ORIGINAL

RESEARCH
E.F. Hauck
B.G. Welch
J.A. White
P.D. Purdy
L.G. Pride
D. Samson

Preoperative Embolization of Cerebral

Arteriovenous Malformations with Onyx
BACKGROUND AND PURPOSE: Preoperative embolization facilitates the surgical management of com-

plex cerebral arteriovenous malformations (cAVMs). This analysis aims to investigate the risks for
preoperative cAVM embolization with Onyx.
MATERIALS AND METHODS: We retrospectively analyzed clinical data of all patients who underwent
embolization with Onyx as a preoperative treatment of cAVMs at our institution since 2005 (US Food
and Drug Administration [FDA] approval). Patients with arteriovenous fistulas were excluded. A total of
107 patients were treated for cAVMs during the study period. Of those patients, 41 underwent cAVM
embolizations with Onyx in 82 procedures.
RESULTS: After the embolization, the cAVM diameter was reduced from 3.71 1.55 cm to 3.06
1.89 cm (P .05). Median volume reduction was 75%. Complete occlusion with embolization alone
was achieved in 4 (10%) cAVMs. The recurrence rate for completely occluded cAVMs was 50% (2
patients). A total of 71% of the 41 patients treated with Onyx underwent surgery, and 15% underwent
radiosurgery. There were 9% who have not yet received definitive treatment of their residual cAVMs.
A new permanent neurologic deficit occurred in 5 patients (6.1% per procedure or 12.2% per patient).
CONCLUSIONS: A considerable risk for a permanent neurologic deficit remains for cAVM embolization
with Onyx. The risk has to be carefully weighted against the benefit of volume reduction in the
treatment of cAVMs.

reoperative embolization of cerebral arteriovenous malformations (cAVMs) has been used successfully for many
decades to facilitate surgical resection.1-6 However, complications secondary to the embolization itself continue to limit its
use. Onyx (ev3, Irvine, Calif) is a newer and promising embolic agent, which has become widely available in North
America only after its approval by the Food and Drug Administration (FDA) in July 2005. Our study analyzes the risks and
success associated with Onyx embolization of cAVMs at the
University of Texas Southwestern Medical Center in Dallas.
Materials and Methods
Patient Selection
Between January 2005 and January 2008, a total of 107 patients were
treated with various modalities for cAVMs at our institution. Patients
with the diagnosis of a cerebral arteriovenous fistula are not included.
Open surgical resection as definitive treatment was the most common
treatment choice (65 patients [61%]). If preoperative imaging suggested a good possibility of early surgical control of feeding vessels,
and/or the location and size of the AVM was favorable, surgical resection was performed without preoperative embolization in 15 patients
(14%). In more complex AVMs with higher Spetzler-Martin grades
(SMGs), preoperative embolization was used in 47 patients (44%). In
3 patients (3%), a small residual AVM after surgical resection was
treated with the gamma knife. Small AVMs, particularly in deeper
locations, were primarily treated with the gamma knife in 26 patients
(24%). Some smaller AVMs with fistulous components were occaReceived July 16, 2008; accepted after revision September 24.
From the Department of Neurosurgery (E.F.H., B.G.W., J.A.W., D.S.), and Department of
Neuroradiology (B.G.W., P.D.P., L.G.P.), University of Texas Southwestern Medical Center,
Dallas, Tex.
Please address correspondence to Erik F. Hauck, MD, PhD, Department of Neurosurgery,
University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390;
e-mail: erik.hauck@utsouthwestern.edu
DOI 10.3174/ajnr.A1376

492

Hauck AJNR 30 Mar 2009 www.ajnr.org

sionally treated with embolization only (7 patients [7%]) or in combination with the gamma knife (9 patients [8%]).
After its approval by the FDA, Onyx quickly became the primary
embolic material for the treatment of cAVMs because of its ease in
handling and superior control, with the exception of fistulous components, where we often prefer n-butyl cyanoacrylate (n-BCA). As a
consequence, 41 of the 107 patients underwent embolization with
Onyx alone or in conjunction with other materials in a total of 82
separate stages. We identified patients retrospectively by reviewing
hospital charts. Data on patient demographics, AVM characteristics,
and endovascular procedures before surgery were analyzed. Approval
of this study was granted by the Institutional Review Board following
the standard protocol for retrospective reviews of patient charts and
electronic records in accordance with the Health Insurance Portability and Accountability Act.

Patient and AVM Characteristics

The patients (n 41) were relatively young with a median age of 41
years (age range, 12 62 years). The female sex was predominant
(male-to-female ratio, 15:26). Nineteen patients (46%) presented
with an intracerebral hemorrhage caused by the malformation. The
remaining patients presented with seizures and/or headaches. The
malformations were located supratentorially in 88% (36 patients). In
19 patients (46%), the lesion was left sided; the lesion was located in
the midline in 2 patients (5%). Eloquent brain5 was involved in 31
patients (77%). Mean maximal diameter of the cAVM nidus was
3.71 1.55 cm. The measurements of the maximum AVM diameter
were based on diagnostic imaging including angiograms, MR imaging, and CT scans. We measured the postembolization maximal
cAVM diameter using the relative maximal diameter change as observed by comparison of preembolization and postembolization arteriograms and absolute measurements of the maximal preembolization diameter by CT or MR imaging. The average volume reduction of
the cAVM was analyzed in categories according to Weber et al7 in
reference to the postembolization angiogram. Deep venous drainage
was present in 22 patients. Median preoperative Spetzler-Martin

Table 1: Demographics and arteriovenous malformation specifics in

36 patients who underwent preoperative embolization with Onyx
Characteristics
Age (y); median (range)
Sex, no. pts (%)
Male
Female
Presentation, no. pts (%)
Hemorrhage
Other (seizure/headache)
AVM specifics
Side, no. pts (%)
Left
Right
Midline
Location, no. pts (%)
Supratentorial
Infratentorial
Eloquence*, no. pts (%)
Size (cm), mean SD
Deep venous drainage,
no. pts (%)
No. (%) pts with reduced SMG

Before
Embolization
41 (1262)

After
Embolization

P value
Complication
Death
Permanent neurological deficit ( 6 mo)
Transient neurological deficit ( 48 h)
Technical complication without deficit
Asymptomatic small infarction
Perforation
Catheter entrapment
No complication
Total

15 (37)
26 (63)
19 (46)
22 (54)

19 (46)
20 (49)
2 (5)
36 (88)
5 (12)
31 (77)
3.70 1.55
20 (54)

Table 2: Complications of preoperative embolization of cerebral

arteriovenous malformations with Onyx

18 (50)
3.06 1.89
18 (50)

1.00
.05
.16

10 (24)

.05

Note:AVM indicates arteriovenous malformation; no. pts, number of patients; SMG,

Spetzler-Martin grade.
*Eloquent regions included deep structures such as the hypothalamus, thalamus, brainstem, and cerebellar peduncles as well as cortical regions such as the sensorimotor,
language, and primary visual areas.
Maximal diameter of the perfused nidus.

Interventions
In 41 patients, a total of 82 separate embolization procedures were
performed. Onyx was used in all 82 embolizations, either as the sole
agent (62%) or in conjunction with other materials (38%) such as
polyvinyl alcohol particles, detachable coils, and n-BCA. On the basis
of preembolization arteriography information, the goal of the embolization was cure in only a few cases. Instead, the treatment intent
was to reduce the flow in the cAVM by occlusion of feeding arteries
while preserving the draining veins to reduce the risk and increase the
success of a subsequent treatment, typically surgery. A median of 2
embolization procedures was performed per patient (range, 15).
Most the patients (29 patients [71%]) then underwent surgical resection of the cAVM. Six patients (15%) were treated with stereotactic
radiosurgery. Four patients (8%) did not yet receive additional
treatment.

Statistics
We performed the statistical analysis using the statistics program
SPSS 15.0 (SPSS, Chicago, Ill). Patient demographics, cAVM characteristics, endovascular procedures, results, and complications were
described for all patients. Frequency and percentage values of categoric variables as well as median and range of continuous variables
were determined. We compared data using the paired t test, Wilcoxon
test, and 2 test. A probability value of less than 5% was considered
significant.

No. of
Patients
(%)
0 (0)
5 (12.2)
2 (4.8)
15 (36.6)
3 (7.3)
9 (22.0)
4 (9.8)
19 (46.3)
41 (100)

Results
Embolization with Onyx alone or in conjunction with other
embolic materials reduced the size of the perfused cAVM by
75% (median; range, 8% to 100%). In 7 patients, the cAVM
was reduced by less than 50%. Volume reduction was 50% to
59% in 4 patients, 60% to 69% in 2 patients, and 70% to 79%
in 9 patients. The volume reduction reached 80% to 89% in 7
patients and 90% to 99% in 8 patients. The maximal cAVM
diameter was reduced from 3.71 1.55 cm before embolization to 3.06 1.89 cm postembolization (P .05; paired t
test). Complete occlusion was achieved in 4 patients (10%).
The cAVM partly recanalized in 2 of those patients (50%),
who then underwent surgical resection. The presence of deep
venous drainages was not significantly affected by the embolization. Permanent complications occurred in 5 patients
(12.2% per patient) or in 5 of 82 procedures (6.1% per procedure). The first patient experienced a small but bithalamic
infarct with a mild cognitive deficit and difficulties with calculations after the initial embolization of a pineal region cAVM.
Two patients with cerebellar AVMs experienced mild ataxia
and dysmetria after the initial embolization. CT scan revealed
a new cerebellar hypoattenuation. Another patient experienced a new persistent hemiparesis after a fourth embolization
of a large right frontal cAVM. The embolization was complicated by a perforation with a small intraparenchymal hemorrhage in the posterior frontal lobe.
A 59-year-old woman underwent partial Onyx embolization of a large, left-sided frontotemporoparietal cAVM with
the intent to treat an intranidal feeding aneurysm, presumably
the cause of the intraparenchymal hemorrhage. Although the
feeding aneurysm was completely occluded, the procedure
was complicated by a new aphasia and hemiparesis. Two patients experienced transient deficits that resolved within 48
hours after the procedure. Asymptomatic technical complications occurred in 16 patients (44%). An asymptomatic small
vessel perforation (contrast extravasation or hemorrhage on
the postprocedure CT) was observed in 9 patients. A new lacunar infarct without neurologic deficit was documented on
postprocedure MR imaging or CT in 3 patients. The microcatheter was retained in situ in 4 procedures and was then
removed during surgery without neurologic consequences.
The effects of the preoperative cAVM embolization with Onyx
are summarized in Table 1, and the complications are summarized in Table 2.
AJNR Am J Neuroradiol 30:49295 Mar 2009 www.ajnr.org

493

INTERVENTIONAL ORIGINAL RESEARCH

grade (SMG) was III (range, IV). Three patients presented with an
SMG I cAVM, 14 patients with an SMG II cAVM, 15 patients with an
SMG III cAVM, 7 patients with an SMG IV cAVM, and 3 patients
harbored an SMG V cAVM. Demographics and AVM characteristics
are summarized in Table 1.

No. of
Procedures
(%)
0 (0)
5 (6.1)
2 (2.4)
16 (19.5)
3 (3.7)
9 (11.0)
4 (4.9)
59 (72.0)
82 (100)

Table 3: Literature reporting on preoperative embolization of cerebral arteriovenous malformations with Onyx
st

1 Author
Jahan
Hamada
van Rooij
Weber
Weber
Mounayer
Hauck

Year
2001
2002
2007
2007
2007
2007
Current

N
23
57
44
47
93
94
41

Surgery

Radiosurgery

Cured*

Mean

No. Pts (%)

12 (52)
54 (95)
10 (23)
47 (100)
56 (60)
7 (7)
29 (71)

No. Pts (%)

11 (48)
3 (5)
20 (45)
0 (0)
13 (14)
20 (21)
6 (15)

No. Pts (%)

0 (0)
0 (0)
7 (16)
0 (0)
17 (18)
26 (28)
2 (5)

Size (cm)
3.6

3.9
3.4

3.7

Per Procedure
Morb. (%)
3.0
3.4
3.8
12
5
3.8
6.1

Mort. (%)
0
0
1.9
0
0
1.4
0

Per Patient
Morb. (%)
4
5.3
4.6
28
12
8.5
12

Mort. (%)
0
0
2.3
0
0
3.2
0

Note:N indicates total number of patients; Morb., morbidity; Mort., mortality.

*Complete occlusion without recurrence.
Only patients treated with Onyx are included.

Discussion
Our study demonstrates that a considerable risk remains with
the application of Onyx for the embolization of cAVMs. Advantageous is the ease in handling, which allows a controlled
and significant size and flow reduction of cAVMs to the degree
of complete occlusion. After significant size reduction, surgery
or radiosurgery may be used to treat the malformation (Table
3). It is important to acknowledge that many of the complications associated with Onyx embolizations are not primarily
related to Onyx as the embolic agent itself, but rather procedure related. This includes vessel perforations with microguidewires or microcatheters. Furthermore, sometimes it is difficult to decide if an embolic complication was indeed the
result of Onyx injection or possibly from injection of a second
embolic agent. This particularly applies for procedures done
with the patient under general anesthesia, in which occasionally an immediate neurologic change might not be recognized,
despite electrophysiologic monitoring.
Morbidity and Mortality Rates after Preoperative
Embolization with Onyx
Various studies have investigated the morbidity and mortality rates associated with preoperative cAVM embolization before the introduction of Onyx.1,4,8-12 At that time,
the morbidity and mortality rates with preoperative embolization ranged from 3.8% and 1.0% to 28% and 3.7%,
respectively. Similar to other investigators, with use of
Onyx, our results of a permanent morbidity rate of 12.2%
and no deaths after preoperative Onyx embolization are
within that range (Table 3). Ann H. Costello and Judy Chen
presented clinical data regarding Onyx embolization collected from 17 centers (including Dallas) at the Neurologic
Devices Advisory Panel Meeting on August 5, 2003 (http://
www.fda.gov/OHRMS/DOCKETS/AC/03/slides/3975s1
02-fda.ppt). In this unpublished study, 51 patients were
treated with Onyx embolizations of cAVMs. The observed
stroke rate was 4.3%. The mortality rate documented in this
analysis was 4.3% as well. The only experimental study
from a single North American center was published by Jahan et al13 in 2001. They found a permanent morbidity rate
of 4% in 23 patients treated. In their series, there were no
deaths. Other similar results are documented by Mounayer
et al13 in 2007 in a series of 94 patients. They described an
overall morbidity rate of 8.5% per patient and a mortality
rate of 3.2%. Hamada et al15 documented a morbidity rate
of 5.3% in a series of 57 patients, with no deaths occurring.
van Rooij et al16 experienced a morbidity rate of 4.6% per
494

Hauck AJNR 30 Mar 2009 www.ajnr.org

patient and a mortality of 2.3%. Two additional studies

with contradictory results have been published by Weber et
al.7,17 In the first study, Weber et al examined the morbidity
rate in 93 patients treated with Onyx for cAVMs, which
apparently was 12% per patient, with no deaths. In a second
consecutive study with 47 patients, they documented an
even higher morbidity rate of 28%. The results of the second study seem less representative because the observed
high morbidity rate of 28% is not only in contrast to the
authors own experience but also the other published and
unpublished data presented above. In summary, the morbidity and mortality rates of preoperative Onyx embolization of cAVMs may be similar to the morbidity and mortality rates of embolization before the introduction of Onyx
(overall range, 10% per patient [5% per procedure]).
Volume Reduction
Volume reduction is difficult to measure precisely. Unlike selective angiography, CT, MR imaging, or CT angiography allows depiction of the entire nidus with injection of all feeding
vessels simultaneously. Furthermore, with CT, CT angiography, or MR imaging, no calibration is required to determine
precise dimensions. However, the resolution of the angio-architecture details is inferior compared with standard angiography. Finally, measurements of a diffuse nidus may vary considerably depending on the investigator. Possibly because of
these limitations, some studies with Onyx did not present precise measurements regarding volume reduction.14-16 We measured the volume reduction of the cAVM postembolization in
categories according to Weber et al7,17 They calculated the
AVM volume on the basis of standard arteriography with the
formula suggested by Pasqualin et al18: V 12 w h l
[V volume, w width, h height, l length]. Volume
reduction was then determined on the basis of postembolization angiography. Weber et al7,17 achieved an 80% to 84%
volume reduction with a considerably high morbidity rate. In
their presentation, Ann H. Costello and Judy Chen quote a
volume reduction of 50% in 97.6% (41/42) of the patients
after embolization with Onyx compared with 84.3% (43/51)
with n-BCA (unpublished data). Jahan et al13 reported an average volume reduction of 63% with a low morbidity rate and
no deaths. On the basis of these reports, a volume reduction of
50% to 65% may be a desirable target with acceptable associated morbidity and mortality rates. The addition of Onyx to
the endovascular armamentarium seems to improve occlusion rates.

Cure
Complete occlusion of a cAVM without recurrence on follow-up angiography may represent a cure.14,19 A prerequisite for cure is the complete occlusion of the draining veins.
However, the primary principle of preoperative embolization is the preservation of the draining veins because the risk
for hemorrhage increases with partial embolization and obstructed venous outflow.13-16 As a consequence, a cure may
occur either as the result of the planned sacrifice of the venous
outflow in small cAVMs or is accidental with increased risk for
the patient. With 2 competing goals, sacrifice of the draining
vein(s) for cure versus preservation of the venous outflow with
size reduction in preparation for additional treatment, the reported cure rates between different centers varies considerably (from 0% to 28%).7,13-17,19 At our institution, cure is
typically not the goal of the preoperative embolization. Complete occlusion of the cAVM was achieved in 10%. Half of
these patients were found to have recurrence of cAVM during
follow-up and then underwent surgery. Therefore, our cure
rate was only 5%. Mounayer et al14 as well as Katsaridis et al19
achieved the highest percentage of cure, both in 28%. Mounayer et al14 report a morbidity rate of 8.5% and a mortality
rate of 3.2%, respectively. Katsaridis et al19 document a similar
morbidity rate with 8% and a mortality rate of 3%. It is noteworthy that a higher cure rate as a result of more aggressive
embolization may be associated with a higher mortality rate.
However, no final conclusions can be drawn at this point.
Additional experience with Onyx and long-term follow-up of
cured cAVMs is required to determine the value of complete
cAVM occlusion after embolization as stand-alone treatment.
Definitive Treatment
Surgical resection of the cAVM with or without preoperative
embolization is the most definitive treatment of cAVMs with
total removal of the lesion.1,4-6,8-12 Accordingly, surgery was
performed as definitive treatment after preoperative Onyx
embolization in more than 50% of the study patients in 5 of 7
centers.7,13-17 However, with improving obliteration rates
with Onyx and/or n-BCA, radiosurgery is more frequently
used as an alternative definitive treatment for small residual
cAVMs.13,14,16 Jahan et al13 treated 48% of their patients with
radiosurgery after embolization. At the time of the report, they
had limited follow-up and documented complete occlusion
after 20 months in only 1 patient. van Rooij et al16 treated 45%
of their patients with radiosurgery after Onyx embolization.
There were 5 of 20 patients who had confirmed cAVM occlusion on follow-up arteriograms; the other 15 patients were still
pending follow-up. Mounayer et al14 treated 20 patients with
radiosurgery as well but did not quote any occlusion rates. Our
patients treated with radiosurgery are awaiting follow-up as
well. In summary, embolization with Onyx followed by radiosurgery is a promising concept. More clinical data will be necessary for a more definitive analysis. Surgical resection re-

mains the reference standard for residual cAVM after

embolization.
Conclusions
On the basis of our results, a considerable risk (6.1% per procedure or 12.2% per patient) for a permanent neurologic deficit remains for cAVM embolization with Onyx. The risk has
to be carefully weighted against the benefit of volume reduction in the treatment of cAVMs.
Acknowledgments
We thank Kim Dutton and Jerri Thomas for their support.
References
1. Haw CS, terBrugge K, Willinsky R, et al. Complications of embolization of
arteriovenous malformations of the brain. J Neurosurg 2006;104:226 32
2. Purdy PD, Batjer HH, Risser RC, et al. Arteriovenous malformations of the
brain: choosing embolic materials to enhance safety and ease of excision.
J Neurosurg 1992;77:21722
3. Purdy PD, Batjer HH, Samson D, et al. Intraarterial sodium amytal administration to guide preoperative embolization of cerebral arteriovenous malformations. J Neurosurg Anesthesiol 1991;3:103 06
4. Samson D, Ditmore QM, Beyer CW Jr. Intravascular use of isobutyl
2-cyanoacrylate: Part 1. Treatment of intracranial arteriovenous malformations. Neurosurgery 1981;8:4351
5. Spetzler RF, Martin NA. A proposed grading system for arteriovenous malformations. J Neurosurg 1986;65:476 83
6. Taylor CL, Dutton K, Rappard G, et al. Complications of preoperative embolization of cerebral arteriovenous malformations. J Neurosurg 2004;100:810
7. Weber W, Kis B, Siekmann R, et al. Endovascular treatment of intracranial
arteriovenous malformations with Onyx: technical aspects. AJNR Am J Neuroradiol 2007;28:37177
8. Debrun GM, Aletich V, Ausman JI, et al. Embolization of the nidus of brain
arteriovenous malformations with n-butyl cyanoacrylate. Neurosurgery 1997;
40:11220; discussion 120 21
9. Frizzel RT, Fisher WS 3rd. Cure, morbidity, and mortality associated with
embolization of brain arteriovenous malformations: a review of 1246 patients
in 32 series over a 35-year period. Neurosurgery 1995;37:103139; discussion
1039 40
10. Hartmann A, Pile-Spellman J, Stapf C, et al. Risk of endovascular treatment of
brain arteriovenous malformations. Stroke 2002;33:1816 20
11. n-BCA Trial Investigators. N-butyl cyanoacrylate embolization of cerebral arteriovenous malformations: results of a prospective, randomized, multi-center trial. AJNR Am J Neuroradiol 2002;23:748 55
12. Valavanis A, Yasargil MG. The endovascular treatment of brain arteriovenous malformations. Adv Tech Stand Neurosurg 1998;24:131214
13. Jahan R, Murayama Y, Gobin YP, et al. Embolization of arteriovenous malformations with Onyx: clinicopathological experience in 23 patients. Neurosurgery 2001;48:984 95; discussion 99597
14. Mounayer C, Hammami N, Piotin M, et al. Nidal embolization of brain arteriovenous malformations using Onyx in 94 patients. AJNR Am J Neuroradiol
2007;28:518 23
15. Hamada J, Kai Y, Morioka M, et al. A mixture of ethylene vinyl alcohol copolymer and ethanol yielding a nonadhesive liquid embolic agent to treat cerebral arteriovenous malformations: initial clinical experience. J Neurosurg
2002;97:881 88
16. van Rooij WJ, Sluzewski M, Beute GN. Brain AVM embolization with Onyx.
AJNR Am J Neuroradiol 2007;28:17277; discussion 178
17. Weber W, Kis B, Siekmann R, et al. Preoperative embolization of intracranial
arteriovenous malformations with Onyx. Neurosurgery 2007;61:244 52; discussion 25254
18. Pasqualin A, Barone G, Cioffi F, et al. The relevance of anatomic and hemodynamic factors to a classification of cerebral arteriovenous malformations.
Neurosurgery 1991;28:370 79
19. Katsaridis V, Papagiannaki C, Aimar E. Curative embolization of cerebral arteriovenous malformations (AVMs) with Onyx in 101 patients. Neuroradiology 2008;50:589 97

AJNR Am J Neuroradiol 30:49295 Mar 2009 www.ajnr.org

495