Surg Neurol 179
1992 ;38 :179-85
Transcranial Doppler Sonography in Carotid-Cavernous Fistulas:
Analysis of Five Cases
Zainal Muttagin, M .D., Kazunori Arita, M .D., Tohru Uozumi, M .D.,
Satoshi Kuwabara, M .D., Shuichi Oki, M.D., Sinji Ohba, M.D., Kaoru Kurisu, M.D .,
Toshinori Nakahara, M.D ., Hiroaki Kohno, M.D., and Hideki Satoh, M .D.
Department of Neurosurgery, Hiroshima University School of Medicine, Hiroshima, Japan
Muttagin Z, Arira K, Uozumi T, Kuwabara S, Oki S, Ohba S,
Kutisu K, Nakahara T, Kohno H, Satoh H . Transcranial doppler
sonography in carotid-cavernous fistulas : analysis of five cases . Surg
Neurol 1992 ;38 :179-85 .
Transcranial doppler sonography was performed transor-
bitally in five patients clinically diagnosed as unilateral
carotid-cavernous fistula. Dural arteriovenous malfor-
mation related-shunts were detected in all the patients .
In the normal eyes, the only doppler signals observed at
an insonation depth of 45 to 55 mm were those of the
ophthalmic artery . In the affected eyes, abnormal doppler
signals with relatively higher flow velocity and lower re-
sistance were observed . In three of the cases, these abnor-
mal signals showed a flow directed anteriorly or away
from the cavernous sinus, consistent with changes in the
ophthalmic veins caused by the presence of the shunts .
In two cases, however, the observed flows were directed
posteriorly, the normal direction of these veins . The possi-
ble explanations for this discrepancy are discussed in rela-
tion with angiographic findings . The use of transcranial
doppler might provide a better understanding about he-
modynamic changes in carotid cavernous fistulas .
KEY WORDS : Transcranial doppler; Carotid-cavernous fistula
The introduction of a high-energy, pulsed doppler sys-
tem by Aaslid et al [1,2] enables one particular vessel to
be examined in isolation, at a defined depth, and this is
worked out by means of an electronic gate, in which
doppler shifts are registered only at certain distances
from the probe, within a defined sample volume [1] .
Aaslid et al [21 and Hennerici et al [10] recommended
the mean flow velocity values to be used to discriminate
normal from abnormal conditions . Using one of the
three transcranial windows mentioned by many authors
Address reprint requests to : Zaina[ Muttagin, M .D., Department of
Neurosurgery, Hiroshima University School of Medicine, 1-2-3 Ka-
sumi, Minami-ku, Hiroshima 734, Japan .
Received October 25, 1991 ; accepted February 21, 1992 .
® 1992 by Elsevier Science Publishing Co ., Inc .
[1,2,8,14], it is possible to measure the normal flow
velocity of the basal cerebral arteries and their main
branches . The orbital window may be used to insonate
the carotid siphon, ipsilateral ophthalmic artery, and
sometimes even the precommunicating segment of the
contralateral anterior cerebral artery [1,7,14] . The oph-
thalmic artery can normally be sampled at a depth of 40
to 55 mm, with its characteristic of higher resistance than
other intracranial vessels and its flow directed toward
the doppler probe [6,7,14] . Normal mean flow velocity
of the ophthalmic artery is 21 + 5 cm/second [14] .
Carotid-cavernous fistulas (CCF5) are abnormal com-
munications between the carotid artery or its dural
branches and the cavernous sinus . Symptoms and signs
are usually related to orbital congestion caused by retro-
grade filling of the ophthalmic veins from the cavernous
sinus [3,4] . The diagnosis of CCF can be made by com-
puted tomographic (CT) scan or magnetic resonance
imaging (MRI) scan [5,12], but the clear definition de-
pends on the demonstration of direct or indirect shunt-
ing from the carotid artery or its dural branches into
the cavernous sinus on cerebral angiography [8,14] . Its
classification into high or low flow still depends on angio-
graphic results alone [3,4] . On the other hand, angiogra-
phy is invasive and has some potential risks that make it
impractical for continuous monitoring and for repeated
follow-up in patients with CCF . Transarterial injection
of contrast material also gives an additional pressure,
which may cause a temporary hemodynamic change lo-
cally, and therefore influences the real angiographic pic-
tures, especially in low-flow arteriovenous shunts . Re-
ported here are the results of transcranial doppler
examination in five CCF cases ; they are analyzed in rela-
tion to the clinical and radiological findings .
Subjects and Methods
The study involved five patients clinically diagnosed as
having unilateral, spontaneous CCF . They were all ad-
mitted to the Department of Neurosurgery, Hiroshima
0090-3019/92/95 .00


180 Surg Neurol

1992 ;38 :179-85


TCD in Carotid-Cavernous Fistulas Surg Neurol 181

1992 ;38 :179-85

Figure 1 ; Case 1 . (A) MRI angiography (upper left) showed the right
superior ophthalmic vein asterisk) . Angiograpby, showed the early opacifi-
cation of the right cavernous sinus (small arrow) and the right superior
ophthalmic vein (large arrow), and revealed that this dural AVM-related
CCF is fed by the ducal branches of the right internal and external carotid
arteries directly, and also by the left external and internal carotid arteries
through intercavernous sinus. (B) Transcranial doppler result of the right
eye showed a relatively high flow velocity and low resistance signals directed
anteriorly (notice the direction of the small arrow away from doppler probe
on the upper right side of each recording), which is consistent with the
arterialized right superior ophthalmic vein . That of the left eye demonstrated
only signals of the normal ophthalmic artery . Normal signals of the right
ophthalmic artery were seen overlapped by the higher signal of the shunted
vein (large white arrow).
University Hospital, between July 1990 and June 1991 .
Transarterial carotid angiography using conventional or
digital subtraction angiography was performed, selec-
tively if possible, on all patients. Transcranial doppler
examinations were done through the orbital window, in
supine position, using a TC2-64 Transcranial Doppler
(EME, Uberlingen, Germany) . Insonations were done
from the superior palpebrae, over a closed eye, with
depths of 45, 50, and 55 mm . Doppler intensity was set
at 10% of spatial peak temporal average, as recom-
mended by the Bioeffect Committee of the American
Institute of Ultrasound in Medicine [8) . Doppler signals
of the ophthalmic arteries in the normal eyes served as
normal controls .
The patients consisted of three men and two women,
age 56 to 67 years (mean 61 .6 years) . All had spontane-
ous, unilateral, ducal arteriovenous malformation
(AVM)-related CCFs, four on the right side and one on
the left side . All patients complained of orbital conges-
tion of the involved eyes . Exophthalmos was found in
two patients, but they all experienced or had once expe-
rienced some kind of tinnitus or bruit, four ipsilaterally
and one contralaterally (case 1) . Table 1 describes details
of patients' signs and symptoms and their angiographic
findings . CT had been performed in two patients and
MRI angiography in another one, and all the results
showed an enlarged intraorbital vein in the affected side .
Results
The results of transcranial doppler examination were
detailed in Table 2 . In all cases, normal signals from the
ophthalmic artery were the only flow signals observed
in the healthy side. These were directed anteriorly and
had mean flow velocities between 16 and 32 cm/s (mean
23 .6 em/s) and pulsatility indices between 1 .0 and 1 .9
(mean 1 .32) . On the pathological side, besides signals of
the normal ophthalmic artery, there were other signals
with relatively higher flow velocity and lower resistance .
These abnormal signals were seen to overlap the signals
of the normal ophthalmic artery (cases 1 and 2) or were
seen as separate signals by moving the insonation angle
slightly (cases 3, 4, and 5) . In the first three cases, these
abnormal signals indicated anteriorly directed flow, or a
direction away from the cavernous sinus (Figure 1 B) .
In the other two cases, however, the abnormal signals
indicated a posteriorly directed flow, exactly the same as
the normal flow direction of the normal ophthalmic vein
(Figure 3 B) . The vessel's resistance, as shown by the
pulsatility index values, ranges from 0 .39 to 0 .89 (mean
0 .62), and flow velocity varies between 20 to 70 cm/s
(mean 42 .4 cm/s) . In case 4, internal carotid angiography
showed a slight opacification of the superior ophthalmic
vein in the late arterial phase (Figure 3 A) . External

Table 1 . Characteristics of Patients with Carotid-Cavernous Fistula

Signs and symptoms
Case Sex Age, yr CCF side Initial Present Feeding arteries
I M 56 Right Bruit (contralateral), Orbital headache, Cavernous branches of
orbital headache, chemosis of right eye bilateral ICAs, bilateral
chemosis of right eye IMAs, bilateral MMAs
F 61 Right Chemosis, exophthalmos, Chemosis, exophthalmos, Cavernous branch of
double vision, and and rerroorbital pain right ICA
right retroorbital pain
3 M 67 Left Bruit, chemosis, double Bruit, chemosis, double Cavernous branch(es) of
vision, and mild ptosis vision, and mild ptosis left ICA
of the left eye of the left eye
F 59 Right Bruit, chemosis, Chemosis, exophthalmos Cavernous branch of
exophthalmos, and right ICA, right IMA,
retroorbital pain of the right MMA
right eye
5 M 65 Right Bruit, chemosis, double Chemosis, bruit Right ascending
vision, and mild prosis (preauricular), double pharyngeal, right IMA
of the right eye vision, mild ptosis
Abbreviations . CCF, carotid-cavernous fistula ; M, male ; ICA, internal carotid artery ; IMA, internal maxillary artery; MMA, middle meningeal artery ; F, female .


182 Surg Neurol

1992 ;38 :179-85


TCD in Carotid-Cavernous Fistulas Surg Neurol 183

1992 ;38 : 1 79-85

Figure 2 ; Case 4. (A) Right internal carotid arteriography opacified
ipsilateral superior ophthalmic vein (large arrow) in the laterarterial phase.
Right external carotid arteriograpby showed feeding arteries from the inter-
nal maxillary artery and middle meningeal arteries . The right cavernous
sinus (small arrow) was weakly opacified, and drained posteriorly to the
inferior pet royal sinus and the jugular vein (arrowhead) . (B) Signals of the
normal ophthalmic arteries were observed in both eyes 4eLt and middle) .
However, a little change of the insonation angle in the right eye revealed
relatively high flow velocity and low resistance signals and indicated a
posteriorlydirectedflow ri bt toward the cavernous sinus (notice The direc-
tion of the small arrow in the right upper corner) . This flow direction is in
accordance with the normal flora direction of the superior ophthalmic vein .
carotid angiography opacified the cavernous sinus
slightly, and it drained posteriorly into the inferior petro-
sal vein . There was no opacification of the superior oph-
thalmic vein . In case 5, angiography revealed numerous
small abnormal vessels from the right ascending pharyn-
geal and the right internal maxillary arteries without a
clear visualization of the cavernous sinus itself. In this
case, the inferior ophthalmic vein drained first to the
inferior petrosal sinus and then to the internal jugular
vein .
Discussion
Angiographically, CCFs can be classified according to,
first, the velocity of the blood flow through the shunt
and, second, the anatomical origin of the artery supply-
ing the CCFs [3,4] . Most of the CCF cases result from
trauma, but about 25%-30% of all cases occur sponta-
neously [3,4,111 . Using the above criteria, Barrow et al
[3) mentioned four types of CCFs . Among these types,
their type D includes most of the dural AVM-related
CCFs . They characteristically have insidious onset, are
fed by dural branches of both external and internal ca-
rotid arteries, have a low flow, and about 10%-60%
have the tendency to resolve without treatment . Sponta-
neous CCFs will result in enlargement of the superior
ophthalmic vein, accompanied by arterialization and re-
versal of its flow away from the cavernous sinus [4,6,9).
Under normal conditions, superior ophthalmic veins
were difficult to insonate, as shown by the result of
insonation of the normal eyes in our cases . Doppler
signals recorded at the depth of 45 to 55 mm of the
normal eyes were merely those of the normal ophthalmic
arteries . Abnormal doppler signals found in the CCF
side of cases 1, 2, and 3, which indicated anteriorly
directed flow, showed changes of flow patterns-the so-
called arterialization-of the superior ophthalmic veins
caused by the presence of arteriovenous shunt in the
cavernous sinus . Compared to the normal signals of the
ophthalmic arteries, these abnormal doppler signals of
the ophthalmic veins had almost twice the mean flow
velocity values and less than half of pulsatility index
values . The same observation had also been reported by
Gomez et al (6) .
Contrary to this "classic" picture, cases 4 and 5
showed the presence of posteriorly directed abnormal
flow, or the same flow direction as the normal ophthalmic
vein (Figures 2B) . This unusual flow direction of the
ophthalmic vein involved in CCF may be explained with
the following hypothesis : The direct draining vein of

Table 2 . Results of Angiography and Transcranial Doppler Sonography

TCD signals Abnormal signals Normal signals
of normal eye of affected eye of affected eye
Case Angiography" MFV, cm/s PI° Direction MFV, cm/s Pt" Direction MFV, cm/s PP Direction

I Rt cavernous sinus and ophthalmic vein 16 1 .0 Anterior 42 0 .41 Anterior Overlapped by abnormal
were seen on midarterial phase . signal
2 Rt cavernous sinus was seen on early 16 1 .9 Anterior 70 0.39 Anterior Overlapped by abnormal
arterial phase, but superior ophthalmic signal
vein was weakly stained later .
3 Is cavernous sinus, ophthalmic vein, and 32 1 .2 Anterior 20 0.89 Anterior 26 .15
1 Anterior
inferior petrosal vein were strongly
stained at early arterial phase .
4 Rr cavernous sinus and superior 30 1 .3 Anterior 0 .53 Posterior 28 1 .3 Anterior
ophthalmic vein were weakly stained
at late arterial phase of ICAG, but
ECAG didn't opacify any ophthalmic
vein .
5 Rt inferior ophthalmic vein and inferior 24 1 .2 Anterior 36 0 .89 Posterior 26 1 .113 Anterior
petrosal vein were weakly stained at
midarterial phase of ECAG .
Abbreviations : TCD, transcranial doppler: MFV, mean flow velocity ; Pl. pulsatility index; Rt, right; Lt, left, ICAG, internal carotid arreriography ; ECAG, external
carotid arteriography.
Opacihcation of the cavernous sinus and its draining veins during angiography (timing and thickness of staining) .
Gosling's pulsatility index = peak systolic velocity - end diastolic velocity/time mean velocity .


184 Surg Neurol

1992 ;38 :179-85

Figure 3 : Case 3 . (A) CT scan (lam) showed the enlarged left superior ophthalmic vein, while angiography revealed the presence of a large shunt from
dural branches of the left internal carotid artery to the left cavernous sinus (open arrow) . which then drained anteriorly to the ipsilateral superior
ophthalmic vein (arrow and posteriorly to the inferior petrosal sinus (arrowhead) . Note the opacification of the ptetygopalatine plexus (double arrows) .
(B) Insonation of the right (normal) eye showed only doppler signals of the right ophthalmic artery !right) . In the left eye, besides signals of the left ophthalmic
artery (lt), a little change of the insonation angle disclosed signals with relatively low resistance directed away from the cavernous sinus, presumed to be
the signals of the left superior ophthalmic vein (middle) .


TCD in Carotid-Cavernous Fistulas Surg Neurol 18 5

1992 ;38 :179-85

a CCF, such as the ophthalmic vein, will gradually be only noninvasive means to evaluate hemodynamic
arterialized [4,6,9] . It is possible that this pressure- changes in CCF .
related physical change of the venous wall will provide
negative feedback for the shunt itself, especially the very
low flow type . This consideration might be related to the
References
fact that some low-flow CCFs tend to resolve spontane-
1 . Aaslid R . Transcranial doppler examination techniques . In: Aaslid
ously, without treatment or after repeated manual com- R, ed : Transcranial doppler sonography . Vienna : Springer-Verlag,
pression of the carotid artery and internal jugular vein 1986 :39-59 .
in the neck . Transarterial injection of contrast material, 2 . Aaslid R, Markwalder TM, Nornes H . Noninvasive transcrania]
doppler ultrasound recording of flow velocity in basal cerebral
especially during selective carotid angiography, gives an
arteries . J Neurosurg 1982 ;57 :769-74 .
additional flow pressure temporarily as noted by Greitz
3 . Barrow DL, Spector RH, Braun IF, Landman JA, Tindall SC,
and Tornell [7] . This additional pressure changed the Tindall GT . Classification and treatment of spontaneous ca-
local hemodynamics temporarily and, as a consequence, rotid-cavernous fistulas. J Neurosurg 1985 ;62 :248-56.
resulted in opacification of an otherwise "balanced" 4 . Debrun GM, Vinuela, F, Fox AJ, Davis KR, Ahn HS . Indications
shunt . Cases 4 and 5 might represent some of the low- for treatment and classification of 132 carotid-cavernous fistulas .
Neurosurgery 1988 ;22 :285-9 .
flow, dural AVM-related carotid-cavernous fistula
5 . Elster AD, Chen MYM, Richardson DN, Yeatts PR . Dilated
cases with arrested or balanced arteriovenous shunts . intercavernous sinuses : an MR sign of carotid-cavernous and
The present clinical presentations of cases 4 and 5 were carotid-dural fistulas . AJNR 1991 ;12 :641-5 .
very mild and were far better than the initial signs and 6 . Gomez CR, Gomez SM, Yoon K-WP, Kraus GE . Evaluation and
symptoms as described by the patients . On angiography, follow-up of carotid-cavernous fistulas by transcranial doppler
sonography : illustrative case . Neurosurgery 1989 ;24 :749-53 .
the ophthalmic veins and/or the cavernous sinus were
7 . Greitz T, Tornell G . Bradycardial reactions during cerebral angi-
weakly opacified during the mid to late arterial phase of ography : a comparison of Isopaque sodium, Isopaque B, Hypaque,
both of these cases . Angiography with contrast material and Urografin . Acta Radio/ Suppl (Stockh) 1967 ;270 :75-86 .
injected transvenously or at a more proximal arterial 8. Harders A . Neurosurgical applications of transcranial doppler
trunk would be agood method of obtaining a real picture sonography. Vienna: Springer-Verlag, 1986 :12-5 .
of the very low flow arteriovenous shunt without giving 9. Hassler W . Haemodynamic aspects of cerebral angiomas . Area
Neurochir Suppl (Wien) 1986 ;37 :38-108 .
any additional pressure to the shunt itself.
10 . Hennerici M, Rautenberg W, Schwartz A . Transcranial doppler
Case 3 showed an intense opacification of the cavern- ultrasound for the assessment of intracranial arterial flow velocity .
ous sinus and all its draining veins in the early arterial Part 2 : Evaluation of intracranial arterial disease . Surg Neurol
phase on angiography (Figure 2 A) . Doppler signals of 1987 ;27 :523-32 .
the ophthalmic vein, however, showed a relatively low 11 . Keltner JL, Satterfield D, Dublin AB, Lee BCP . Dural and
carotid-cavernous fistulas : diagnosis, management, and complica-
flow velocity (Table 2) . This is related to the principle tions . Ophthalmology 1987 ;94 :1585-600 .
that transcranial doppler measures flow velocity, not flow 12 . Leppien A, Eppen R, Pohlenz O . CT-findings of carotid cavernous
volume, so it will be difficult to relate transcranial dopp- fistula . A case report . Neurosurg Rev 1988 ; 11 :293-6.
ler results to the volume of the shunted blood . 13 . Phelp CD, Thomson HS, Ossoirig KC . The diagnosis and progno-
From this observation, it seemed that transcranial sis of atypical carotid-cavernous fistula (red eye shunt syndrome).
Am J Ophthalmol 1982 ;93 :423-36 .
doppler has the ability to record changes of the flow
14 . Ringelstein EB . A practical guide to transcranial doppler so-
velocity, the vessel's resistance, and direction of flow in nography . In : Noninvasive imaging of cerebrovascular disease.
the ophthalmic vein in patients with CCF, and it is the New York : Alan R . Liss, 1989 :75-121 .