2014 | VOLUME 10 www.nature.com/nrneurol

NATURE REVIEWS | NEUROLOGY VOLUME 10 | JANUARY 2014 | 57

58 | JANUARY 2014 | VOLUME 10 www.nature.com/nrneurol

NATURE REVIEWS | NEUROLOGY VOLUME 10 | JANUARY 2014 | i

You might also like