Professional Documents
Culture Documents
Textbook Arteriovenous and Cavernous Malformations Robert F Spetzler Ebook All Chapter PDF
Textbook Arteriovenous and Cavernous Malformations Robert F Spetzler Ebook All Chapter PDF
https://textbookfull.com/product/scale-development-theory-and-
applications-robert-f-devellis/
https://textbookfull.com/product/precalculus-essentials-5th-ed-
robert-f-blitzer/
https://textbookfull.com/product/pocket-atlas-of-emergency-
ultrasound-robert-f-reardon/
https://textbookfull.com/product/vascular-malformations-advances-
and-controversies-in-contemporary-management-byung-boong-lee/
Atlas of Fetal Ultrasound: Normal Imaging and
Malformations 1st Edition Victor Bunduki
https://textbookfull.com/product/atlas-of-fetal-ultrasound-
normal-imaging-and-malformations-1st-edition-victor-bunduki/
https://textbookfull.com/product/information-governance-for-
healthcare-professionals-a-practical-approach-1st-edition-robert-
f-smallwood-author/
https://textbookfull.com/product/the-knee-injury-bible-
everything-you-need-to-know-about-knee-injuries-how-to-treat-
them-and-how-they-affect-your-life-1st-edition-robert-f-laprade/
https://textbookfull.com/product/java-closures-and-lambda-
fischer-robert/
https://textbookfull.com/product/soil-and-environmental-
chemistry-william-f-bleam-auth/
HANDBOOK OF CLINICAL
NEUROLOGY
Series Editors
VOLUME 143
ELSEVIER
Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and retrieval system, without permission in writing
from the publisher. Details on how to seek permission, further information about the Publisher’s permissions
policies, and our arrangements with organizations such as the Copyright Clearance Center and the Copyright
Licensing Agency can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the Publisher (other than
as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience broaden our
understanding, changes in research methods, professional practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any
information, methods, compounds, or experiments described herein. In using such information or methods they
should be mindful of their own safety and the safety of others, including parties for whom they have a professional
responsibility.
With respect to any drug or pharmaceutical products identified, readers are advised to check the most current
information provided (i) on procedures featured or (ii) by the manufacturer of each product to be administered, to
verify the recommended dose or formula, the method and duration of administration, and contraindications. It is the
responsibility of practitioners, relying on their own experience and knowledge of their patients, to make diagnoses,
to determine dosages and the best treatment for each individual patient, and to take all appropriate safety precautions.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for
any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any
use or operation of any methods, products, instructions, or ideas contained in the material herein.
ISBN: 978-0-444-63640-9
Available titles
Vol. 79, The human hypothalamus: basic and clinical aspects, Part I, D.F. Swaab, ed. ISBN 9780444513571
Vol. 80, The human hypothalamus: basic and clinical aspects, Part II, D.F. Swaab, ed. ISBN 9780444514905
Vol. 81, Pain, F. Cervero and T.S. Jensen, eds. ISBN 9780444519016
Vol. 82, Motor neurone disorders and related diseases, A.A. Eisen and P.J. Shaw, eds. ISBN 9780444518941
Vol. 83, Parkinson’s disease and related disorders, Part I, W.C. Koller and E. Melamed, eds. ISBN 9780444519009
Vol. 84, Parkinson’s disease and related disorders, Part II, W.C. Koller and E. Melamed, eds. ISBN 9780444528933
Vol. 85, HIV/AIDS and the nervous system, P. Portegies and J. Berger, eds. ISBN 9780444520104
Vol. 86, Myopathies, F.L. Mastaglia and D. Hilton Jones, eds. ISBN 9780444518996
Vol. 87, Malformations of the nervous system, H.B. Sarnat and P. Curatolo, eds. ISBN 9780444518965
Vol. 88, Neuropsychology and behavioural neurology, G. Goldenberg and B.C. Miller, eds. ISBN 9780444518972
Vol. 89, Dementias, C. Duyckaerts and I. Litvan, eds. ISBN 9780444518989
Vol. 90, Disorders of consciousness, G.B. Young and E.F.M. Wijdicks, eds. ISBN 9780444518958
Vol. 91, Neuromuscular junction disorders, A.G. Engel, ed. ISBN 9780444520081
Vol. 92, Stroke – Part I: Basic and epidemiological aspects, M. Fisher, ed. ISBN 9780444520036
Vol. 93, Stroke – Part II: Clinical manifestations and pathogenesis, M. Fisher, ed. ISBN 9780444520043
Vol. 94, Stroke – Part III: Investigations and management, M. Fisher, ed. ISBN 9780444520050
Vol. 95, History of neurology, S. Finger, F. Boller and K.L. Tyler, eds. ISBN 9780444520081
Vol. 96, Bacterial infections of the central nervous system, K.L. Roos and A.R. Tunkel, eds. ISBN 9780444520159
Vol. 97, Headache, G. Nappi and M.A. Moskowitz, eds. ISBN 9780444521392
Vol. 98, Sleep disorders Part I, P. Montagna and S. Chokroverty, eds. ISBN 9780444520067
Vol. 99, Sleep disorders Part II, P. Montagna and S. Chokroverty, eds. ISBN 9780444520074
Vol. 100, Hyperkinetic movement disorders, W.J. Weiner and E. Tolosa, eds. ISBN 9780444520142
Vol. 101, Muscular dystrophies, A. Amato and R.C. Griggs, eds. ISBN 9780080450315
Vol. 102, Neuro-ophthalmology, C. Kennard and R.J. Leigh, eds. ISBN 9780444529039
Vol. 103, Ataxic disorders, S.H. Subramony and A. Durr, eds. ISBN 9780444518927
Vol. 104, Neuro-oncology Part I, W. Grisold and R. Sofietti, eds. ISBN 9780444521385
Vol. 105, Neuro-oncology Part II, W. Grisold and R. Sofietti, eds. ISBN 9780444535023
Vol. 106, Neurobiology of psychiatric disorders, T. Schlaepfer and C.B. Nemeroff, eds. ISBN 9780444520029
Vol. 107, Epilepsy Part I, H. Stefan and W.H. Theodore, eds. ISBN 9780444528988
Vol. 108, Epilepsy Part II, H. Stefan and W.H. Theodore, eds. ISBN 9780444528995
Vol. 109, Spinal cord injury, J. Verhaagen and J.W. McDonald III, eds. ISBN 9780444521378
Vol. 110, Neurological rehabilitation, M. Barnes and D.C. Good, eds. ISBN 9780444529015
Vol. 111, Pediatric neurology Part I, O. Dulac, M. Lassonde and H.B. Sarnat, eds. ISBN 9780444528919
Vol. 112, Pediatric neurology Part II, O. Dulac, M. Lassonde and H.B. Sarnat, eds. ISBN 9780444529107
Vol. 113, Pediatric neurology Part III, O. Dulac, M. Lassonde and H.B. Sarnat, eds. ISBN 9780444595652
Vol. 114, Neuroparasitology and tropical neurology, H.H. Garcia, H.B. Tanowitz and O.H. Del Brutto, eds.
ISBN 9780444534903
Vol. 115, Peripheral nerve disorders, G. Said and C. Krarup, eds. ISBN 9780444529022
Vol. 116, Brain stimulation, A.M. Lozano and M. Hallett, eds. ISBN 9780444534972
Vol. 117, Autonomic nervous system, R.M. Buijs and D.F. Swaab, eds. ISBN 9780444534910
Vol. 118, Ethical and legal issues in neurology, J.L. Bernat and H.R. Beresford, eds. ISBN 9780444535016
Vol. 119, Neurologic aspects of systemic disease Part I, J. Biller and J.M. Ferro, eds. ISBN 9780702040863
Vol. 120, Neurologic aspects of systemic disease Part II, J. Biller and J.M. Ferro, eds. ISBN 9780702040870
Vol. 121, Neurologic aspects of systemic disease Part III, J. Biller and J.M. Ferro, eds. ISBN 9780702040887
Vol. 122, Multiple sclerosis and related disorders, D.S. Goodin, ed. ISBN 9780444520012
Vol. 123, Neurovirology, A.C. Tselis and J. Booss, eds. ISBN 9780444534880
vi AVAILABLE TITLES (Continued)
Vol. 124, Clinical neuroendocrinology, E. Fliers, M. Korbonits and J.A. Romijn, eds. ISBN 9780444596024
Vol. 125, Alcohol and the nervous system, E.V. Sullivan and A. Pfefferbaum, eds. ISBN 9780444626196
Vol. 126, Diabetes and the nervous system, D.W. Zochodne and R.A. Malik, eds. ISBN 9780444534804
Vol. 127, Traumatic brain injury Part I, J.H. Grafman and A.M. Salazar, eds. ISBN 9780444528926
Vol. 128, Traumatic brain injury Part II, J.H. Grafman and A.M. Salazar, eds. ISBN 9780444635211
Vol. 129, The human auditory system: Fundamental organization and clinical disorders, G.G. Celesia
and G. Hickok, eds. ISBN 9780444626301
Vol. 130, Neurology of sexual and bladder disorders, D.B. Vodušek and F. Boller, eds. ISBN 9780444632470
Vol. 131, Occupational neurology, M. Lotti and M.L. Bleecker, eds. ISBN 9780444626271
Vol. 132, Neurocutaneous syndromes, M.P. Islam and E.S. Roach, eds. ISBN 9780444627025
Vol. 133, Autoimmune neurology, S.J. Pittock and A. Vincent, eds. ISBN 9780444634320
Vol. 134, Gliomas, M.S. Berger and M. Weller, eds. ISBN 9780128029978
Vol. 135, Neuroimaging Part I, J.C. Masdeu and R.G. González, eds. ISBN 9780444534859
Vol. 136, Neuroimaging Part II, J.C. Masdeu and R.G. González, eds. ISBN 9780444534866
Vol. 137, Neuro-otology, J.M. Furman and T. Lempert, eds. ISBN 9780444634375
Vol. 138, Neuroepidemiology, C. Rosano, M.A. Ikram and M. Ganguli, eds. ISBN 9780128029732
Vol. 139, Functional neurologic disorders, M. Hallett, J. Stone and A. Carson, eds. ISBN 9780128017722
Vol. 140, Critical care neurology Part I, E.F.M. Wijdicks and A.H. Kramer, eds. ISBN 9780444636003
Vol. 141, Critical care neurology Part II, E.F.M. Wijdicks and A.H. Kramer, eds. ISBN 9780444635990
Vol. 142, Wilson disease, A. Członkowska and M.L. Schilsky, eds. ISBN 9780444636256
All volumes in the 3rd Series of the Handbook of Clinical Neurology are published electronically, on Science Direct:
http://www.sciencedirect.com/science/handbooks/00729752.
Foreword
The varying clinical presentations of patients with vascular malformations of the central nervous system and its
coverings can make their diagnosis challenging. Refinements in imaging procedures have facilitated the recognition
of these lesions, however, and advances in surgical and endovascular techniques have led to remarkable changes in the
management of affected patients. These developments have generated new questions concerning the need to treat these
malformations in asymptomatic patients in whom they have been encountered incidentally by the sophisticated
imaging procedures now in widespread use. A new understanding of the genetic underpinnings of various vascular
malformations, their pathophysiology, and the manner in which these malformations affect neurologic function has
also led to a more informed approach to patients harboring these lesions. Much has yet to be learned, of course,
and a convenient up-to-date review might well be useful – we hoped – in pointing the way to future advances. For
these various reasons, then, we felt the need to devote a volume of the Handbook of Clinical Neurology to the topic.
We were particularly glad that Professor Robert F. Spetzler, the J.N. Harber Chairman of Neurological Surgery
and Director of the Barrow Neurological Institute in Phoenix, Arizona, agreed to edit this volume because he is
an internationally recognized authority on these lesions, has contributed much to the relevant literature, and has
unrivaled experience in their surgical management. He has been aided in this endeavor by two of his former trainees,
Dr. Karam Moon and Dr. Rami O. Almefty, both also in Phoenix. Together, they have produced a most comprehensive
and wide-ranging text.
The volume is divided into several parts. The first two-thirds, containing some 20 chapters, deals with arteriovenous
malformations from many different viewpoints. Twelve chapters focus on intracranial lesions, covering both cerebral
(eight chapters) and dural vascular anomalies (four chapters), and another eight chapters are devoted to intramedullary
or dural spinal lesions. Different chapters discuss the epidemiology, clinical manifestations, and surgical and endovas-
cular management of these arteriovenous anomalies. The last one-third of the volume contains 11 chapters devoted to
all aspects of cranial and spinal cavernous malformations.
We are grateful to the volume editors, and to the various contributors whom they enlisted as coauthors, for crafting
such a splendid and well-illustrated volume. As series editors, we reviewed all of the chapters for scope, substance, and
style, making suggestions for improvement as needed. Based on our review, it is our belief that the volume will serve as
a valuable reference work for neurologists, neurosurgeons, and interventional radiologists, as well as providing a
practical guide to the management of patients with these lesions.
Elsevier has been the publisher of the Handbook series since its inception, and we are grateful for the continued
support that we have received from the publisher. We acknowledge with particular pleasure, however, our personal
indebtedness to Michael Parkinson in Scotland and to Mara Conner and Kristi Anderson in California for their
assistance in seeing these volumes to fruition.
Michael J. Aminoff
François Boller
Dick F. Swaab
Preface
Vascular malformations, specifically arteriovenous and cavernous malformations, have long fascinated cerebrovascu-
lar surgeons with their complexity and management options. At Barrow Neurological Institute, we have studied the vast
continuum of these malformations affecting our patients, and we have refined the techniques for managing these
lesions. Although considerable knowledge and new treatment strategies have been added to our armamentarium in
recent decades, many questions remain about the optimal management of individual vascular malformations.
Because these lesions can vary widely in size, location, and clinical presentation, the care of individual patients must
be tailored to provide the least risky and most efficacious management available. For the best treatment options, the
judicious use of microsurgery, radiosurgery, and embolization requires expertise from multiple specialists. However,
nonintervention remains preferable for the nonsymptomatic patient who has a grade 5 arteriovenous malformation or
a cavernous malformation under the floor of the fourth ventricle. Thus, the aims of this volume are to improve recog-
nition of the pathology, awareness of the natural history, and understanding of the risks of treatment and nontreatment
of these lesions.
In developing this volume for the Handbook of Clinical Neurology, we have drawn on the experience and expertise
of established leaders in the field to update surgeons, neurologists, trainees, and others on the best current knowledge to
diagnose and treat these often complex and challenging lesions. The chapters in this volume specifically address arte-
riovenous malformations and cavernous malformations. Their subject matter comprises the full spectrum of diagnostic
modalities for the evaluation of patients, considerations crucial to clinical decision-making, the risks and benefits of
treatment, and the latest neurosurgery literature on natural history and outcomes.
The volume is divided into two sections. The first section encompasses arteriovenous malformations, beginning
with chapters that discuss genetics, natural history, and clinical presentation. The next several chapters discuss the
intricacies of treatment of arteriovenous malformations, including the indications for treatment and the multiple treat-
ment modalities that are available. This section also discusses arteriovenous fistulas, which are a close relative of arte-
riovenous malformations. Finally, discussions of spinal arteriovenous malformations and fistulas are included as well.
The second section of the book is devoted to cavernous malformations. The first few chapters are devoted to the
natural history, epidemiology, clinical presentation, and pathophysiology of these lesions. A discussion of develop-
mental venous anomalies, which are closely associated with cavernous malformations, is also included. Finally,
because the presentation of cavernous malformations and the indications for their treatment depend on the location
of these lesions, several chapters are dedicated to their surgical management.
We believe that readers will find this volume of the Handbook of Clinical Neurology to be a valuable contribution to
the existing neurosurgery literature on these complex lesions. Although questions remain about spinal arteriovenous
malformations and fistulas, ongoing research, from the molecular to the clinical level, is focused on providing more
answers.
We sincerely thank all the contributors to this volume for the key roles they have played in bringing this endeavor to
fruition. We hope that our readers glean much from the collective expertise within these pages.
Robert F. Spetzler, MD
Karam Moon, MD
Rami O. Almefty, MD
Contributors
L.F. Gonzalez
M.T. Lawton
Department of Neurosurgery, Duke University, Durham,
Department of Neurological Surgery, University of
NC, USA
California, San Francisco, CA, USA
C. Grady
Department of Neurosurgery, NYU Langone Medical D.D.M. Lin
Center, New York, NY, USA Division of Neuroradiology, Russell H. Morgan
Department of Radiology and Radiological Science,
Johns Hopkins University School of Medicine,
B.A. Gross
Baltimore, MD, USA
Department of Neurosurgery, Brigham and Women’s
Hospital and Harvard Medical School, Boston, MA and
Department of Neurosurgery, Barrow Neurological T.K. Maiti
Institute, St. Joseph’s Hospital and Medical Center, Department of Neurosurgery, Louisiana State University
Phoenix, AZ, USA Health Sciences Center, Shreveport, LA, USA
M.A. Mooney
Department of Neurosurgery, Barrow Neurological C.E. Sarris
Institute, St. Joseph’s Hospital and Medical Center, Department of Neurosurgery, Barrow Neurological
Phoenix, AZ, USA Institute, St. Joseph’s Hospital and Medical Center,
Phoenix, AZ, USA
M.K. Morgan
Department of Clinical Medicine, Macquarie University, J. Schramm
Sydney, New South Wales, Australia Department of Neurosurgery, University of Bonn, Bonn,
Germany
C.B. Mulholland
Department of Neurosurgery, Barrow Neurological J.P. Sheehan
Institute, St. Joseph’s Hospital and Medical Center, Department of Neurosurgery, University of Virginia,
Phoenix, AZ, USA Charlottesville, VA, USA
P. Nakaji
R.A. Solomon
Department of Neurosurgery, Barrow Neurological
Department of Neurosurgery, The Neurological Institute,
Institute, St. Joseph’s Hospital and Medical Center,
Columbia University Medical Center, New York, NY, USA
Phoenix, AZ, USA
Chapter 1
Abstract
Arteriovenous malformations (AVMs) are vascular deformities involving fistula formation of arterial to
venous structures without an intervening capillary bed. Such anomalies can prove fatal as the high arterial
flow can disrupt the integrity of venous walls, thus leading to dangerous sequelae such as hemorrhage.
Diagnosis of these lesions in the central nervous system can often prove challenging as intracranial AVMs
represent a heterogeneous vascular pathology with various presentations and symptomatology. The liter-
ature suggests that most brain AVMs (bAVMs) are identified following evaluation of the etiology of acute
cerebral hemorrhage, or incidentally on imaging associated with seizure or headache workup. Given the
low incidence of this disease, most of the data accrued on this pathology comes from single-center expe-
riences. This chapter aims to distill the most important information from these studies as well as examine
meta-analyses on bAVMs in order to provide a comprehensive introduction into the natural history, clas-
sification, genetic underpinnings of disease, and proposed pathophysiology. While there is yet much to be
elucidated about AVMs of the central nervous system, we aim to provide an overview of bAVM etiology,
classification, genetics, and pathophysiology inherent to the disease process.
*Correspondence to: Alp Ozpinar, MD, Department of Neurological Surgery, University of Pittsburgh Medical Center, Suite B-400,
200 Lothrop Street, Pittsburgh PA, USA. Tel: +1-412-647-3685, E-mail: ozpinara@upmc.edu
6 A. OZPINAR ET AL.
To date, bAVM data have been derived from single- of brain adjacent to the malformation. The five-point
center experiences. From epidemiological studies, the scale assigns a score as follows: for size, <3 cm ¼ 1 point,
incidence of AVMs ranges from 1.12 to 1.42 cases per 3–6 cm ¼ 2 points, >6 cm ¼ 3 points; for location,
100 000 person-years, with 38–68% of new cases pre- eloquent brain ¼ 1 point, noneloquent ¼ 0 points; for
senting as first-ever hemorrhage (Abecassis et al., venous drainage, deep venous drainage ¼ 1 point, not
2014). Annual rates of hemorrhage in untreated bAVMs deep venous drainage ¼ 0 points. Applying this scheme,
have been estimated at 2.10–4.12% (Abecassis et al., higher scores translate to higher Spetzler–Martin grade,
2014). Various studies have tried to identify factors asso- which corresponds to greater risk of morbidity and
ciated with increased risk of AVM rupture. Meta- mortality with microsurgical removal of the lesion. (See
analyses on the topic have found that increased risk of Figure 1.1 for an example of Spetzler–Martin grade 3
future rupture has been associated with factors such as AVM on preoperative imaging, intraoperative photos,
previous rupture, location of AVM in deep brain struc- and intraoperative fluorescence angiography processing.)
tures, and exclusive deep venous drainage (Stapf et al., While the Spetzler–Martin grade is the most-quoted
2006; Gross and Du, 2013). Gross and Du (2013) calcu- benchmark/standard in determining outcome and man-
lated for observed AVMs an overall yearly risk of hem- agement, other grading systems have been introduced
orrhage rate of 3.0%, with an annual rate of 2.2% for the as adjuncts. Lawton et al. proposed a supplementary
unruptured subset and 4.5% for ruptured AVMs. Stapf and complementary system to the Spetzler–Martin
et al. (2006) showed that the annual rate of rupture for grade in 2010 that takes into account age, hemorrhagic
an AVM is 35.5% when the AVM has the triad of deep presentation, and diffuseness of lesion, and yields
venous drainage, deep location, and prior hemorrhage. improved accuracy in predictability of neurologic out-
The other large meta-analysis (Kim et al., 2014) exam- come when combined with the Spetzler–Martin scale
ined four AVM cohorts and found an overall annual hem- (Table 1.1). This supplementary scale in conjunction
orrhage rate of 2.3%, with a rate of 1.3% for unruptured with Spetzler–Martin is a 10-point scale that takes the
AVMs and 4.8% for the ruptured group. These estimates five-point Spetzler–Martin scheme and, in addition,
have held up as results of the multicenter prospective ran- accounts for the following: for age, <20 years ¼ 1 point,
domized trial ARUBA showed a 2.2% annual risk of rup- 20–40 ¼ 2 points, >40 ¼ 3 points; for hemorrhagic pre-
ture for unruptured bAVMs (Mohr et al., 2014). sentation, hemorrhage ¼ 0 points, no hemorrhage ¼ 1
Data for the natural history of ruptured AVMs are less point; for lesion nidus diffuseness, compact ¼ 0 points,
prevalent as the decision to observe ruptured lesions is a diffuse ¼ 1 point. Lawton et al. (2010) have shown
riskier enterprise. Some studies (Kondziolka et al., 1995; that application of their scale can help with preoperative
Brown, 2000) have gone so far as to extrapolate lifetime risk prediction and extrapolation of outcome (Kim
risk of hemorrhage by using annual rates and applying et al., 2014).
the following equation: In 2011, Spetzler and Ponce suggested a consolida-
tion of the Spetzler–Martin scale into a three-tiered sys-
Annual rupture risk ¼ 1 ðrisk of no hemorrhageÞðlife expectancyÞ
tem by combining grades I and II, and grades IV and
and rupture risk ¼ 105 patientage V into classes A and C, respectively. Grade III lesions
were classified as class B in the consolidated system.
While the aforementioned studies have shown concor- Grouping into these new classes was based on similar
dance in estimates of annual risk of rupture, variability surgical results and was intended to provide simplified
among lesions and patient populations presents inherent management recommendations as well as superior sta-
challenges to generalizability and comprehensively pre- tistical power for comparative studies (Spetzler and
dicting risk of rupture, as illustrated in various studies on Ponce, 2011).
the natural history of bAVMs (Pollock et al., 1996; Stapf Another predictive model of neurologic outcome
et al., 2001; Stefani et al., 2002; Fullerton et al., 2005; following bAVM surgery is the University of Toronto
Kim et al., 2010; Laakso et al., 2010; Gross and Brain AVM Study Group’s scale. Proposed by Spears
Du, 2013). and collaborators (2006), the scale is a nine-point strat-
ified risk score where the predictive variable’s influence
is a function of its relative weight (eloquence ¼ 4, dif-
CLASSIFICATION SYSTEMS
fuse nidus ¼ 3, deep venous drainage ¼ 2). Applying
Multiple scales have been formulated to predict the mor- the score, the probability of suffering a disabling neuro-
bidity and mortality associated with AVMs and the asso- logic outcome with surgery is as follows: low risk (0–2
ciated risk of intervention. Spetzler and Martin (1986), points) ¼ 1.8%, moderate risk (3–5 points) ¼ 17.4%,
in their seminal paper on bAVMs, proposed a grading high risk (6–7 points) ¼ 31.6%, very high risk (>7
system based on size, venous drainage, and eloquence points) ¼ 52.9% (Spears et al., 2006).
EPIDEMIOLOGY, GENETICS, PATHOPHYSIOLOGY, AND PROGNOSTIC CLASSIFICATIONS 7
Fig. 1.1. (A) A representative Spetzer–Martin grade 3 arteriovenous malformation (AVM) (3 points for size) of the right frontal
operculum and premotor region is demonstrated on anterior–posterior angiography. (B) Intraoperative photo demonstrates the
dilated draining veins of Labbe and Trolard on the surface of the AVM with the AVM nidus located deep to the draining veins.
(C) Intraoperative flow 800 processing of indocyanine green angiography demonstrates the arrival time of dye to the AVM with red
demonstrating the vessels that fill first (feeders). Seconds in the figure key indicates latency in arrival time of dye. (D) AVM nidus
seen deep to the draining veins along the anterior noneloquent border of the AVM. (E) Intraoperative photo demonstrating the
surgical bed following resection. Note the change in color of the draining veins which are no longer arterialized. Change in color
from red (before resection) to blue (following resection). (F) Postoperative anterior–posterior cerebral angiography demonstrating
complete resection of AVM.
Table 1.1 Along with surgery, other methods have been adopted
to treat bAVMs. Namely, radiosurgical modalities have
Comparing the Spetzler–Martin scale and Lawton
become an established therapeutic option. Given that fac-
supplementary grade (Kim et al., 2014)
tors determining risk of procedure as well as variables
Spetzler–Martin associated with successful surgical resection of intracra-
grading Points Supplementary grading nial AVMs differ from those for radiosurgical manage-
ment, development of a grading system for AVM
Size (cm) Age (years) radiosurgery was conceived. The radiosurgery-based
<3 1 <20 AVM score (RBAS) was introduced following collabora-
3–6 2 20–40 tive efforts between the University of Pittsburgh and the
>6 3 >40 Mayo Clinic. Multiple validation studies have followed
Venous drainage Bleeding since then analyzing its application to lesions in various
Superficial 0 Yes
locations (including brainstem, deep structures, and all
Deep 1 No
locations for AVMs), as well as with different radiosur-
Eloquence Compactness
No 0 Yes gery techniques, including gamma knife, linac and
Yes 1 No CyberKnife (Pollock, 2013). The evolution of the grad-
Total 5 ing scale has reflected a trend towards simplification, as
the original Pittsburgh AVM radiosurgery scale (Pollock
et al., 1997) takes into account AVM volume, patient age,
Of note, the Toronto scale has been found to have a AVM location, embolization status, and number of drain-
superior predictive ability to even the supplementary- ing veins, while more recent iterations of the scale,
Spetzler–Martin scale when looking at area under receiver namely the modified radiosurgery-based AVM score,
operating characteristic curve, but is not as widely adopted focus on AVM location and volume, and patient age
as the multiple iterations of the Spetzler–Martin scheme. (Pollock and Flickinger, 2008) (See Fig. 1.2 for
8 A. OZPINAR ET AL.
60 power of the new grading scale is in helping surgeons
% Modified Rankin Scale Decline
20 GENETICS
Animal models have allowed researchers to observe
10
characteristic changes in nidal vessels of cerebral AVMs,
including nonuniform changes in the thickness of vessel
0
£1.00 1.01-1.50 1.51-2.00 >2.00 walls, lack of tight and adherent junctions, and splitting
Pollock-Flickinger AVM Score of elastic lamina (Tu et al., 2010). These observed
Fig. 1.2. Relationship of modified radiosurgery-based arterio- changes illustrate the manifestation of molecular and cel-
venous malformation (AVM) grading system and decline in lular underpinnings that drive the development of cere-
modified Rankin Scale. Error bars illustrate 95% confidence bral AVMs. The genetic basis of intracranial AVMs is
interval for each point along the curve. (Adapted from still being elucidated, but multiple candidate genes and
Pollock and Flickinger, 2008.) pathways have been identified, both in syndromic and
sporadic AVMs. Cerebral AVMs that result from associ-
illustration of relationship between change in modified ated syndromes provide insight into the etiology of AVM
Rankin Scale and Pollock–Flickinger AVM score.) development. For example, the most common syndrome
The most recent grading scale in AVM management associated with bAVMs is Osler–Weber–Rendu (also
was created to address outcomes in endovascular treat- known as hereditary hemorrhagic telangiectasia), which
ment of intracranial AVMs. Dumont and colleagues has been linked to haplo-insufficiency of transforming
(2015) devised the Buffalo score and retrospectively growth factor (TGF)-b pathway signaling genes like
applied the metric to 50 bAVM patients treated with ENG and SMAD4 (Rangel-Castilla et al., 2014). Another
endovascular embolization, comparing accuracy of com- syndrome of note with increased incidence of AVM
plication prediction to that seen with the Spetzler–Martin formation is Cobb’s syndrome, in which patients are
system. The proposed Buffalo score grade is determined diagnosed with spinal AVMs and exhibit abnormal
by accounting for arterial pedicle number, arterial pedicle expression of platelet endothelial cell adhesion molecule
diameter, and eloquence of nearby cortex (Table 1.2; (PECAM-1), vascular endothelial growth factor (VEGF),
Dumont et al., 2015). Results from this initial retrospec- and matrix metalloproteinase (MMP)-9 (Rangel-Castilla
tive study showed superior correlation of grade and com- et al., 2014).
plication incidence when compared to the Spetzler– Although there is no grand unifying signaling path-
Martin scale for this particular subset of AVM patients way for AVM proliferation in the context of associated
(Dumont et al., 2015). The authors note that part of the syndromes, the dysregulation of angiogenesis, vasculo-
genesis, and inflammation have all been implicated
Table 1.2 (Sturiale et al., 2013). These recurring themes may also
play a leading role in sporadic AVM formation and
Buffalo grading system for endovascular treatment of brain
development. One proposed mechanism associated with
arteriovenous malformations (Dumont et al., 2015)
AVM development is single-nucleotide polymorphisms
Graded feature Points assigned (SNPs) of inflammatory factors like TGF-b and
interleukin-6, and SNPs of vascular growth factors like
Number of arterial pedicles angiopoietin-like 4 glycoprotein (Rangel-Castilla et al.,
1 or 2 1 2014). In the same vein, studies have found overexpres-
3 or 4 2 sion of VEGF and angiopoietin-2 (a factor promoting
5 or more 3 vascular remodeling and destabilization) work in concert
Diameter of arterial pedicles to assist in intracranial AVM development (Kim et al.,
Most >1 mm 0 2009; Moftakhar et al., 2009). Notch4 signaling activa-
Most 1 mm 1
tion has also been characterized as sufficient to induce
Nidus location
AVM phenotype in the developing mouse brain as it pro-
Noneloquent 0
Eloquent 1 duces pathologically large vessels, and shunting physiol-
ogy consistent with AVM (Murphy et al., 2008).
EPIDEMIOLOGY, GENETICS, PATHOPHYSIOLOGY, AND PROGNOSTIC CLASSIFICATIONS 9
Along with AVM development, genetic anomalies findings were a difference of 43.4% in mean arterial
have been found to predispose to AVM rupture. In partic- pressure between ruptured versus unruptured AVMs.
ular, alterations in MMPs can lead to compromise of vas- The nidus or functional unit of the AVM creates a tan-
cular stability and irregular angiogenesis. Studies have gled vascular network that connects feeding arteries and
shown that even plasma levels of MMP-9 are elevated draining veins. Located proximal to the nidus are perini-
in AVM cases relative to controls prior to intervention dal vessels that form a capillary network which some
(Starke et al., 2010). Authors from the University of Cal- believe may contribute to postoperative hemorrhage or
ifornia have also shown the potential for agents which recurrence upon attempted surgical resection of the nidal
block MMP activity; antibiotic medications, including unit (Ogilvy et al., 2001). It is important to note that
doxycycline, may have the potential for inducing a static endovascular embolization of an AVM, which is often
state in which AVM growth or activity can be mitigated done to prevent complications during surgical resection,
(Frenzel et al., 2008). can possibly increase the potential risk of rupture by
Along with a better understanding of the etiology of altering local hemodynamics and re-routing blood flow
bAVMs, genetic analysis allows for identification of tar- within the AVM (Henkes et al., 2004). This may also
gets for future molecular therapies, and therapeutic inhi- be in part due to angiogenesis via upregulation of VEGF
bition of aberrant pathways. and hypoxia-inducible factor (HIF) following partial
embolization (Meyer et al., 1999).
Compartmentalization of AVMs, first described by
PATHOPHYSIOLOGY
Yasargil (1987), is a concept pertaining to a single hemo-
The origin of AVMs remains unclear and is an area of dynamic unit (i.e., compartment) served by one or more
ongoing investigation. Eliciting how these lesions arise feeding pedicles with one or more draining veins. Intrao-
can improve clinical management based on the presen- peratively, as feeding arteries are resected, the compart-
tation, especially in assessment of the risk of rupture ment will lose blood flow, causing it to collapse. Thus,
leading to intracranial hemorrhage. Though intracranial theoretically, if feeders are removed and all compart-
hemorrhage is the most common presentation, others ments of the AVM collapse, hemorrhage would reason-
include seizure and neurologic deficit without underly- ably be avoided. However, Pellettieri et al. (1997) first
ing rupture (Zivin, 2012; Josephson et al., 2015). Norris proposed the idea of hidden compartments following sur-
et al. (1999) conducted a study on 31 patients showing gery to explain the occurrence of postoperative hemor-
that alterations in contrast dilution curves (decreased rhage and edema. If a compartment was unaccounted
time to peak contrast versus increased time to peak con- for, bleeding could be explained if that compartment
trast) are correlated with seizure or hemorrhage. The became subsequently filled. There has been development
interplay of shear force and altered flow dynamics of methods to detect the presence of underlying hidden
mixed with architectural vascular anomalies constitutes compartments, including serial selective digital subtrac-
physiologic changes that shape the understanding of tion angiography and serial high-resolution MR angiog-
AVM development in addition to underlying molecular raphy (Homan et al., 1986; Hashimoto and Nozaki,
mechanisms (Moftakhar et al., 2009). Each respectively 1999). Yamada et al. (2004) established a protocol to
requires further validation. potentially outline compartments to ensure preservation
Feeding artery pressure has been correlated with of surrounding brain tissue.
clinical presentation in conjunction with AVM size Abnormal venous architecture is a factor some have
(Mckissock and Paterson, 1956; Henderson and implicated in AVM pathogenesis (Mullan, 1994; Mullan
Gomez, 1967; Houser et al., 1973; Waltimo, 1973; et al., 1996a). Theories involve architectural changes
Guidetti and Delitala, 1980; Parkinson and Bachers, that either give rise to de novo AVM production or
1980; Itoyama et al., 1989; Spetzler et al., 1992). The perpetuate existing native arteriovenous connections.
relationship of size to risk of rupture has been postu- On histologic examination, endothelial cells seen within
lated to be inversely proportional, or, in other words, AVMs resemble those of early embryologic develop-
smaller AVMs have greater risk of rupture resulting ment despite arterial vasculature of feeding arteries
in intracranial hemorrhage (Norris et al., 1999). How- lacking analogous primitive features (Deshpande and
ever, other studies have failed to support this theory Vidyasagar, 1980). Though invalidated, embryologic
(Gross and Du, 2013). One can infer feeding artery anomalies leading to venous occlusion, stenosis, or
pressures are higher in AVMs that rupture compared agenesis seek to explain subsequent de novo AVM
to those that do not due to the wall stress exerted on development due to venous hypertension (Bederson
the vessels. Spetzler et al. (1992) measured intraopera- et al., 1991; Herman et al., 1995; Lawton et al.,
tive perfusion pressure plus mean arterial pressure on 1997). As the role of venous hypertension has yet to
24 patients while documenting the AVM size. Key be definitively determined, an alternative hypothesis
10 A. OZPINAR ET AL.
describes angiogenesis originating from venous struc- compensation following degeneration of normal vessels
tures due to venous hypertension causing transformation serving redundant areas, AVM production may occur as
to AVM (Bederson et al., 1991; Wilson, 1992). Hypo- blood flow through shunted vessels increases. Related to
xia resulting from vascular pressure changes may stimu- the earlier discussion of perinidal vessels, hemodynamic
late the formation of arteriovenous fistulas thought to overload state can result due to a hypervascular perinidal
arise from venous structures (Wilson, 1992). This would network potentially connected to the nidus termed
establish abnormal vascular architecture that increases “modja-modja” (Takemae et al., 1993). Solely following
venous pressure. Lastly, fistula formation of low-flow surgical resection of AVM, intravascular pressure can
malformations has been implicated based on similar char- increase in this feeble vascular network, leading to rup-
acteristics with select AVMs (abnormal surface and deep ture and hemorrhage.
vein drainage plus the presence of deep collecting veins), Dilation of feeding arteries deprives alternate vessels of
accounting for the belief that low-flow malforma- blood flow, shunting blood to the AVM under increased
tions provide an architectural platform for AVMs to arise pressure (Taylor et al., 2002). The presence of vascular
(Mullan, 1994; Mullan et al., 1996b; Pietil€a et al., 2000). steal phenomenon, where blood is “stolen” from other
Abnormal venous drainage, specifically solely deep regions and redirected toward the AVM, can be defined
drainage, is a supported finding in the literature for pos- as such. Support for this theory comes from CT imaging
sible formation of intracranial AVMs, and interestingly, studies, both single-photon emission and Xe-CT, that dem-
de novo AVMs have often been found to drain into onstrate decreased blood flow in areas adjacent to AVMs
low-flow malformations (Nataf et al., 1997). It is thus (Okabe et al., 1983; Homan et al., 1986). One observation
hypothesized that low-flow malformations may contrib- describes cerebral calcifications arising, which is theorized
ute to the generation of AVMs. Previous authors have to arise from steal phenomenon (Yu et al., 1987). Addition-
also demonstrated that AVMs, cavernous malformations, ally, neuropsychologic studies have described a relation-
and capillary teleangiectasias may exist on a spectrum ship between the potential occurrence of vascular steal
and may be related, as demonstrated by the sequential phenomenon with lesion size, peripheral venous drainage,
formation of all three entities in a single patient (Abla and the development of abnormal feeding arteries (Marks
et al., 2008). et al., 1991). In theory, steal may play a role in risk of cere-
Native arteriovenous connections are thought to bral edema postoperatively. Following lesion resection,
become pathogenic following a venous vaso-occlusive increased blood flow through vessels previously deprived
event progressing to AVM as blood flow is redirected of fluid volume could account for this occurrence along
through preformed connections from occluded vessels with underlying molecular mechanisms of injury/ischemia
(Hasegawa et al., 1967). This postulate has been supported (i.e., HIF and VEGF) (Moftakhar et al., 2009). This may
in animal studies, but is still in early investigational stages be due to chronic hypoxia with loss of capillaries and
as to its validity in human models. In regard to rupture risk, hypoperfusion, leading to increased production of both
those abnormalities that increase venous pressure (venous HIF and VEGF (Meyer et al., 1999).
stenosis, predominant deep drainage, and venous reflux) Some conflicting data surrounding the actual pres-
have been associated with hemorrhage (Awad et al., ence of AVM steal come from studies that provide no
1993; al-Rodhan, 1995; Mullan et al., 1996a). Venous support for the presence of steal in cerebral vascular net-
abnormalities are thus evidenced to have a role in AVM works with AVMs (Mast et al., 1995; Meyer et al., 1998).
pathogenesis, yet other factors still remain pertinent. Meyer and collaborators (1998) conducted a study that
One such factor is the role of autoregulation. showed that blood flow is maintained at relatively nor-
The relationship between hemodynamics and AVM mal levels, calling to question the existence of steal phe-
pathogenesis remains unclear. One postulate pertains to nomenon in AVM physiology. Advanced imaging
abnormal autoregulation, which may lead to changes techniques have been used to further investigate these
in perfusion pressures. Spetzler et al. (1978) describe a discrepancies, namely MR perfusion studies, which have
theory whereby the presence of an AVM leads to local described abnormal blood flow regulation (Bambakidis
loss of autoregulation, described as normal perfusion et al., 2001; Guo et al., 2004). Additional studies are still
pressure breakthrough theory. Those in opposition to this needed to provide further insight into the in vivo charac-
theory, however, believe paradoxically that autoregula- teristics of “steal.”
tion (or dysautoregulation) gives rise to the AVM rather
than being caused by the presence of one (Quick et al.,
CONCLUSION
2001). In this instance, loss of autoregulation causes
venous hypertension that decreases perfusion pressure, In conclusion, AVMs of the CNS represent a fascinating
ultimately resulting in reduction of blood flow and and complex disease entity with a presentation profile
AVM development. In essence, as a mode of consisting of with hemorrhage, seizure, or headache.
EPIDEMIOLOGY, GENETICS, PATHOPHYSIOLOGY, AND PROGNOSTIC CLASSIFICATIONS 11
Various classification systems are described, most nota- arteriovenous malformations. Stroke; a Journal of Cere-
bly the Spetzler–Martin grading system, the Lawton– bral Circulation 36: 2099–2104.
Young supplementary grading scheme, the Pollock– Gross BA, Du R (2013). Natural history of cerebral arteriove-
Flickinger AVM score for radiosurgery, and the newly nous malformations: a meta-analysis. J Neurosurg 118:
437–443.
described Buffalo score for endovascular treatment.
Guidetti B, Delitala A (1980). Intracranial arteriovenous mal-
AVMs can occur in association with hereditary genetic
formations. Conservative and surgical treatment.
syndromes; they can also involve sporadic development J Neurosurg 53: 149–152.
and formation, which may relate to abnormalities in cere- Guo W-Y, Wu Y-T, Wu H-M et al. (2004). Toward normal
bral vascular autoregulation as well as abnormalities in perfusion after radiosurgery: perfusion MR imaging
venous architecture. Finally, the entities of normal perfu- with independent component analysis of brain arterio-
sion pressure breakthrough as part of a global process of venous malformations. AJNR Am J Neuroradiol 25:
loss of cerebral autoregulation and the process described 1636–1644.
as “steal” are fascinating pathophysiology described in Hasegawa T, Ravens JR, Toole JF (1967). Precapillary arterio-
AVM literature and are the subjects of ongoing study. venous anastomoses. “Thoroughfare channels” in the
brain. Arch Neurol 16: 217–224.
Hashimoto N, Nozaki N (1999). Do cerebral arteriovenous
REFERENCES
malformations recur after angiographically confirmed total
Abecassis IJ, Xu DS, Batjer HH et al. (2014). Natural history of extirpation? Critical reviews in neurosurgery: CR 9:
brain arteriovenous malformations: a systematic review. 141–146.
Neurosurg Focus 37: E7. Henderson WR, Gomez RD (1967). Natural history of cerebral
Abla A, Wait SD, Uschold T et al. (2008). Developmental angiomas. Br Med J 4: 571–574.
venous anomaly, cavernous malformation, and capillary Henkes H, Gotwald TF, Brew S et al. (2004). Pressure mea-
telangiectasia: spectrum of a single disease. Acta surements in arterial feeders of brain arteriovenous malfor-
Neurochir 150: 487–489; discussion 489. mations before and after endovascular embolization.
Al-Rodhan NR (1995). Occlusive hyperemia remains the most Neuroradiology 46: 673–677.
logical explanation for the hemodynamic complications of Herman JM, Spetzler RF, Bederson JB et al. (1995). Genesis of
resected intracerebral arteriovenous malformations. a dural arteriovenous malformation in a rat model.
J Neurosurg Anesthesiol 7: 208–210. J Neurosurg 83: 539–545.
Awad IA, Robinson JR, Mohanty S et al. (1993). Mixed vas- Hofmeister C, Stapf C, Hartmann A et al. (2000).
cular malformations of the brain: clinical and pathogenetic Demographic, morphological, and clinical character-
considerations. Neurosurgery 33: 179–188; discussion 188. istics of 1289 patients with brain arteriovenous malfor-
Bambakidis NC, Sunshine JL, Faulhaber PF et al. (2001). mation. Stroke; a Journal of Cerebral Circulation 31:
Functional evaluation of arteriovenous malformations. 1307–1310.
Neurosurg Focus. 11e2. Homan RW, Devous MD, Stokely EM et al. (1986).
Bederson JB, Wiestler OD, Br€ ustle O et al. (1991). Intracranial Quantification of intracerebral steal in patients with arterio-
venous hypertension and the effects of venous outflow venous malformation. Arch Neurol 43: 779–785.
obstruction in a rat model of arteriovenous fistula. Houser OW, Baker HL, Svien HJ et al. (1973). Arteriovenous
Neurosurgery 29: 341–350. malformations of the parenchyma of the brain.
Brown RD (2000). Simple risk predictions for arteriovenous Angiographic aspects Radiology 109: 83–90.
malformation hemorrhage. Neurosurgery 46: 1024. Itoyama Y, Uemura S, Ushio Y et al. (1989). Natural course of
Conger A, Kulwin C, Lawton MT et al. (2015). Diagnosis and unoperated intracranial arteriovenous malformations:
evaluation of intracranial arteriovenous malformations. study of 50 cases. J Neurosurg 71: 805–809.
Surg Neurol Int 6: 76. Josephson CB, Rosenow F, Al-Shahi Salman R (2015).
Deshpande DH, Vidyasagar C (1980). Histology of the persis- Intracranial vascular malformations and epilepsy. Semin
tent embryonic veins in arteriovenous malformations of Neurol 35: 223–234.
brain. Acta Neurochir 53: 227–236. Kim H, Pawlikowska L, Chen Y et al. (2009). Brain arteriove-
Dumont TM, Kan P, Snyder KV et al. (2015). A propo- nous malformation biology relevant to hemorrhage and
sed grading system for endovascular treatment of cere- implication for therapeutic development. Stroke; a
bral arteriovenous malformations: Buffalo score. Surg Journal of Cerebral Circulation 40: S95–S97.
Neurol Int 6: 3. Kim H, McCulloch CE, Johnston SC et al. (2010). Comparison
Frenzel T, Lee CZ, Kim H et al. (2008). Feasibility of minocy- of 2 approaches for determining the natural history risk of
cline and doxycycline use as potential vasculostatic therapy brain arteriovenous malformation rupture. Am J Epidemiol
for brain vascular malformations: pilot study of adverse 171: 1317–1322.
events and tolerance. Cerebrovascular Diseases (Basel, Kim H, Al-Shahi Salman R, McCulloch CE et al. (2014).
Switzerland) 25: 157–163. Untreated brain arteriovenous malformation: patient-level
Fullerton HJ, Achrol AS, Johnston SC et al. (2005). Long-term meta-analysis of hemorrhage predictors. Neurology 83:
hemorrhage risk in children versus adults with brain 590–597.
12 A. OZPINAR ET AL.
Kondziolka D, McLaughlin MR, Kestle JR (1995). Simple risk Norris JS, Valiante TA, Wallace MC et al. (1999). A simple
predictions for arteriovenous malformation hemorrhage. relationship between radiological arteriovenous malfor-
Neurosurgery 37: 851–855. mation hemodynamics and clinical presentation: a pro-
Laakso A, Dashti R, Juvela S et al. (2010). Natural history of spective, blinded analysis of 31 cases. J Neurosurg 90:
arteriovenous malformations: presentation, risk of hemor- 673–679.
rhage and mortality. Acta Neurochirurg Suppl 107: 65–69. Ogilvy CS, Stieg PE, Awad I et al. (2001). AHA Scientific
Lawton MT, Jacobowitz R, Spetzler RF (1997). Redefined role Statement: Recommendations for the management of intra-
of angiogenesis in the pathogenesis of dural arteriovenous cranial arteriovenous malformations: a statement for
malformations. J Neurosurg 87: 267–274. healthcare professionals from a special writing group of
Lawton MT, Kim H, McCulloch CE et al. (2010). A supplemen- the Stroke Council, American Stroke Association.
tary grading scale for selecting patients with brain arteriove- Stroke; a Journal of Cerebral Circulation 32: 1458–1471.
nous malformations for surgery. Neurosurgery 66: 702–713; Okabe T, Meyer JS, Okayasu H et al. (1983). Xenon-enhanced
discussion 713. CT CBF measurements in cerebral AVM’s before and after
Marks MP, Lane B, Steinberg G et al. (1991). Vascular char- excision. Contribution to pathogenesis and treatment.
acteristics of intracerebral arteriovenous malformations in J Neurosurg 59: 21–31.
patients with clinical steal. AJNR Am J Neuroradiol 12: Parkinson D, Bachers G (1980). Arteriovenous malformations.
489–496. Summary of 100 consecutive supratentorial cases.
Mast H, Mohr JP, Osipov A et al. (1995). ‘Steal’ is an unestab- J Neurosurg 53: 285–299.
lished mechanism for the clinical presentation of cerebral Pellettieri L, Svendsen P, Wikholm G et al. (1997). Hidden
arteriovenous malformations. Stroke; a Journal of compartments in AVMs – a new concept. Acta
Cerebral Circulation 26: 1215–1220. Radiologica (Stockholm, Sweden: 1987) 38: 2–7.
Mckissock W, Paterson JH (1956). A clinical survey of intra- Pietil€a TA, Zabramski JM, Thèllier-Janko A et al. (2000).
cranial angiomas with special reference to their mode of Animal model for cerebral arteriovenous malformation.
progression and surgical treatment: a report of 110 cases. Acta Neurochir 142: 1231–1240.
Brain: A Journal of Neurology 79: 233–266. Pollock BE (2013). Development and testing of a
Meyer B, Schaller C, Frenkel C et al. (1998). Physiological radiosurgery-based arteriovenous malformation grading
steal around AVMs of the brain is not equivalent to cortical system. Prog Neurol Surg 27: 58–66.
ischemia. Neurol Res 20 (Suppl 1): S13–S17. Pollock BE, Flickinger JC (2008). Modification of the
Meyer B, Schaller C, Frenkel C et al. (1999). Distributions of radiosurgery-based arteriovenous malformation grading
local oxygen saturation and its response to changes of mean system. Neurosurgery 63: 239–243; discussion 243.
arterial blood pressure in the cerebral cortex adjacent to Pollock BE, Flickinger JC, Lunsford LD et al. (1996). Factors
arteriovenous malformations. Stroke; a Journal of that predict the bleeding risk of cerebral arteriovenous
Cerebral Circulation 30: 2623–2630. malformations. Stroke; a Journal of Cerebral Circulation
Moftakhar P, Hauptman JS, Malkasian D et al. (2009). 27: 1–6.
Cerebral arteriovenous malformations. Part 2: physiology. Pollock BE, Flickinger JC, Lunsford LD et al. (1997). The
Neurosurg Focus 26: E11. Pittsburgh arteriovenous malformation radiosurgery
Mohr JP, Parides MK, Stapf C et al. (2014). Medical manage- (PAR) grading scale. Radiosurgery 2: 137–146.
ment with or without interventional therapy for unrup- Quick CM, Hashimoto T, Young WL (2001). Lack of flow reg-
tured brain arteriovenous malformations (ARUBA): a ulation may explain the development of arteriovenous mal-
multicentre, non-blinded, randomised trial. Lancet 383: formations. Neurol Res 23: 641–644.
614–621. Rangel-Castilla L, Russin JJ, Martinez-Del-Campo E et al.
Mullan S (1994). Reflections upon the nature and management (2014). Molecular and cellular biology of cerebral arterio-
of intracranial and intraspinal vascular malformations and venous malformations: a review of current concepts and
fistulae. J Neurosurg 80: 606–616. future trends in treatment. Neurosurg Focus 37: E1.
Mullan S, Mojtahedi S, Johnson DL et al. (1996a). Cerebral Spears J, Terbrugge KG, Moosavian M et al. (2006).
venous malformation-arteriovenous malformation transi- A discriminative prediction model of neurological outcome
tion forms. J Neurosurg 85: 9–13. for patients undergoing surgery of brain arteriovenous mal-
Mullan S, Mojtahedi S, Johnson DL et al. (1996b). formations. Stroke; a Journal of Cerebral Circulation 37:
Embryological basis of some aspects of cerebral vascular 1457–1464.
fistulas and malformations. J Neurosurg 85: 1–8. Spetzler RF, Martin NA (1986). A proposed grading sys-
Murphy PA, Lam MTY, Wu X et al. (2008). Endothelial tem for arteriovenous malformations. J Neurosurg 65:
Notch4 signaling induces hallmarks of brain arteriovenous 476–483.
malformations in mice. Proc Natl Acad Sci U S A 105: Spetzler RF, Ponce FA (2011). A 3-tier classification of cere-
10901–10906. bral arteriovenous malformations. Clinical article.
Nataf F, Meder JF, Roux FX et al. (1997). Angioarchitecture J Neurosurg 114: 842–849.
associated with haemorrhage in cerebral arteriovenous mal- Spetzler RF, Wilson CB, Weinstein P et al. (1978). Normal
formations: a prognostic statistical model. Neuroradiology perfusion pressure breakthrough theory. Clin Neurosurg
39: 52–58. 25: 651–672.
EPIDEMIOLOGY, GENETICS, PATHOPHYSIOLOGY, AND PROGNOSTIC CLASSIFICATIONS 13
Spetzler RF, Hargraves RW, McCormick PW et al. (1992). Taylor CL, Selman WR, Ratcheson RA (2002). Steal affecting
Relationship of perfusion pressure and size to risk the central nervous system. Neurosurgery 50: 679–688;
of hemorrhage from arteriovenous malformations. discussion 688–689.
J Neurosurg 76: 918–923. Tu J, Karunanayaka A, Windsor A et al. (2010). Comparison of
Stapf C, Mohr JP, Pile-Spellman J et al. (2001). Epidemiology an animal model of arteriovenous malformation with
and natural history of arteriovenous malformations. human arteriovenous malformation. Journal of Clinical
Neurosurg Focus. 11. e1. Neuroscience: Official Journal of the Neurosurgical
Stapf C, Mast H, Sciacca RR et al. (2006). Predictors of hem- Society of Australasia 17: 96–102.
orrhage in patients with untreated brain arteriovenous mal- Waltimo O (1973). The relationship of size, density and local-
formation. Neurology 66: 1350–1355. ization of intracranial arteriovenous malformations to the
Starke RM, Komotar RJ, Hwang BY et al. (2010). Systemic type of initial symptom. J Neurol Sci 19: 13–19.
expression of matrix metalloproteinase-9 in patients with Wilson CB (1992). Cryptic vascular malformations. Clin
cerebral arteriovenous malformations. Neurosurgery 66: Neurosurg 38: 49–84.
343–348; discussion 348. Yamada S, Brauer FS, Colohan ART et al. (2004). Concept of
Stefani MA, Porter PJ, Terbrugge KG et al. (2002). Large and arteriovenous malformation compartments and surgical
deep brain arteriovenous malformations are associated management. Neurol Res 26: 288–300.
with risk of future hemorrhage. Stroke; a Journal of Yasargil MG (1987). Pathological considerations. Microneuro-
Cerebral Circulation 33: 1220–1224. surgery: AVM of the Brain, History, Embryology, Patholo-
Sturiale CL, Puca A, Sebastiani P et al. (2013). Single nucle- gical Considerations, Hemodynamics, Diagnostic Studies,
otide polymorphisms associated with sporadic brain arte- Microsurgical Anatomy, Thieme Verlag, Stuttgart.
riovenous malformations: where do we stand? Brain: Yu YL, Chiu EK, Woo E et al. (1987). Dystrophic intra-
A Journal of Neurology 136: 665–681. cranial calcification: CT evidence of ‘cerebral steal’
Takemae T, Kobayashi S, Sugita K (1993). Perinidal hyper- from arteriovenous malformation. Neuroradiology 29:
vascular network on immediate postoperative angiogram 519–522.
after removal of large arteriovenous malformations located Zivin J (2012). Hemorrhagic Cerebrovascular Disease.
distant from the arterial circle of Willis. Neurosurgery 33: Goldman’s Cecil Medicine, 24th ed. Saunders Elsevier,
400–405; discussion 405–406. Philadelphia, PA.
Handbook of Clinical Neurology, Vol. 143 (3rd series)
Arteriovenous and Cavernous Malformations
R.F. Spetzler, K. Moon, and R.O. Almefty, Editors
http://dx.doi.org/10.1016/B978-0-444-63640-9.00002-3
© 2017 Elsevier B.V. All rights reserved
Chapter 2
Abstract
Cerebral arteriovenous malformations (AVMs) are composed of a complex tangle of abnormal arteries and
veins and are a significant source of cerebral hemorrhage and consequent morbidity and mortality in young
adults, representing a diagnostic and therapeutic challenge. Current natural-history studies of cerebral
AVMs report overall annual rates of 1% and 3% for the risk of epilepsy and hemorrhage, respectively.
Unruptured AVMs have an annual hemorrhage rate of 2.2% while ruptured lesions have an annual hem-
orrhage rate of 4.5%. These hemorrhage rates are can change over time, particularly for hemorrhagic
lesions, with the rebleed rate ranging from 6% to 15.8% in the first year after rupture across several studies.
Besides hemorrhage, other significant risk factors for AVM hemorrhage include deep location, deep
venous drainage, associated aneurysms, and pregnancy. Other factors include patient age, sex, and small
AVM size, which are not currently considered significant risk factors for AVM hemorrhage. In addition to
hemorrhage risk and seizure risk, the natural history of an AVM also encompasses the daily psychologic
burden that a patient must endure knowing that he or she possesses an untreated AVM. This chapter
reviews the epidemiology, clinical features, and natural history of cerebral AVMs.
*Correspondence to: Rose Du, MD, PhD, Department of Neurosurgery, Brigham and Women’s Hospital and Harvard Medical
School, 75 Francis St, Boston MA 02115, USA. Tel: +1-617-732-6600, Fax: 617-734-8342, E-mail: rdu@bwh.harvard.edu
16 A. CAN ET AL.
the anticipated hemorrhage risk and accompanying hemorrhagic stroke. Population-based natural-history
morbidity and mortality; it also includes the future risk studies have reported an approximately 50% rate of hem-
of seizures as well as the underreported daily psycho- orrhagic presentation, with intracerebral hemorrhage as
logic burden and consequent anxiety of harboring a the most common type of bleed, followed by intraven-
life-threatening intracranial abnormality. tricular and subarachnoid hemorrhages (Stapf et al.,
2006; Kim et al., 2007; da Costa et al., 2009; Gross
DEMOGRAPHICS and Du, 2013). In a recent meta-analysis, the overall rate
of hemorrhagic presentation was 52% (95% CI 48–56%)
The true incidence and prevalence of cerebral AVMs
(Gross and Du, 2013). AVM rupture typically presents in
remain unknown, due to the relative rarity of the disease
the third or fourth decade of life, and in a retrospective,
and lack of large-scale epidemiologic studies. Across
hospital-based study of patients under 40 years of age
autopsy studies, the prevalence of AVMs has ranged from
with intracerebral hemorrhage, AVMs were the leading
5 to 613 cases per 100 000 (Courville, 1967; Stapf et al.,
cause of intracerebral hemorrhage, affecting 33% (95%
2001). Al-Shahi and colleagues (2002) found a crude
CI 40–85%) of patients (Ruiz-Sandoval et al., 1999).
AVM prevalence rate of 15–18 per 100 000 adults in a ret-
In addition to hemorrhage, seizures have been reported
rospective community-based study in Scotland. In the
as the second most common presentation modality, seen
New York Islands study, a prospective population-based
in 27% of patients (95% CI 24–30%) (Gross and Du,
survey in which the incidence of AVM-related hemorrhage
2013). Other less common presenting symptoms include
and the associated rates of morbidity and mortality were
headaches or focal neurologic deficits due to mass effect
determined, the annual AVM detection rate was 1.34 per
or, though rare, as sources of ischemic symptoms as a
100 000 person-years (95% confidence interval (CI)
result of steal phenomena.
1.18–49) (Stapf et al., 2001). Other studies reported detec-
Headaches in the absence of hemorrhage are reported
tion rates of 1.12 (95% CI 0.90–1.37) and 1.11 (95% CI
to occur in approximately 6–14% of patients with AVMs
0.6–1.8) per 100 000 person-years (Brown et al., 1996;
and Waltimo et al. (1975) reported that 58% of women
Al-Shahi et al., 2003).
with AVMs were found to have migraine with or without
Despite their presumed congenital origin, cerebral
aura (Fults and Kelly, 1984; Ondra et al., 1990; Brown
AVMs are usually not hereditary lesions, and a recent
et al., 2005). The headaches are usually unilateral;
study evaluating the prevalence of AVMs in first-degree
however, more generalized headaches due to elevated
relatives of patients harboring these lesions suggested
venous pressures have also been reported (Al-Shahi
against a familial risk factor (van Beijnum et al.,
and Warlow, 2001). However, the lack of prospective
2014). However, aside from cases of hereditary hemor-
studies with validated diagnostic criteria hampers draw-
rhagic telangiectasia, rare cases of familial occurrence
ing definite conclusions regarding the association
have been reported, supporting the hypothesis that
between headaches and AVMs, and more prospective
genetic factors may play a role in the etiology (van
longitudinal natural-history studies with standardized
Beijnum et al., 2007).
diagnostic headache criteria are needed.
Six large natural-history studies with at least 200
Although rare, cerebral AVMs may cause focal neuro-
patients with cerebral AVMs reported a mean age in a
logic deficits and signs of ischemia in the absence of prior
tight range of 32–39 years of age (Pollock et al., 1996;
or concomitant cerebral hemorrhage in approximately
Stapf et al., 2006; Kim et al., 2007; Yamada et al.,
3–10% of patients (Crawford et al., 1986; Itoyama
2007; Hernesniemi et al., 2008; da Costa et al., 2009).
et al., 1989; Kader et al., 1994; Mast et al., 1997;
Indeed, in a recent meta-analysis of cerebral AVMs,
Al-Shahi and Warlow, 2001; Halim et al., 2004; Stapf
the mean age at presentation was reported to be
et al., 2006; Yamada et al., 2007). Focal neurologic def-
33.7 years (95% CI 31.1–36.2), underscoring that this
icits can present with sudden onset and their course may
is an abnormality discovered in younger patients
vary from transient to progressive. The multifactorial
(Gross and Du, 2013).
pathophysiology includes vascular steal phenomenon,
Most studies do not report a significant sex predispo-
resulting from high-flow shunting through the AVM with
sition for patients with cerebral AVMs, including studies
consequent low blood pressure in the surrounding arter-
of pediatric patients (Stapf et al., 2006; Kim et al., 2007;
ies and brain tissue (Mast et al., 1995). Other contributing
da Costa et al., 2009; Darsaut et al., 2011).
factors are venous hypertension causing hypoperfusion
of the surrounding brain parenchyma, and mass effect
CLINICAL PRESENTATION
of the AVM or hydrocephalus. Other modes of AVM pre-
Regardless of study design (natural history vs. surgical/ sentation that have been reported in the literature include
interventional series), the most commonly reported pre- cognitive dysfunction (Olivecrona and Riives, 1948),
sentation modality for patients with cerebral AVMs is learning and behavioral disorders (Lazar et al., 1999),
THE NATURAL HISTORY OF CEREBRAL ARTERIOVENOUS MALFORMATIONS 17
pulsatile tinnitus (Sabra, 1959), raised intracranial pres- sex, and frontotemporal lesion location were significant
sure (Chimowitz et al., 1990), and specific deficits asso- risk factors for seizures on presentation; deep artery per-
ciated with location of AVMs, such as movement forators were associated with postoperative seizures
disorders by AVMs located in the basal ganglia (Englot et al., 2012).
(Lobo-Antunes et al., 1974) or cranial nerve palsies Although several studies have demonstrated clear
(Hatori et al., 1991), trigeminal neuralgia (Johnson and risk factors for seizures at presentation, this important
Salmon, 1968), and hemifacial spasm caused by AVMs feature is not as well represented in the prospective
in the posterior fossa (Kim et al., 1991). natural-history literature (Englot et al., 2012; Galletti
In addition, with advances in diagnostic radiology and et al., 2014). A recent observational analysis of 229
the availability of high-resolution imaging techniques, adults with a new diagnosis of an AVM reported a
asymptomatic AVMs are being detected more often, with 72% 5-year risk of recurrent seizures in patients present-
an overall rate of 10% in modern series (Pollock et al., ing with an initial seizure. Of patients presenting with
1996; Stapf et al., 2006; Kim et al., 2007; Yamada seizures, by 2 years, 76% will develop epilepsy
et al., 2007; Hernesniemi et al., 2008; da Costa et al., (Josephson et al., 2012). In patients presenting with hem-
2009; Gross and Du, 2013). orrhage, the 5-year risk of a first unprovoked seizure was
26%. In a natural-history study of 217 patients with
AVMs, the risk of epilepsy for all-comers by 10-year
SEIZURE RISK
follow-up was 11%, increasing to 18% by 20-year
The second most common presenting manifestation of follow-up. This corresponds to an annual de novo seizure
cerebral AVMs is seizures (Thorpe et al., 2000), with risk of 1% (Crawford et al., 1986). Twenty-two percent
approximately 17–30% of AVM patients presenting with of patients presenting with hemorrhage developed epi-
this symptom (Galletti et al., 2014; Spetzler et al., 2015). lepsy within 20 years of diagnosis, representing the
However, the exact pathogenesis of seizures caused by greatest risk factor for de novo epilepsy. In addition,
cerebral AVMs remains unclear. Hemosiderin deposi- the younger the patient was at diagnosis, the higher the
tion, mass effect with cortical irritation, hemodynamic risk of developing epilepsy during follow-up, with rates
modifications, and/or vascular remodeling leading to ranging from 44% for patients aged 10–19 and 6% for
steal, ischemia, and neuronal damage have all been sug- patients over the age of 30. Size or depth of the AVM
gested as pathophysiologic etiologies to epilepsy result- did not influence de novo epilepsy development
ing from cerebral AVMs (Turjman et al., 1995a; (Crawford et al., 1986).
Moftakhar et al., 2009).
Some studies have reported risk factors for seizure
HEMORRHAGE RISK
presentation in patients with AVMs. In one study, frontal,
temporal, and/or superficial topography were significant Estimating the natural history of AVMs has been chal-
risk factors for seizure presentation, with the strongest lenged by the heterogeneity of AVM lesions and diverse
association for temporal-lobe location (odds ratio (OR) patient populations, with a wide range of clinical presen-
3.48; 95% CI 1.77–6.85) (Galletti et al., 2014). Other tations and outcomes, in addition to surgeon or institu-
studies also confirmed frontal and temporal location to tional bias toward or against treatment. Hemorrhage is
be correlated with seizure presentation (Perret and often considered the most common source of morbidity
Nishioka, 1966; Morello and Borghi, 1973) in addition and mortality from an AVM and is therefore the main
to other angioarchitectural features such as cortical loca- focus of most natural-history studies that seek to identify
tion of the AVM, feeding by the middle cerebral artery, risk factors predicting hemorrhage. In one long-term
cortical location of the feeder, absence of aneurysms, follow-up study of 168 patients with unruptured cerebral
and presence of varix/varices (Turjman et al., 1995a). AVMs, morbidity and mortality from AVM hemorrhage
In Galletti et al.’s (2014) study, 31% of patients presented were 35% and 29%, respectively (Brown et al., 1988). In
with seizures as the first clinical presentation of cerebral another recent retrospective study of 51 patients sustain-
AVM, while 80% of these patients had epilepsy which ing a hemorrhage from a previously untreated AVM,
led to the discovery of the lesion, and the remaining 74% of patients who survived had neurologic deficits
20% had a single seizure at the time of diagnosis. Nearly upon discharge, with 25% of patients having a severe
half of the patients had simple partial seizures, 13% com- deficit (Majumdar et al., 2015). On follow-up, 55% of
plex partial, 16% partial evolving to secondary general- patients were independent in their daily activities of
ized seizures, and only 3% generalized seizures (Galletti living. The mortality rate was 8%. In a pediatric study
et al., 2014). In the remaining 23% of patients, the clin- of 115 patients, 68% of those sustaining a hemorrhage
ical pattern could not be established (Galletti et al., were independent (Gross et al., 2015). Overall, estimated
2014). In another study, prior AVM hemorrhage, male hemorrhage-related mortality rates range from 10% to
18 A. CAN ET AL.
30%, and morbidity rates from 20% to 30% (Crawford hemorrhage (Yamada et al., 2007; Hernesniemi et al.,
et al., 1986; Brown et al., 1988; Ondra et al., 1990). 2008; Gross and Du, 2013). This was confirmed in a
Numerous studies have evaluated AVM hemorrhage recent meta-analysis where small AVM size was not
risk, and there is general consensus that the overall associated with an increased risk of hemorrhage (Gross
annual hemorrhage rate for AVMs ranges from 2% to and Du, 2013). Significant risk factors for hemorrhage
4%. In a retrospective series of 238 patients with include prior hemorrhage, exclusively deep venous
untreated AVMs and an average follow-up of 13.5 years, drainage, deep location, associated aneurysms, and
the average annual risk of hemorrhage from AVM was pregnancy (da Costa et al., 2009; Gross and Du,
2.4% (Hernesniemi et al., 2008). A recent meta-analysis 2012a, 2013).
confirmed these reports and defined the overall rate
at 3.0% (95% CI 2.7–3.4%) (Gross and Du, 2013).
Prior hemorrhage
However, this overall rate has little meaning for individ-
ualized counseling as it is well known that it varies Although the original Finnish and University of Toronto
widely depending on various risk factors. AVM natural-history studies did not report an increased
Many early studies evaluated the influence of demo- risk of hemorrhage after AVM rupture (Ondra et al.,
graphic and AVM angioarchitectural features on the risk 1990; Stefani et al., 2002b), subsequent reports from
of hemorrhagic presentation, a relatively simplistic the same and other groups with longer follow-up
approach as it does not require follow-up. Langer et al. clearly identified previous rupture to be the most signif-
(1998) found hypertension, small size, and deep venous icant risk factor for subsequent hemorrhage during
drainage to be significantly associated with AVM hemor- follow-up (Forster et al., 1972; Crawford et al., 1986;
rhagic presentation. Other authors reported deep location Mast et al., 1997; Halim et al., 2004; Stapf et al.,
(Stefani et al., 2002b; Stapf et al., 2006), nonborderzone/ 2006; Yamada et al., 2007; Hernesniemi et al., 2008;
watershed location (Stapf et al., 2006), associated da Costa et al., 2009). The prospective University of
aneurysms (Stapf et al., 2006), infratentorial location Toronto AVM study demonstrated an annual hemor-
(Khaw et al., 2004), small number of draining veins rhage rate of 4.61% per year for the entire cohort of
(Stefani et al., 2002b), high feeding artery pressure 678 patients and an annual hemorrhage rate of 7.48%
(Duong et al., 1998), and venous ectasias (Stefani for patients with initial hemorrhagic presentation, result-
et al., 2002b) as independent factors significantly associ- ing in prior hemorrhage as a significant independent pre-
ated with hemorrhagic AVM presentation. However, this dictor of future hemorrhage (hazard ratio 2.15, p ¼ 0.07)
type of analysis is largely influenced by presentation (da Costa et al., 2009). A recent meta-analysis of AVM
bias, as quiescent AVMs that may otherwise only present natural-history studies confirmed these findings by dem-
with hemorrhage, such as small lesions, were perceived onstrating hemorrhage to be a statistically significant risk
as having a greater risk of hemorrhage (Itoyama et al., factor for subsequent bleeding, with a hazard ratio of 3.2
1989; Turjman et al., 1995b; Stapf et al., 2002; Stefani (95% CI 2.1–4.3) (Gross and Du, 2013). Unruptured
et al., 2002a; Majumdar et al., 2015). Results from this AVMs had an annual hemorrhage rate of 2.2% (95%
approach have little validity for patient counseling as fac- CI 1.7–2.7%) in this study while ruptured AVMs had
tors associated with hemorrhagic presentation are not an annual re-rupture rate of 4.5% (95% CI 3.7–5.5%)
equal to independent risk factors for future hemorrhage. (Gross and Du, 2013). The former is consistent with
Therefore, such decisions should be based on the per- the recent ARUBA trial results that demonstrated an
ceived prospective risk of hemorrhage after presentation. annual hemorrhage rate of 2.2% for untreated, unrup-
Thus, recent studies have evaluated annual rates of hem- tured AVMs (Mohr et al., 2014).
orrhage and hemorrhage risk factors on this annual rate. Although rebleed rates are higher for AVMs, it is
Most modern studies have not demonstrated a signif- important to underscore that the rebleed rate is dynamic,
icant impact of demographic features such as patient age with the highest hemorrhage rate during the first year of
and sex on the risk of hemorrhage (Kim et al., 2007; rupture, and declines thereafter. This may explain why
Hernesniemi et al., 2008; da Costa et al., 2009; Gross studies with longer follow-up periods generally report
and Du, 2013). Though studies based on hemorrhagic lower average annual rupture rates than studies with
presentation suggested a greater risk of hemorrhagic pre- shorter follow-up times. Re-rupture rates within the first
sentation for small AVMs (Itoyama et al., 1989; Spetzler year range from 6% to 15.8% across multiple studies
et al., 1992; Kader et al., 1994; Turjman et al., 1995b; (Mast et al., 1997; Yamada et al., 2007; Hernesniemi
Stapf et al., 2002; Stefani et al., 2002b), as discussed, this et al., 2008; da Costa et al., 2009; Gross and Du,
was likely an artifact of presentation bias as studies based 2012b), a fact that has important implications for both
on prospective annual hemorrhage rates have not demon- the timing of surgical treatment and the choice of radio-
strated small size to be a significant risk factor for surgery for ruptured AVMs. In one study comparing the
THE NATURAL HISTORY OF CEREBRAL ARTERIOVENOUS MALFORMATIONS 19
risk of hemorrhage in the clinical course of AVMs in 281 1.01–5.67) and 4.1 (95% CI 1.2–14.9), respectively. In
patients with and without initial hemorrhage, re-rupture a recent meta-analysis, the overall hazard ratio for hem-
occurred in 13% versus 2% in the nonhemorrhage group, orrhage from AVMs with deep venous drainage was 3.25
leading to an annual risk of hemorrhage of 17.8% and (95% CI 1.01–5.67) (Stapf et al., 2006).
2.2%, respectively (Mast et al., 1997). However, in a sub-
sequent report of the same cohort with 622 consecutive Deep location
patients, with a mean follow-up of 829 days during
Approximately one-third of AVMs in natural-history
which 39 patients developed AVM hemorrhage, the
studies are deep (Yamada et al., 2007; Hernesniemi
annual rupture rate for unruptured AVMs was 1.3% ver-
et al., 2008; Gross and Du, 2013). It is important to
sus 5.9% for those after hemorrhagic AVM presentation
underscore that deep location has been implicated as a
(Stapf et al., 2006). In another study, two-thirds of AVMs
risk factor for hemorrhage independent of the tendency
that rebled within the first year were associated with
of deep AVMs to possess deep venous drainage. Deep
aneurysms, suggesting that associated aneurysms are
location may be an independent risk factor due to a
probably a contributing factor to this high re-rupture rate.
greater tendency for deep AVMs to have supply from
The permanent morbidity from a rebleed was 45%
fragile perforators or potentially from a lack of
(Gross and Du, 2012b). Interestingly, after reviewing
“tamponade effect” from adjacent parenchyma for those
the literature, this report illustrated that the rebleed rate
presenting to a ventricular surface (Hernesniemi et al.,
within the first year was generally double the overall
2008; Gross and Du, 2013).
re-hemorrhage rate provided in each natural-history
Overall, annual hemorrhage rates for deep AVMs
study (Gross and Du, 2012b).
range from 4.1% to 8.6% in the literature (Yamada
et al., 2007; Hernesniemi et al., 2008). Deep location
Deep venous drainage was found to correlate with clinical presentation of hem-
orrhage in some retrospective reports (Turjman et al.,
Deep drainage has been shown to be correlated with
1995b; Mansmann et al., 2000), but not in others
hemorrhagic presentation by multiple research groups
(Duong et al., 1998; Langer et al., 1998). Across two
(Marks et al., 1990; Kader et al., 1994; Turjman et al.,
natural-history studies, the annual rupture rate for deep
1995b). In one retrospective study, deep drainage was
AVMs ranged from 3.1% to 4.5% for unruptured lesions
significantly associated with hemorrhagic presentation
and increased to 11.4–14.8% for ruptured AVMs (Stapf
(odds ratio 5.77; p ¼ 0.009) and 25% of patients had
et al., 2006; Yamada et al., 2007). Deep location was an
additional intraventricular hemorrhage (Langer et al.,
independent risk factor for subsequent AVM hemorrhage
1998). In addition to these small retrospective studies
in several studies, with hazard ratios ranging from 3.07
with significant limitations, most natural-history studies
(95% CI 1.28–7.40) (Yamada et al., 2007) to 3.25
consistently demonstrate that approximately half of
(95% CI 1.30–8.16) (Stapf et al., 2006). In a recent
AVMs possess deep venous drainage (Stapf et al.,
meta-analysis, the overall meta-analytic hazard ratio of
2006; Yamada et al., 2007; Hernesniemi et al., 2008;
deep AVM location was 2.4 (95% CI 1.4–3.4) (Gross
da Costa et al., 2009). It has been postulated that the
and Du, 2013). However, similar to small AVMs, deep
increased pressure of the deep venous system, and sub-
lesions may be less likely to cause symptoms such as sei-
sequent increase in the pressure gradient in the AVM
zures and neurologic deficits due to their deep location,
nidus, is the etiologic cause of greater reported hemor-
and are therefore prone to selection bias.
rhage rates for AVMs with this feature (Hernesniemi
et al., 2008; da Costa et al., 2009). Overall hemorrhage
Associated aneurysms
rates for AVMs with deep venous drainage range from
3.4% to 5.4% in the literature (Yamada et al., 2007; The pathophysiology, incidence, and classification of
Hernesniemi et al., 2008; da Costa et al., 2009). Across intracranial aneurysms in relation to cerebral AVMs have
two natural-history studies, annual hemorrhage rates for been the topic of controversy and interest, and the subject
AVMs with deep venous drainage ranged from 2.4% to of ongoing discussions in the medical literature. Aneu-
2.6% for unruptured lesions and increased to rysms may be located in the nidus, on feeding arteries,
7.2–11.4% for ruptured AVMs (Stapf et al., 2006; or they may occur on peripheral arteries seemingly unre-
Yamada et al., 2007). Deep venous drainage was reported lated to the AVM lesion. They have been reported to
as an independent predictor of subsequent hemorrhage in occur in association with AVMs in approximately 10%
a prospective study of 622 patients with untreated AVMs of patients (Al-Shahi and Warlow, 2001), and across
from the Columbia AVM database (Stapf et al., 2006) several natural-history studies, approximately 20% of
and in an earlier reported analysis of the same cohort cerebral AVMs were associated with aneurysms
(Mast et al., 1997), with hazard ratios of 3.25 (95% CI (Brown et al., 1988; da Costa et al., 2009; Gross and
20 A. CAN ET AL.
Du, 2013). In these studies, approximately 50% of associ- system, although the exact underlying mechanisms in rela-
ated aneurysms were on feeding arteries while 25% were tion to AVMs have yet to be clarified (Stieg et al., 2007).
intranidal and 25% were in a remote location (Gross and Early studies report AVMs to be responsible for intra-
Du, 2013). cranial hemorrhage in 21–48% of cases of spontaneous
Multiple studies found an association between clini- intracranial hemorrhage during pregnancy (Fliegner
cal presentation of AVMs and the presence of related et al., 1969; Amias, 1970; Robinson et al., 1972,
intracranial aneurysms (Marks et al., 1990; Turjman 1974). In a study of 154 pregnant women with verified
et al., 1995b). In a cross-sectional study of 463 prospec- intracranial hemorrhage, 23% of the hemorrhages were
tively enrolled patients, Stapf et al. (2002) demonstrated secondary to ruptured AVMs (Dias and Sekhar, 1990).
in a multivariate model an independent effect of feeding The overall maternal and fetal mortality from ruptured
artery aneurysms on hemorrhagic AVM presentation AVMs in patients that were observed were 32% and
(OR 2.11, 95% CI 1.18–3.78). 23%, respectively. More than half of patients with rup-
Although some of these studies demonstrated a clear tured AVMs were moribund or comatose at time of
correlation between hemorrhagic presentation and asso- presentation.
ciated aneurysms, this important angiographic feature is In another report comprised of 24 women with AVMs,
underreported in the prospective natural-history litera- hemorrhage was associated with a 49% fetal complica-
ture. With intuitive preliminary evidence that AVMs tion rate and a 26% mortality rate, in contrast to a 33%
associated with aneurysms have a greater risk of rupture fetal complication rate associated with aneurysmal
(Turjman et al., 1995b; Stapf et al., 2002), and with addi- subarachnoid hemorrhage (Robinson et al., 1974).
tional evidence suggesting greater morbidity from rup- A recent retrospective study of 54 women with an angio-
ture of AVMs associated with aneurysms (Gross et al., graphic diagnosis of an AVM quantified a hemorrhage
2013), it is likely that these lesions are selected out early rate of 8.1% per pregnancy or an annual hemorrhage rate
from natural-history studies and are therefore subject to of 10.8% versus 1.1% in nonpregnant women. The sta-
selection bias, unless the aneurysms seem angiographic- tistically significant hazard ratio for hemorrhage during
ally benign and/or small. Nevertheless, two studies pregnancy was 7.91 in this study (Gross and Du,
reported relatively consistent annual hemorrhage rates 2012a). These findings are consistent with another recent
ranging from 6.9% to 8.3% for AVMs associated with study that reported an annual hemorrhage rate of 11.1%
aneurysms. In the prospective Toronto AVM study, mul- for women who become pregnant during the 3-year
tivariate analysis did not reveal associated aneurysms as latency interval between stereotactic radiosurgery and
independent predictors for subsequent hemorrhage. documented AVM obliteration (Tonetti et al., 2014).
However, in univariate analysis for specific aneurysm Among nonpregnant women, the annual hemorrhage rate
subtypes, both feeding artery aneurysms (HR 1.7, was 2.5% during the latency interval (Tonetti et al.,
p ¼ 0.03) and intranidal aneurysms (HR 2.1, p ¼ 0.03) 2014). Timing of hemorrhage during pregnancy varies
significantly increased the risk of hemorrhage, in con- widely among studies, with reported cases during the
trast with remote aneurysms (da Costa et al., 2009). first (Horton et al., 1990; Tonetti et al., 2014), second
A subsequent recent meta-analysis reinforced associated (Horton et al., 1990), and third trimester (Horton et al.,
aneurysms as a statistically significant risk factor for 1990; Gross and Du, 2012a), and puerperium (Horton
hemorrhage (Gross and Du, 2013). et al., 1990).
Due to preliminary studies that strongly suggest ele-
vated hemorrhage risks during pregnancy and the obvious
ethical dilemmas related to prospective natural-history
Pregnancy
studies in pregnant women harboring AVMs, no such
Although data regarding risk of AVM hemorrhage during studies have been performed to date and the annual hem-
pregnancy are sparse and inconclusive, pregnancy has orrhage rates for these patients are consequently lacking.
long been implicated as a significant risk factor for
AVM hemorrhage and the presence of an AVM signifi-
Age and sex
cantly complicates the management of pregnant women
(Crawford et al., 1986; Lanzino et al., 1994; Skidmore Although age was shown to be a significant predictor of
et al., 2001). In the early natural-history study by AVM hemorrhage in some studies (Karlsson et al., 1997;
Crawford et al. (1986), 25% of women aged 20–29 years Stapf et al., 2006), this finding was not confirmed by
old presenting with AVM hemorrhage were pregnant. others (Yamada et al., 2007; da Costa et al., 2009;
Pregnancy may be a risk factor for AVM hemorrhage Gross and Du, 2013). In one prospective study, patients
due to several major physiologic changes that occur within who experienced an AVM rupture were younger at
the cardiovascular, hormonal, and hemostatic-thrombotic admission (on average 5 years), but these results were
THE NATURAL HISTORY OF CEREBRAL ARTERIOVENOUS MALFORMATIONS 21
not significant in univariate and multivariate Cox models In addition, women desiring to bear children with inci-
(Hernesniemi et al., 2008). However, a recent meta- dentally discovered AVMs should also be considered
analysis of combined data from four large cohorts found for curative treatment prior to planned conception.
increasing age to be significantly associated with an Although the details of AVM treatment go beyond the
approximately 30% increase in risk for every 10-year scope of this chapter, the choice of management for
increase in age (HR 1.34 per decade, 95% CI patients harboring these lesions should consider
1.17–1.53). However, the lack of consensus in the litera- treatment-related risks as related to patient age, location
ture, combined with a higher risk of poor outcome after sur- of the lesion, size of the AVM, and vascular topography,
gery in elderly patients (Ding et al., 2015), highlights the as well as the natural history of the individual patient.
need for more well-designed larger prospective studies. Therapeutic options for AVMs include operative obliter-
Similar to data regarding age, current studies do not ation or resection, endovascular embolization, and radio-
provide clear associations between sex and risk of surgery. Given the high risk of rebleeding within 1 year,
AVM-related hemorrhage. One clear exception was a ruptured AVMs are generally considered for surgical
natural-history study of 305 consecutive patients with treatment as it provides the most expedient and definitive
AVMs that showed a significantly increased hemorrhage means for angiographic obliteration. Similarly, AVMs
risk among female patients (HR 2.93 95% CI 1.20–7.16). with deep venous drainage or associated aneurysms
However, the smaller portion of female patients and should also be considered for early surgery, if feasible
skewed distributions in this study compared with other at low morbidity. Deep AVMs (e.g., deep capsular, brain-
studies may have influenced the results (Fults and stem, or basal ganglia) may pose a challenge for surgery
Kelly, 1984; Mast et al., 1997; Stapf et al., 2003; given their tendency to involve eloquent tissue; radiosur-
Halim et al., 2004). Karlsson et al. (1997) reported that gery may thus serve as the optimal therapeutic option.
the risk of hemorrhage was higher in women during their Although one early prospective study of 73 consecutive
fertile years as compared to males in the same age group. patients with grades IV and V AVMs suggested that
Although Mast et al. (1997) initially reported male sex to high-grade AVMs may in fact have a more benign
be significantly associated with subsequent hemorrhage natural history than their low-grade counterparts (Han
(HR 9.2, 95% CI 2.1–41.3), these findings were not con- et al., 2003), more recent studies have reported conflicting
firmed in a more recent report from the same group (Stapf results. One retrospective study of 61 consecutive patients
et al., 2006). In line with the discussed differences in pre- with high-grade AVMs reported an annual hemorrhage rate
viously published reports, Kim et al. (2007) reported bor- of 10.4% (95% CI 2.2–15.4) (Jayaraman et al., 2007),
derline associations between female sex and hemorrhage while another reported an overall annual hemorrhage rate
risk (HR 1.49, 95% CI 0.96–2.30) and a recent meta- of 3.3% that increased to 6.0% for ruptured, high-grade
analysis showed female sex not to be associated with lesions (Laakso et al., 2008). Though they present a signi-
hemorrhage (Gross and Du, 2013). ficant potential therapeutic challenge, these recent studies
encourage creative, multimodality approaches when con-
sidering treatment of high-grade AVMs.
DISCUSSION
As an increasing number of unruptured cerebral
The wide variation in the clinical course of patients with AVMs are identified due to the increasing use of
AVMs and the heterogeneity of different patient popula- high-resolution imaging, this will lead to a shift in the
tions hamper generalizations and make estimations of the AVM population, creating a diagnostic and treatment
natural history of AVMs the subject of controversy. The challenge for clinicians. For a better understanding of
natural history of AVMs must incorporate a perceived, the natural history of cerebral AVMs, future studies with
prospective risk of epilepsy, hemorrhage, and accompa- longer follow-up times are needed.
nying daily psychologic burden. The risk of epilepsy is
considerably less explored than the risk of hemorrhage
CONCLUSION
in the literature, and the importance of seizure control
is often undervalued in the surgical treatment of AVMs; The natural history of cerebral AVMs encompasses over-
however, epilepsy may result in significant morbidity all annual rates of 2–4% for the risk of hemorrhage, and
and decreased quality of life. Nevertheless, the decision an annual rate of 1% for the development of de novo sei-
to treat an AVM is often predicated on the presumed, pro- zures. Significant risk factors for hemorrhage include
spective risk of hemorrhage from the lesion. This chapter prior rupture, deep location, deep venous drainage, asso-
gave an overview of risk factors for hemorrhage that ciated aneurysms, and pregnancy. Patient age, sex, and
should lower the threshold for treatment – specifically small AVM size are not currently considered significant
for hemorrhagic AVMs, deep AVMs, those with deep risk factors for AVM hemorrhage. In addition to hemor-
venous drainage, and those with associated aneurysms. rhage risk and seizure risk, the natural history of an AVM
22 A. CAN ET AL.
also encompasses the daily psychologic burden that a Dias MS, Sekhar LN (1990). Intracranial hemorrhage from
patient must endure knowing he or she possesses an aneurysms and arteriovenous malformations during preg-
untreated AVM. All of the above factors need to be taken nancy and the puerperium. Neurosurgery 27: 855–865;
into consideration when determining the optimal treat- discussion 865–856.
Ding D, Xu Z, Yen CP et al. (2015). Radiosurgery for cerebral
ment of an AVM.
arteriovenous malformations in elderly patients: effect of
REFERENCES advanced age on outcomes after intervention. World
Neurosurg 84: 795–804.
Al-Rodhan NR, Al-Mefty O, Rifai A et al. (1986). Persistence Duong DH, Young WL, Vang MC et al. (1998). Feeding artery
of primitive cerebral vasculature in a newborn. A case pressure and venous drainage pattern are primary determi-
report of whole brain AVM. Clin Neurol Neurosurg 88: nants of hemorrhage from cerebral arteriovenous malfor-
283–287. mations. Stroke 29: 1167–1176.
Al-Shahi R, Warlow C (2001). A systematic review of the fre- Englot DJ, Young WL, Han SJ et al. (2012). Seizure
quency and prognosis of arteriovenous malformations of predictors and control after microsurgical resection of
the brain in adults. Brain 124: 1900–1926. supratentorial arteriovenous malformations in 440 patients.
Al-Shahi R, Fang JS, Lewis SC et al. (2002). Prevalence Neurosurgery 71: 572–580; discussion 580.
of adults with brain arteriovenous malformations: a com- Fliegner JR, Hooper RS, Kloss M (1969). Subarachnoid hae-
munity based study in Scotland using capture-recapture morrhage and pregnancy. J Obstet Gynaecol Br
analysis. J Neurol Neurosurg Psychiatry 73: 547–551. Commonw 76: 912–917.
Al-Shahi R, Bhattacharya JJ, Currie DG et al. (2003). Forster DM, Steiner L, Hakanson S (1972). Arteriovenous
Prospective, population-based detection of intracranial malformations of the brain. A long-term clinical study.
vascular malformations in adults: the Scottish Intracranial J Neurosurg 37: 562–570.
Vascular Malformation Study (SIVMS). Stroke 34: Fults D, Kelly Jr DL (1984). Natural history of arteriovenous
1163–1169. malformations of the brain: a clinical study. Neurosurgery
Amias AG (1970). Cerebral vascular disease in pregnancy. I. 15: 658–662.
Haemorrhage. J Obstet Gynaecol Br Commonw 77: 100–120. Galletti F, Costa C, Cupini LM et al. (2014). Brain arteriove-
Brown Jr RD, Wiebers DO, Forbes G et al. (1988). The natural nous malformations and seizures: an Italian study. J Neurol
history of unruptured intracranial arteriovenous malforma- Neurosurg Psychiatry 85: 284–288.
tions. J Neurosurg 68: 352–357. Gonzalez LF, Bristol RE, Porter RW et al. (2005). De novo
Brown Jr RD, Wiebers DO, Torner JC et al. (1996). Incidence presentation of an arteriovenous malformation. Case report
and prevalence of intracranial vascular malformations in and review of the literature. J Neurosurg 102: 726–729.
Olmsted County, Minnesota, 1965 to 1992. Neurology Gross BA, Du R (2012a). Hemorrhage from arteriovenous
46: 949–952. malformations during pregnancy. Neurosurgery 71:
Brown Jr RD, Flemming KD, Meyer FB et al. (2005). Natural 349–355; discussion 355–356.
history, evaluation, and management of intracranial vascu- Gross BA, Du R (2012b). Rate of re-bleeding of arteriovenous
lar malformations. Mayo Clin Proc 80: 269–281. malformations in the first year after rupture. J Clin
Buis DR, van den Berg R, Lycklama G et al. (2004). Spontaneous Neurosci 19: 1087–1088.
regression of brain arteriovenous malformations – a clinical Gross BA, Du R (2013). Natural history of cerebral arteriove-
study and a systematic review of the literature. J Neurol 251: nous malformations: a meta-analysis. J Neurosurg 118:
1375–1382. 437–443.
Chimowitz MI, Little JR, Awad IA et al. (1990). Intracranial Gross BA, Lai PM, Du R (2013). Hydrocephalus after arterio-
hypertension associated with unruptured cerebral arterio- venous malformation rupture. Neurosurg Focus 34: E11.
venous malformations. Ann Neurol 27: 474–479. Gross BA, Storey A, Orbach DB et al. (2015). Microsurgical
Courville CB (1967). Intracranial tumors. Notes upon a series treatment of arteriovenous malformations in pediatric
of three thousand verified cases with some current observa- patients: the Boston Children’s Hospital experience.
tions pertaining to their mortality. Bull Los Angeles Neurol J Neurosurg Pediatr 15: 71–77.
Soc 32 (Suppl 2): 1–80. Halim AX, Johnston SC, Singh V et al. (2004). Longitudinal
Crawford PM, West CR, Chadwick DW et al. (1986). risk of intracranial hemorrhage in patients with arteriove-
Arteriovenous malformations of the brain: natural history nous malformation of the brain within a defined population.
in unoperated patients. J Neurol Neurosurg Psychiatry Stroke 35: 1697–1702.
49: 1–10. Han PP, Ponce FA, Spetzler RF (2003). Intention-to-treat anal-
da Costa L, Wallace MC, Ter Brugge KG et al. (2009). The ysis of Spetzler-Martin grades IV and V arteriovenous mal-
natural history and predictive features of hemorrhage from formations: natural history and treatment paradigm.
brain arteriovenous malformations. Stroke 40: 100–105. J Neurosurg 98: 3–7.
Darsaut TE, Guzman R, Marcellus ML et al. (2011). Hatori K, Urabe T, Kanazawa A et al. (1991). A case of
Management of pediatric intracranial arteriovenous mal- brainstem vascular malformation with isolated trochlear
formations: experience with multimodality therapy. nerve palsy as the initial symptom. No To Shinkei 43:
Neurosurgery 69: 540–556; discussion 556. 965–968.
THE NATURAL HISTORY OF CEREBRAL ARTERIOVENOUS MALFORMATIONS 23
Hernesniemi JA, Dashti R, Juvela S et al. (2008). Natural his- Lanzino G, Jensen ME, Cappelletto B et al. (1994). Arteriovenous
tory of brain arteriovenous malformations: a long-term malformations that rupture during pregnancy: a manage-
follow-up study of risk of hemorrhage in 238 patients. ment dilemma. Acta Neurochir (Wien) 126: 102–106.
Neurosurgery 63: 823–829; discussion 829–831. Lazar RM, Connaire K, Marshall RS et al. (1999).
Horton JC, Chambers WA, Lyons SL et al. (1990). Pregnancy Developmental deficits in adult patients with arteriovenous
and the risk of hemorrhage from cerebral arteriovenous malformations. Arch Neurol 56: 103–106.
malformations. Neurosurgery 27: 867–871; discussion Lee SK, Vilela P, Willinsky R et al. (2002). Spontaneous
871–872. regression of cerebral arteriovenous malformations: clini-
Itoyama Y, Uemura S, Ushio Y et al. (1989). Natural course of cal and angiographic analysis with review of the literature.
unoperated intracranial arteriovenous malformations: Neuroradiology 44: 11–16.
study of 50 cases. J Neurosurg 71: 805–809. Lobo-Antunes J, Yahr MD, Hilal SK (1974). Extrapyramidal
Jayaraman MV, Marcellus ML, Do HM et al. (2007). dysfunction with cerebral arteriovenous malformations.
Hemorrhage rate in patients with Spetzler-Martin grades J Neurol Neurosurg Psychiatry 37: 259–268.
IV and V arteriovenous malformations: is treatment justi- Majumdar M, Tan LA, Chen M (2015). Critical assessment of
fied? Stroke 38: 325–329. the morbidity associated with ruptured cerebral arteriove-
Johnson MC, Salmon JH (1968). Arteriovenous malformation nous malformations. J Neurointerv Surg 8 (2): 163–167.
presenting as trigeminal neuralgia. Case report. J Mansmann U, Meisel J, Brock M et al. (2000). Factors associ-
Neurosurg 29: 287–289. ated with intracranial hemorrhage in cases of cerebral arte-
Josephson CB, Bhattacharya JJ, Counsell CE et al. (2012). riovenous malformation. Neurosurgery 46: 272–279;
Seizure risk with AVM treatment or conservative manage- discussion 279–281.
ment: prospective, population-based study. Neurology 79: Marks MP, Lane B, Steinberg GK et al. (1990). Hemorrhage in
500–507. intracerebral arteriovenous malformations: angiographic
Kader A, Young WL, Pile-Spellman J et al. (1994). The influ- determinants. Radiology 176: 807–813.
ence of hemodynamic and anatomic factors on hemorrhage Mast H, Mohr JP, Osipov A et al. (1995). ‘Steal’ is an unestab-
from cerebral arteriovenous malformations. Neurosurgery lished mechanism for the clinical presentation of cerebral
34: 801–807; discussion 807–808. arteriovenous malformations. Stroke 26: 1215–1220.
Kader A, Goodrich JT, Sonstein WJ et al. (1996). Recurrent Mast H, Young WL, Koennecke HC et al. (1997). Risk of spon-
cerebral arteriovenous malformations after negative post- taneous haemorrhage after diagnosis of cerebral arteriove-
operative angiograms. J Neurosurg 85: 14–18. nous malformation. Lancet 350: 1065–1068.
Karlsson B, Lindquist C, Johansson A et al. (1997). Annual McCormick WF (1966). The pathology of vascular (“arterio-
risk for the first hemorrhage from untreated cerebral arte- venous”) malformations. J Neurosurg 24: 807–816.
riovenous malformations. Minim Invasive Neurosurg 40: Moftakhar P, Hauptman JS, Malkasian D et al. (2009).
40–46. Cerebral arteriovenous malformations. Part 2: physiology.
Khaw AV, Mohr JP, Sciacca RR et al. (2004). Association Neurosurg Focus 26: E11.
of infratentorial brain arteriovenous malformations Mohr JP, Parides MK, Stapf C et al. (2014). Medical manage-
with hemorrhage at initial presentation. Stroke 35: ment with or without interventional therapy for unruptured
660–663. brain arteriovenous malformations (ARUBA): a multicen-
Kilbourn KJ, Spiegel G, Killory BD et al. (2014). Case report tre, non-blinded, randomised trial. Lancet 383: 614–621.
of a de novo brainstem arteriovenous malformation in an Morello G, Borghi GP (1973). Cerebral angiomas. A report of
18-year-old male and review of the literature. Neurosurg 154 personal cases and a comparison between the results of
Rev 37: 685–691. surgical excision and conservative management. Acta
Kim Y, Tanaka A, Kimura M et al. (1991). Arteriovenous mal- Neurochir (Wien) 28: 135–155.
formation in the cerebellopontine angle presenting as hemi- Olivecrona H, Riives J (1948). Arteriovenous aneurysms of the
facial spasm – case report. Neurol Med Chir (Tokyo) 31: brain, their diagnosis and treatment. Arch Neurol Psychiatry
109–112. 59: 567–602.
Kim H, Sidney S, McCulloch CE et al. (2007). Racial/Ethnic Ondra SL, Troupp H, George ED et al. (1990). The natural his-
differences in longitudinal risk of intracranial hemorrhage tory of symptomatic arteriovenous malformations of the
in brain arteriovenous malformation patients. Stroke 38: brain: a 24-year follow-up assessment. J Neurosurg 73:
2430–2437. 387–391.
Laakso A, Dashti R, Seppanen J et al. (2008). Long-term Pasqualin A, Vivenza C, Rosta L et al. (1985). Spontaneous
excess mortality in 623 patients with brain arteriovenous disappearance of intracranial arterio-venous malforma-
malformations. Neurosurgery 63: 244–253; discussion tions. Acta Neurochir (Wien) 76: 50–57.
253–255. Perret G, Nishioka H (1966). Report on the cooperative study
Langer DJ, Lasner TM, Hurst RW et al. (1998). Hypertension, of intracranial aneurysms and subarachnoid hemorrhage.
small size, and deep venous drainage are associated with Section VI. Arteriovenous malformations. An analysis
risk of hemorrhagic presentation of cerebral arteriovenous of 545 cases of cranio-cerebral arteriovenous malforma-
malformations. Neurosurgery 42: 481–486; discussion tions and fistulae reported to the cooperative study.
487–489. J Neurosurg 25: 467–490.
24 A. CAN ET AL.
Pollock BE, Flickinger JC, Lunsford LD et al. (1996). Factors Stefani MA, Porter PJ, terBrugge KG et al. (2002b). Large and
that predict the bleeding risk of cerebral arteriovenous mal- deep brain arteriovenous malformations are associated
formations. Stroke 27: 1–6. with risk of future hemorrhage. Stroke 33: 1220–1224.
Robinson JL, Hall CJ, Sedzimir CB (1972). Subarachnoid Steinheil SO (1895). Ueber einen Fall von Varix aneurysma-
hemorrhage in pregnancy. J Neurosurg 36: 27–33. ticus im Bereich der Gehirngefaesse, F. Fromme,
Robinson JL, Hall CS, Sedzimir CB (1974). Arteriovenous Wurzburg, pp. 1–56.
malformations, aneurysms, and pregnancy. J Neurosurg Stieg PE, Batjer HH, Samson DS (2007). Intracranial arterio-
41: 63–70. venous malformations, Informa Healthcare, New York.
Ruiz-Sandoval JL, Cantu C, Barinagarrementeria F (1999). Thorpe ML, Cordato DJ, Morgan MK et al. (2000).
Intracerebral hemorrhage in young people: analysis of risk Postoperative seizure outcome in a series of 114 patients
factors, location, causes, and prognosis. Stroke 30: with supratentorial arteriovenous malformations. J Clin
537–541. Neurosci 7: 107–111.
Sabra F (1959). Observations on one hundred cases of cerebral Tonetti D, Kano H, Bowden G et al. (2014). Hemorrhage
angioma. J Am Med Assoc 170: 1522–1524. during pregnancy in the latency interval after stereotactic
Skidmore FM, Williams LS, Fradkin KD et al. (2001). radiosurgery for arteriovenous malformations. J Neurosurg
Presentation, etiology, and outcome of stroke in pregnancy 121 (Suppl): 226–231.
and puerperium. J Stroke Cerebrovasc Dis 10: 1–10. Turjman F, Massoud TF, Sayre JW et al. (1995a). Epilepsy
Spetzler RF, Hargraves RW, McCormick PW et al. (1992). associated with cerebral arteriovenous malformations: a
Relationship of perfusion pressure and size to risk of hem- multivariate analysis of angioarchitectural characteristics.
orrhage from arteriovenous malformations. J Neurosurg AJNR Am J Neuroradiol 16: 345–350.
76: 918–923. Turjman F, Massoud TF, Vinuela F et al. (1995b). Correlation
Spetzler RF, Kalani Y, Nakaji P (2015). Neurovascular sur- of the angioarchitectural features of cerebral arteriovenous
gery, Thieme, New York. malformations with clinical presentation of hemorrhage.
Stapf C, Mohr JP, Pile-Spellman J et al. (2001). Epidemiology Neurosurgery 37: 856–860; discussion 860–862.
and natural history of arteriovenous malformations. van Beijnum J, van der Worp HB, Schippers HM et al. (2007).
Neurosurg Focus 11: e1. Familial occurrence of brain arteriovenous malformations:
Stapf C, Mohr JP, Pile-Spellman J et al. (2002). Concurrent a systematic review. J Neurol Neurosurg Psychiatry 78:
arterial aneurysms in brain arteriovenous malformations 1213–1217.
with haemorrhagic presentation. J Neurol Neurosurg van Beijnum J, van der Worp HB, Algra A et al. (2014).
Psychiatry 73: 294–298. Prevalence of brain arteriovenous malformations in
Stapf C, Mast H, Sciacca RR et al. (2003). The New York first-degree relatives of patients with a brain arteriovenous
Islands AVM Study: design, study progress, and initial malformation. Stroke 45: 3231–3235.
results. Stroke 34: e29–e33. Waltimo O (1973). The change in size of intracranial arterio-
Stapf C, Mast H, Sciacca RR et al. (2006). Predictors of hem- venous malformations. J Neurol Sci 19: 21–27.
orrhage in patients with untreated brain arteriovenous mal- Waltimo O, Hokkanen E, Pirskanen R (1975). Intracranial
formation. Neurology 66: 1350–1355. arteriovenous malformations and headache. Headache
Stefani MA, Porter PJ, terBrugge KG et al. (2002a). 15: 133–135.
Angioarchitectural factors present in brain arteriovenous Yamada S, Takagi Y, Nozaki K et al. (2007). Risk factors for
malformations associated with hemorrhagic presentation. subsequent hemorrhage in patients with cerebral arteriove-
Stroke 33: 920–924. nous malformations. J Neurosurg 107: 965–972.
Another random document with
no related content on Scribd:
for more than a thousand years, Arabia lay contained within herself,
dammed perhaps by the Persian, Macedonian, Parthian, and Roman
empires, until in the seventh century after Christ Mohammedanism
led forth her peoples. A much earlier movement, at an unknown
time, had brought the forefathers of the Abyssinians across the
mouth of the Red Sea into Africa, and the Hyksos who overthrew the
Middle Kingdom of ancient Egypt may have been Semites.
The Indo-Europeans entered southwest Asia later and permeated
it more locally than the Semites. Soon after 2000 the Kassites or
Kossæans intruded into Babylonia; they seem to have been Indo-
Europeans, perhaps Iranians. Around 1500 the Mitanni were a
power on the upper Euphrates between the Assyrians and the
Hittites of Asia Minor. Their personal and god names as preserved in
Assyrian Cuneiform inscriptions show them to have been an Iranian
people. The latter are not recognizably referred to in their permanent
home on the Iranian plateau until about 1000, but may well have
settled there a thousand years earlier. Their close relatives, the Indic
branch, are believed to have begun their entry of India about 2000-
1500 B.C. or soon after.
250. Europe
With Greece we have entered the realm of what is conventionally
regarded as history. For the rest of Europe, prehistoric archæology
and its record of illiterate peoples abut so closely on history in the
ordinary sense, that a tracing of the transition takes one promptly
into documentary study. There is much in this early historical field
that is of anthropological interest, and just back of it lies more that is
specifically so: where the round headed peoples came from who
began to appear in Europe during the Neolithic; whether peoples like
Ligurians, Sicilians, Scythians, were Indo-Europeans or not, and of
what branch; where the blond Nordic type took shape and whether it
originally spoke dialects of the Germanic group; who built
Stonehenge and the other megalithic monuments of western Europe;
where the first home of the Indo-Europeans lay. But such problems
are intricate, and usually answerable, if at all, from stray indications
scattered among masses of literary and historical data controllable
only by the specialist, whose primary interests tend in other
directions. Where these documented indications fail, the problems
become speculative. We have no clear record of any indubitable
Indo-European people, in or out of Europe, before the second
millenium B.C. When they appear in history, they are already
differentiated into their familiar main divisions. The Bronze Age
Scandinavians seem likely to have been Indo-Europeans and
perhaps of the Germanic branch; for the Neolithic an identification
would be mere guessing. The LaTène Iron culture is
characteristically Keltic, that of the Hallstadt period and area less
certainly Illyrian and Keltic. And after all, such considerations
concern speech, or race, which can be associated with any culture.
Our present concern being primarily with the latter, it will be more
profitable to pass on from these questions and turn to regions
remote from those in which Occidental civilization assumed its
modern form.
251. China
China, far from Europe and known to the outside world only
recently, possesses a civilization so different from ours in a multitude
of aspects, that thought of connection between the cultures seems at
first unreasonable. One thinks of rice, pagodas, bound feet, queues,
silk, tea, ancestor worship, a strange, chopped, singing speech, and
writing in still stranger characters. Yet the Chinese have long had a
civilization identical in many of its constituents with our own: civil
government, rimed poetry, painting, trousers, wheat and barley, our
common domestic animals, bronze and iron, for instance. Since
most of these culture elements are wanting in Africa and Oceania, as
well as in native America, there is no inherent reason why they
should be expectable in China. Their repetition in China and in the
West as the result of independent causes would be remarkable.
Evidently many if not all of this group of common traits represent
absorptions into the civilization of China, or diffusions out of it into
the West, much as the larger part of early European civilization was
imported out of the nearer Orient.
In the broader perspective of culture history, then, China no longer
stands aloof. The roots of her civilization are largely the same as
those of our own. In this light, understanding of Chinese civilization
involves two steps. The first is the tracing of the elements derived
from the west or imparted to it. The second is the recognition of how
these were remodeled and combined with elements of local growth
and thereby given their peculiarly Chinese cast and setting.
Authentic historical records of China go back only three thousand
years, and her archæology is little known. Beyond the beginning of
the Chou dynasty, about 1100 B.C., or more exactly, beyond a point
when this dynasty was about three centuries old, in 827 B.C., the
Chinese possess only legendary history, in which slight strands of
fact are interwoven with fabricated or fabulous constituents. Then,
too, the Chinese have long been genuinely more advanced than
their neighbors, than all of their world, in fact; with the result that they
could hardly escape the conviction of their own superiority and self-
sufficiency, and the belief that they had devised almost everything in
their own culture. This presumption led to the conscientious
manufacture by native historians of dates for inventions which were
really made outside of China.
Beginning, then, in the ninth century B.C., we find the Chinese a
settled and populous people in the lower valley of the great Yellow
river, in what is now the northeastern corner of the “Eighteen
Provinces” or China proper, from about Si-an-fu to Peking. They may
have come from middle Asia or still farther west by a national
migration, as has sometimes been conjectured: there is nothing to
show that they did, and a great deal to suggest that they had lived in
or near their seats of that period long before. It is difficult to imagine
that the Chinese could have moved out of central Asia without
leaving a part of their number behind, or without leaving conspicuous
traces of their culture among their former neighbors. Of this there is
no evidence. For one of the most advanced peoples of its time to
remove itself and its civilization complete and unimpaired would be
without parallel in history, and indeed is inconceivable as soon as
one turns from the vague idea to face specific details of the process.
Nevertheless the Chinese as we first know them had the principal
grains and tamed animals, the metals and plow and wheel of
contemporary Eastern Asia and Europe. While it is scarcely
thinkable that this great complex of culture traits should not have
been due to connections with the west, there is every probability that
these connections were of the sort that have been traced so
frequently in foregoing pages: diffusions unaccompanied by
populational drifts, or at least in the main independent of them. When
it is recalled that western Turkistan held cereal-growing and animal-
raising inhabitants far back in the Neolithic, and that the Bronze
period is definitely represented in Siberia,[36] such transmission will
not seem far fetched. It is also well to remember that all through the
historic period central Asia contained farming populations, cities,
traders, and skilled artisans, some measure of which evidences of
higher culture it is only necessary to project a few thousand years
backward to complete the link in the cultural chain between China
and the west. We tend to overlook this fact because it is the transient
Hun and Mongol invasions that chiefly obtrude into both western and
Chinese history. Whenever the nomads ceased boiling over, they
receded from the historians’ view. Obviously they could not have
migrated and fought and burnt and slain among each other
continuously. The more settled life at home, which they led most of
the time, and into which they were always inclined to take over the
religion, writing, and arts of the Orient, India and China, is the phase
of their existence most likely to be overlooked, but which, from the
point of view of the history of civilization, is far more important than
their evanescent conquests. In this underlying phase they were the
connectors of Near and Far East.
It is interesting in this connection that so far as more recent
transmissions between China and the west are datable or positively
traceable, they took place chiefly by the long land route through
central Asia. The first trade between the Roman empire and China of
the Han dynasty was overland; so was the introduction of Buddhism
and cotton from India. In each case sea communication came later. It
was scarcely before Mohammedan times that ocean trade between
China and the west became important.
On the other hand, the Chinese and other east Asiatics always
lacked, and still lack, several aspects of the grain-cattle-horse-wheel-
metals cluster that are very ancient and practically universal in
Europe, the Near East, and even central Asia. They do not use milk
or its products, wool, nor bread-leaven. It has been suggested that
the cluster was transmitted to China before these traits had been
added to it; and that when they finally might have reached China,
they found its people satisfactorily established in a culture containing
substitutes for these traits, and therefore resistive to them.
It is significant that even to-day northern China, within which the
oldest known China lay, still cultivates wheat, barley, and millet, and
breeds cattle, horses, sheep, goats, and swine, as did the Swiss
lake-dwellers; whereas in southern China the typical grain and
animal are rice and the buffalo, as in Indo-China and Malaysia.
There are evidently two fundamentally distinct economic systems
here, characteristic respectively of Europe and west and north Asia,
and of southeastern Asia; and evolving in the main independently.
The first civilized China, that of the Chou period, that which produced
Confucius as its literary standardizer, and has chiefly shaped
Chinese traditions and institutions ever since, belonged to the great
northern and western cycle; was in fact its easternmost outpost.
This brings up the question whether Chinese writing could not also
have sprung from a western source, notably the Sumerian
Cuneiform, which it superficially resembles in its linear, non-pictorial
strokes, and in its mixed ideographic-phonetic method. The
connection has indeed been asserted, but no satisfactory evidence
of specific correspondences has been adduced. The most that it
seems valid to maintain is that a remote connection is thinkable; a
connection not extending beyond a limited number of characters or
the idea or method of writing. The earliest Chinese characters
preserved on bronzes are nearly two thousand years younger than
the most ancient inscriptions of the system which developed into the
classic Cuneiform. Both systems are fairly crystallized when they first
come to our knowledge. Their formative stages, in which such
connection as they might have would be most apparent, are
obscure. It is well to remember that Cuneiform and Egyptian
hieroglyphic, which were virtually contemporaneous and much
nearer to each other geographically, have not yet been brought into
specific relation as regards their origins.
254. Korea
Korea has repeatedly been under the political authority of China,
more often autonomous, but for three thousand years has been
dependent on China culturally. Non-Chinese influences have also
reached it: such as an alphabet of Indian origin (§ 149); and probably
the earliest iron industry came in from Altaic sources. In its turn, the
peninsula transmitted to Japan: until about a thousand years ago,
Chinese writing and culture reached Japan mainly via Korea. The
spread of Chinese civilization was perhaps largely of the usual, slow,
diffusing kind, but was several times accelerated by the settlement in
Korea of groups of Chinese refugees, colonists, or adventurers; for
instance in Chou and Han times. The center of power and civilization
within Korea has gradually moved southwards, which suggests the
waning of original central Asiatic affiliations as Chinese ones
became stronger. The first realm to be defined was Fuyu on the
Sungari river. Then followed Kokorai or Korai, whence our name
Korea. In the centuries immediately before and after Christ, Shinra
and then Hiaksai, farther south on the peninsula, came into the lead.
By the beginning of this period not only the writing but the classic
books of China had been introduced. Since the fourteenth century
Confucianism has been the state religion—though the people had
long before become Buddhists—and literary examinations for office
of the Chinese type have been in vogue.
255. Japan
Japan is the one country of eastern Asia from which considerable
prehistoric data are available. There are indeed no indubitable
evidences of any Palæolithic culture or race, but shellheaps and
burial mounds abound and have been explored. The shell deposits,
of which 4,000 have been found, are probably the accumulation of
refuse of occupation by the Ainu, the first known inhabitants of
Japan, now surviving only in the extreme north of the island chain
and in Sakhalien, and still a primitive people. This race is different
from the Japanese, and has often been classified as Caucasian (§
27). The shell deposits show the aborigines to have been fishers and
hunters, without agriculture or edible domestic animals. They had the
bow, dog, incised hand-made pottery, and ground stone axes, and
were thus approximately in an early Neolithic stage.
A somewhat dubious bronze age is sparsely represented in
southern Japan. It has been ascribed to an invasion about the