Child’s Nerv Syst (1997) 13: 595–600
© Springer-Verlag 1997

Eric Manceau
Maurice Giroud
Raymond Dumas

Received: 2 April 1997

E. Manceau (½) · M. Giroud · R. Dumas
Consultation de Neurologie Infantile,
Service de Neurologie,
Centre Hospitalier Universitaire,
3, Rue du Faubourg Raines,
F-21033 Dijon Cedex, France

ORIGINAL PAPER

Moyamoya disease in children
A review of the clinical and radiological features
and current treatment

Abstract Two cases of moyamoya
disease observed in two children are
reported. The two cases recall the
frequency of ischemic strokes, transient ischemic attacks, and seizures
revealing the disease. Diagnosis is
made by cerebral arteriography
showing carotid stenosis and the
dense deep arterial collateral as seen
in these two cases. Magnetic resonance imaging makes it possible to
identify this vascular disease as it reveals vascular varicosities in the basal ganglia, as in case 1. Strokes can

Introduction

Moyamoya means ‘‘spirals of smoke’’ in Japanese. This
term was used to name a cerebrovascular disease characterized by a particular angiographical appearance of the cerebral collateral network that it causes. This condition consists in progressive stenosis of both terminal internal carotid arteries in their supraclinoid part, with the development of a network of cerebral collaterals, which are what
define the condition as a specific pathologic entity. Starting with two neuropediatric cases, we give a theoretical
summary of this disease, describe information contributed
by recent radiological techniques, and suggest some pathophysiological hypotheses.

Case reports
Case 1
This girl is an adopted child of Korean origin, whose natural parents
are first cousins. Her first years of life passed without incident, with
normal food, motor and sphincter development, but a slight language

sometimes be prevented by constructing an anastomosis between the
superficial temporal artery and the
middle cerebral artery, as demonstrated in case 1. Thus, the diagnosis
of moyamoya disease has been improved by magnetic resonance imaging and there is now a surgical treatment to prevent ischemic stroke.
Key words Moyamoya disease ·
Carotid stenosis · Deep arterial
collaterals

delay was noted, in contrast with apparently good comprehension.
When she was 3 years old her adoptive parents were surprised to note
that she was starting to have repeated brief episodes of confusion,
mainly on waking.
When she was 4 years old, she was admitted to hospital following several generalized tonic-clonic epileptic fits, followed by a partial clonic motor fit affecting the left arm, which was complicated
by a homolateral hemiparesis that recovered in 45 min. Waking and
sleeping electro-encephalograms (EEG) and CT scan with and without injection of contrast medium did not show any specific features.
However, T2-weighted magnetic resonance imaging (MRI) gave a
picture with a right frontal hyposignal and two nodules with a hypersignal in the white matter of the right lobe. It was decided just to
monitor these abnormalities, since there were no signs of a spaceoccupying lesion. At 5 years old, the child started having sudden
weakness in the legs and to complain of migraine-like headaches. In
view of severe episodes of hyperactivity, normothymic anticonvulsant treatment with carbamazepine (CBZ) was begun. At that time,
she developed a state of neglect of the right side, which forced her
to become left-handed. She continued to have atonic fits which made
her fall down, whose posticteric phase seemed to get longer and longer. The EEG showed a slowing of the background rhythm associated with large slow bifrontal waves, predominantly on the left, which
suggested that the lesion was progressing.
When she was 6 years old, she was readmitted to hospital for investigation, because she woke up with right faciobrachial hemiplegia, which disappeared totally in 1 month. A second cerebral MRI
scan showed diffuse heterogeneous hypersignal, associated with posterior cortical and subcortical ischemic plaques with thinning of the

596

Fig. 1 Case 1: T1-weighted
MRI with horizontal view:
posterior cortical and subcortical hypersignal, associated with
deep arterial varicosities within
basal ganglia
Fig. 2 Case 1: T1-weighted
MRI with frontal view showing
the deep arterial varicosities
ehanced by gadolinium

Fig. 4 Case 2: arteriography showing the terminal carotid artery
stenosis with deep arterial collaterals characteristic of moyamoya
disease

Fig. 3 Case 1: arteriography showing the terminal carotid artery
stenosis with deep arterial collaterals characteristic of moyamoya
disease

cortical layer in the same area, and in particular tortuous images
which were similar to arterial varicosities of the central gray nuclei
(Figs. 1, 2). This typical radiological appearance, together with the
clinical picture, allowed the diagnosis of moyamoya disease to be
made, which was confirmed by a cervico-cerebral angiogram
(Fig. 3). The internal carotid arteries had a classic appearance as far
as the supraclinoid part, where they ended as a narrow stenosis, before giving off narrow anterior and middle cerebral arteries; there
was also markedly abnormal development of the basal ganglia perforating arteries. A left temporo-sylvian artery anastomosis was performed on the left by angio-myo-duro-synangiogenesis, followed

6 months later by the same procedure on the right. Since then, the
patient, now 7 years old, has had no more migraine or epileptic fits;
she shows good psychomotor development and has only slight underusage of her right hand, which remains apraxo-astereognosic. A control angiogram showed perfect collateral formation by the external
carotid network of the autograft, although the internal carotid abnormality continued to progress insidiously to a late stage, without causing the child any harm.
Case 2
The birth of this child of western Caucasian origin was uncomplicatd and his parents do not have a history of any particular hereditary diseases. When he was 3 months old his parents noted that he
was greatly agitated during the night, with a clonic epileptic fit involving the right leg and causing his head and gaze to turn to the left.
When the fit had finished he had right hemiparesis, predominantly

597

facio-brachial, associated with axial hypotonia. There was no fever
or increased pressure at the fontanel. Retinal examination and lumbar puncture were normal, but the EEG showed permanent low voltage of the left hemisphere. A cerebral CT scan showed a large hypodense area in the left fronto temporoparietal region, demonstrating a massive sylvian and left anterior cerebral ischemic cerebrovascular stroke. A left carotid angiogram showed there was a narrow
stenosis of the presylvian internal carotid, as well as an abnormal
view of a lenticulo-striated collateral network at the level of the basal ganglia (Fig. 4) allowing the diagnosis of moyamoya disease.
When the boy was 4 months old psychomotor retardation was noted, with the absence of spontaneous grasping associated with alternating convulsions. CT scan revealed generalised cerebral atrophy,
predominantly in the left hemisphere, associated with severe bilateral ventricular dilatation. A right carotid angiogram revealed that
there was also partial stenosis of the right carotid siphon.
When he was 1 year old, paresis of the left arm and axial hypotonus persisted with microcephaly. His waking seemed to be better,
and his psychomotor developmental age was estimated as equivalent
to 6 months. The EEG showed a focus of left-sided frontorolandic
spikes, and an overall angiographic examination (supra-aortic vessels, coronaries and renal arteries) was asked for in view of the recent discovery of associated labile hypertension with systolic blood
pressure of 170 mmHg and diastolic blood pressure of 100 mmHg.
These tests were all normal, as were a temporal artery biopsy and the
renal function. It was decided that this patient was not suitable for
any neurosurgical operation using a temporo-sylvian shunt, in view
of the length of time he had had the symptoms, their severity, and
the fact that a slight psychomotor improvement had been seen. Nevertheless, the outcome has been marked by mental retardation and
drug-resistant frontal epilepsy.

Discussion

Moyamoya (MM) is a rare occlusive arterial condition,
which causes ischemic strokes in young people, for which
precise classification was only made possible by the advent of cerebral angiography, and which was long considered as an exclusively Japanese disease. This disease was
first described in Japan [18], characterized by progressive
stenosis of both supra-clinoid internal carotids on cerebral
arteriography, associated with abnormal collateral circulation within the basal ganglia. Then, more than 100 cases
from all over the world were collected [12]. In France, Picard et al. [16] appears to have described the largest European series with his 24 cases. By 1979, it had become a
real public health problem in Japan, needing an official
definition in an annual research report commissioned by
the Ministry of Health: “Occlusive disease of the vessels
at the basal ganglia characterized by the association of an
angiographic picture of bilateral stenosis of the supraclinoid carotid arteries and/or their main terminal branches,
and abnormal development of a collateral network around
the circle of Willis.” In 1989, Kitamura et al. [12] stressed
that a primary form existed with an ethnic preference, and
that this form could be called moyamoya disease, and that
there were also secondary forms, which he excluded from
the definition. This distinction seems to us to be fundamental, even though overall, the clinical features, evolution and
the principles of treatment are similar.

Our patient in the first case is a child of Korean origin.
In Korea, the annual incidence of the disease has been studied over a period of 18 years and seems to be half that in
Japan [20] [0.08 (0.072–0.129) per 100,000 population],
with no regional preference. Around 150 new cases per
year are seen by the Japanese, compared with fewer than
3 in France. Yonekawa also noted that Africans were overrepresented amongst Western patients and also have a large
predominance of secondary forms compared with the
Oriental cases. Our second patient continues to be the
youngest case in the world, and was described by Israel et
al. in 1982 [8], when he was only 2 years old. The regular
reports from the Japanese Ministry of Health and the numerous publications or theses show that there are two age
peaks for the time of diagnosis, viz the 1st and 4th decades
of life. The most frequent age at which the condition comes
to light is 6 years old. The number of secondary forms increases in proportion to age, but cases described after
50 years of age are rare, and to our knowledge none have
been described after the age of 60. The sex ratio shows a
slight female predominance (60%). Only 7% of cases are
familial, but the risk is 40 times greater in children with a
first-degree relative affected by the disease. It should be
recalled that the natural parents of our first patient were
first cousins, although we do not have any more information on their family histories. It seems to us to be important to look for any factors responsible for occlusion of the
vasa vasorum (e.g. drepanocytosis or thalassemia) [17], for
enveloping fibrosis of the vessels at the basal ganglia (pachymeningitis, tumors of the brain stem or, especially, cerebral irradiation) and conditions affecting the arterial wall
(systemic elastorrhexis, fibromuscular dysphasia, glycogenesis or homocystinuria), as these may be secondary, potentially curable, forms. Our second patient later also developed Raynaud disease and unexplained early hypertension, which are without doubt part of a multisystem vascular condition. Whatever the causative disease is, close
and long-term monitoring seems to be advisable, even if a
specific treatment seems to have been spectacularly effective, because a relapse is always possible [10]. Apart from
these supposedly secondary forms, the great majority seem
to be primary, as in our first case. This would explain the
ethnic predilection, the increased risk when a family history is present, the forms that are associated with genetically determined diseases, such as phakomatosis, trisomy
21, Turner syndrome, Fanconi disease or multiple malformation syndromes (congenital cataract, transposition of
the great vessels, coarctation of the aorta, polycystic renal
disease, or fibromuscular dysplasia of the renal arteries).
The presence of primary moyamoya disease is revealed
mostly by ischemic strokes or by their immediate or late
complications. Since there is progressive bilateral stenosis, sometimes asymmetrical at the onset, of the end of the
carotids with secondary development of a collateral network through perforating and choroid arteries and also
through the transcortical and transdural networks, it is easy

598

to understand, from a time-scale point of view, why it is
particularly transient ischemic attacks (TIAs) that reveal
moyamoya disease in children [14], and hemorrhagic
strokes in adults. Repeated TIAs, sometimes alternating,
in particular during exercise, are the classic way that the
disease starts [4]. All types of segmental motor deficit have
been seen, as have several descriptions of abnormal movements, according to Suzuki et al. [18] (dyskinesia of the
limbs, chorea and hemiballismus). As in our two cases, the
disease may be revealed by epileptic fits, which are mainly
partial but may later become generalized. Moyamoya represents 12% of pediatric cases and may of course initially
be confused with idiopathic or benign epilepsy and treated
as such [12]. We also see it as important to insist on the
fact that symptoms such as severe headaches that appear
to be migrainous, like those our first patient complained
of, are undoubtedly the result of rapid cerebral hemodynamic alterations with their many associated symptoms,
such as slowing down of mental function or irritability, as
her parents had noted [6, 9]. Finally, as with all causes of
chronic ischemia during childhood development, the absence of curative treatment unavoidably leads to progressive mental deterioration, as in case 2. More than half of
these children will have at least three strokes during their
lives, bearing in mind that ischemic attacks in childhood
generally herald hemorrhagic strokes in adulthood. These
strokes should in theory be less serious than other types of
intracranial hemorrhage, since they are caused by rupture
of pseudo-aneurysms that have developed on small vessels, but in fact they are the main cause of mortality. Thus,
any subarachnoid hemorrhage occurring in this context requires a search for an associated true aneurysm. More than
50% of the patients remain independent in their daily lives;
25% have a greater or lesser degree of residual functional
disability, but the complications may cause disabilities or
may lead to true encephalopathy by way of chronic ischemia and its complications. Apart the disease manifesting
itself by epileptic fits, Kodama et al. [13] described the
characteristics of the EEG with long slow waves indicating a subcortical origin, which were often posterior, that is
in the areas where the blood flow was spared, disappeared
when the eyes were opened and were greatly reduced during sleep, when the spindles had a low voltage. On hyperventilation testing, the slow waves persisted abnormally
for 20–60 s after the end of hyperventilation, sometimes
accompanied by mental confusion. This pattern is only
found in around 80% of pediatric forms; it has not been
found in any adult cases. Kodama et al. [13] tried to explain the EEG abnormalities and studied its progression
with time in a prospective study on 25 children who were
followed up for 72 months from the start of their symptoms. Thanks to this study, we now know that it takes
10 months for the large wide posterior waves to appear,
20 months for them to be found also in the centrotemporal areas and to be less reactive, and 56 months for there
to be diffuse low-amplitude activity. The natural progres-

sion of the disease can thus be assessed indirectly, because
the posterior slow waves correspond to the expression of
deeper activity in a healthy cortical area, whereas their disappearance is due to the degree of ischemia becoming too
severe and/or to a superadded deep lesion. As for hyperventilation, this produces hypocapnia, which causes widespread vasoconstriction, and this naturally affects small
vessels first of all, so involving the network of telangiectases in the brainstem. This could explain the repeated confusional states and the paroxysmal deficits of the lower
limbs by vasoconstriction of the anterior cerebral arteries
during physical exertion or crying, as described in our two
cases. Fujiwara et al. [4], thanks to EEG recordings on ten
patients with simultaneous assessment of their arterial
blood gases, recently proved that administration of O2 soon
after hyperventilation makes it possible to avoid the rebuild-up phenomenon and, especially, its accompanying
symptoms which according to him, may go as far as producing TIAs. Iwama et al. [9], investigating a series of 124
children, have also just proved the importance of maintaining normocapnia and normal blood pressure during the
whole of the perioperative period. TIAs are, for them, an
index of an intracranial hemodynamic lesion and of potential peri- or postoperative complications. Moyamoya owes
its definition and its classification to the development of
angiographic techniques. It is a progressive condition, as
our two cases show, which has radiological features that
are now well known and which allowed Suzuki et al. [18]
to classify it into six successive stages as early as 1966. In
any case, the radiological features recognized as diagnostic criteria remain: the presence of stenoses or occlusions
of the two intracranial internal carotid arteries and/or of
the start of the sylvian or anterior cerebral arteries, the presence of a network of abnormally dilated, anastomotic perforating arteries, the presence of collaterals from the arteries of the external carotid network in place of the supraclinoid internal carotid vessels, thanks to abnormally developed transdural anastomoses, and finally transcortical collaterals with arteries from the vertebro-basilar network. It
is important to note that primary moyamoya disease always
presents with an asymmetrical stenosis, but that the unilateral forms (1/3 of cases) become bilateral later. Handa et
al. [5] described four possible sites for these stenoses: these
are, distal to the ophthalmic artery, between the anterior
choroid and the posterior communicating arteries, beyond
the anterior choroid artery, and lastly between the A1 segment of the anterior cerebral artery and M1 of the sylvian
artery. In addition, the basal ganglia perforating arteries
can be separated into anterior (lenticular) and posterior
(thalamic) arteries at three different levels, according to
their location with respect to the basal ganglia. MRI is a
recent technique, which has made it possible for more to
be known about the radiological features of moyamoya disease. It provides very good precision for detection of small
strokes not seen on CT scan, whether ischemic or hemorrhagic, and for the assessment of the brain stem network

599

[3]. Neuroradiologists recognize the tortuous appearance
devoid of a signal with T1 weighting and the moth-eaten
appearance of the brain stem nuclei, showing the presence
of telangiectases of the moyamoya network. ARM (angiography by magnetic resonance) not only allows a carotid
stenosis to be seen, but transdural anastomoses can also be
identified by trained radiologists with good-quality equipment. Doppler examinations are particularly used for functional assessment after treatment, bearing in mind that before treatment, they notably reveal an acceleration in blood
flow in the ophthalmic arteries and in the vertebral axes,
with a reduction of blood flow in the carotid network. Functional cerebral imaging, which is currently being developed, also provides good quality peri- and postoperative
assessment, as Touho et al. [19] showed in a series of 25
patients. SPECT shows an antero-posterior gradient in
these patients, which is increased with hyperventilation,
and preferential posterior perfusion, as well as better perfusion in the gray nuclei. PET scanning also demonstrates
a reduction in local O2 metabolism in the ischemic areas,
with an increased O2 extraction coefficient in the cortex,
but not in the basal ganglia. Moyamoya was treated symptomatically for many years, and has always had a very poor
prognosis for function and survival. It was not until 1977
that Yonekawa and Yasargil [20] performed the first temporo-sylvian anastomosis, using a neurosurgical operation
that consisted of bypassing the stenosis at the brain stem
by creating a transdural anastomosis, using a vessel from
the external carotid network. He also made an end-to-side
anastomosis between the branches of the superficial temporal artery and the sylvian artery on the same side for anterior moyamoya. Since then, other authors have proposed
the same sort of operation, but between the occipital artery
and the calcarin artery, for the forms with posterior progression of the disease. Many indirect methods have followed on from these direct methods, whose main disadvantage lies in the small size of the sylvian artery for anastomosis. Some of these methods consist of occipito-gastro-epiploic anastomoses by omental transplantation into
the cortex. Others move the temporal muscle inside the
skull, which has the advantage of being possible in patients
at any age, although it has many disadvantages (postoperative epilepsy from compressive irritation, loss of muscle
trophicity in the medium term, a large hole with an unsightly appearance). Inspired by this encephalo-myo-synangiogenesis (EMS) described by Karasawa et al. in 1978
[11], Matsushima et al. [14] perfected a simple technique
in which a piece of epicranial aponeurosis is moved inside
the skull onto the dura mater, together with its branch of
the temporal artery. In our first case, the patient benefited
from a bilateral transdural angiogenesis (TDA), with a
spectacularly successful result. This type of operation is
quick and causes minimal damage, and remains at present
the easiest and most frequently performed in children, with
an angiographic efficacy that was first recognized in 1986
and has been confirmed more recently both by EEG and

by functional imaging techniques [15]. Direct anastomoses
are still very widely performed, with an equally good result in most cases. The neurosurgical contraindications are
rare, the main ones being early onset of the symptoms, a
high frequency of stroke, and the finding of a poor transdural network.. These three criteria were present in our second case, which is why we did not perform a curative surgical operation for this patient. At present, moyamoya
seems to us to be the only acceptable surgical indication
for treatment to prevent strokes, in particular in children.
Even if the mechanism responsible for the development of
basal ganglia collaterals is unanimously recognized as the
result of any type of recent intracranial stenosis, the exact
pathophysiology of the stenosing process in primary moyamoya as yet remains unclear. A malformation hypothesis
has easily been refuted, because the lesions are progressive, as proved by repeated angiograms performed on the
different patients and pathological examination of these areas. The histological description is in fact more or less that
of an atherosclerosed vessel in hypertension, with hyperplasia of the intima, made up of smooth muscle fibers and
elastic fibers, and the media, which has few fibrocytes and
separates it by an external elastic layer that has splits in
several layers or is simply jagged. No parietal inflammatory reaction has ever been seen on these stenosed arterial
walls, which excludes a form of arteritis. Fibrocellular hyperplasia of the intima seems to be the essential lesion, and
Aoyagi et al. [1, 2] and later Hoshimaru et al. [7] suggested
that the cause might be a defect in the synthesis of PDGF
(platelet-derived growth factor) receptors, when the FGF
(fibroblast growth factor) receptors are saturated, possibly
as a result of ischemia. In view of these recent observations and the fact that Kitamura et al. [12], as early as 1991,
described the stigmata of fibromuscular hyperplasia in all
the systemic vessels, and even in the smaller vessels, we
can honestly integrate primary moyamoya into the larger
group of genetically induced angiodysphasia, through the
growth factors of the small muscle fiber. Therefore, moyamoya disease must be known to pediatrician, neurologist
and neuro-surgeon, because of the multiple clinical features, from ischemic to hemorrhagic strokes, and because
there is a possible preventive treatment with brain revascularization.

600

References
1. Aoyagi M, Fugai N, Sakamoto H,
Shinkai T, Yoshiharu M, Yamamoto M,
Yamamoto K (1991) Altered cellular
responses to serum mitogens, including platelet-derived growth factor, in
cultured smooth muscle cell derived
from arteries of patients with moyamoya disease. J Cell Physiol
147:191–198
2. Aoyagi M, Fugai N, Matsushima Y,
Yamamoto M, Yamamoto K (1993) Kinetics of 125 I-PDGF receptor in arterial smooth muscle cells derived from
patients with moya-moya disease.
J Cell Physiol 154:281–288
3. Demearel P, Wilms G, Casteels-Van
Daele M, Marchal G, Van Frayenhoven
L, Bosmans H, Casaer P, Baert A
(1990) Moya-moya. Diagnosis by MRI
and AMR. J Radiol 21:119–123
4. Fujiwara J, Nakahara S, Enomoto T,
Nakata Y, Takita H (1996) The effectiveness of O2 administration for transient ischemic attacks in Moya-moya
disease in children. Child’s Nerv Syst
12:69–75
5. Handa J, Nakano Y, Okuno T, Komuro
H, Hojyo H, Handa H (1977) Computerized tomography in moya-moya
syndrome. Surg Neurol 7:315–319

6. Hoshi H, Ohnishi H, Jinnouchi S, Futami S, Nagamachi S, Kodama T, Watanabe K, Ueda T, Wakisaka T (1994)
Cerebral blood flow study in patients
with moya-moya disease evaluated by
IMP SPECT. J Nucl Med 35:44–50
7. Hoshimaru M, Takahashi J, Nagata I,
Kikuchi H, Nagata I, Hatanaka M
(1991) Possible roles of fibroblast
growth factor in the pathogenesis of
moya-moya disease: an immunohistochemical study. J Neurosurg
75:267–270
8. Israel J, Giroud M, Couailler JF, Alison M (1982) Malade de Moya-moya,
phénomène de Raynaud et déficit en
IgA. Arch Fr Pédiatr 39:825–827
9. Iwana T, Hashimoto N, Yoneka Y
(1996) The relevance of hemodynamic
factors to perioperative ischemic complications in childhood moya-moya
disease. Neurosurgery 38:1120–1125
10. Kalinyak K, Morris C, Ball W, Harris
R, Rucknagel D (1995) Bone-marrow
transplantation in a young child with
sickle cell anemia. Am J Hematol
48:256–261
11. Karawawa J, Kikushi H, Furuse S,
Kawamura J, Sasaki T (1978) Treatment of moya-moya disease with
STA-MCA anastomosis. J Neurosurg
49:679–688
12. Kitamura K, Fukui M, Oka K, Matsushima T, Kurokawa T, Hasuo K (1989)
Moya-moya disease. Handbook of
clinical neurology, vol 11(55): Vascular disease, part III, pp 293–306
13. Kodama M, Aoki Y, Hiraga H, Wada
T, Suzuki J (1979) Electroencephalography findings in children with moyamoya disease. Arch Neurol 36:16–29

14. Matsushima Y, Fukai N, Tanaka K,
Tsuruoka S, Inaba Y, Aoyagi M, Ohno
K (1981) New surgical treatment of
moya-moya disease in children:
a preliminary report. Surg Neurol
15:313–320
15. Ohashi K, Fernandz-Ulloa M, Hall L
(1992) SPECT, magnetic resonance
and angiographic features in a moyamoya patient before and after externalto-internal carotid artery bypass.
J Nucl Med 33:1692–1695
16. Picard L, André JM, Tridon P (1974)
Le syndrome de moya-moya. J Neuroradiol 1:47–54
17. Seeler RA, Royal JE, Power L, Goldbar HR (1978) Moya-moya in children
with sickle cell anemia and cerebrovascular occlusions. J Pediatr 93:808
18. Suzuki J, Takaku A (1969) Cerebrovascular moya-moya disease. Arch
Neurol 20:288–299
19. Touho H, Karasawa J, Ohnishi H
(1996) Pre-operative and post-operative evaluation of cerebral perfusion
and vasodilatory capacity with TcHMPAO SPECT and acetazolamide in
childhood moya-moya disease. Stroke
27:282–289
20. Yonekawa Y, Yasargil MG (1977)
Brain revascularization omentum: a
possible treatment of cerebral ischemia. Neurosurgery 1:256–259