Full Download Arteriovenous and Cavernous Malformations 1st Edition Edition Robert F. Spetzler PDF

Download More ebooks [PDF]. Format PDF ebook download PDF KINDLE.
Full download ebooks at ebookmass.com
Arteriovenous and Cavernous

Malformations 1st Edition Edition Robert
F. Spetzler
For dowload this book click BUTTON or LINK below
https://ebookmass.com/product/arteriovenous-and-
cavernous-malformations-1st-edition-edition-
robert-f-spetzler/
OR CLICK BUTTON
DOWLOAD NOW
Download More ebooks from https://ebookmass.com

More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Revision Lumbar Spine Surgery 1st Edition Robert F.

Heary
https://ebookmass.com/product/revision-lumbar-spine-surgery-1st-
edition-robert-f-heary/
Algebra and Trigonometry, 7th Global Edition Robert F.

Blitzer
https://ebookmass.com/product/algebra-and-trigonometry-7th-
global-edition-robert-f-blitzer/
Operations and Supply Chain Management: The Core 6th

Edition F. Robert Jacobs
https://ebookmass.com/product/operations-and-supply-chain-
management-the-core-6th-edition-f-robert-jacobs-2/
Algebra and Trigonometry, 4e - Instructor’s Solutions

Manual 4th Edition Robert F. Blitzer
https://ebookmass.com/product/algebra-and-
trigonometry-4e-instructors-solutions-manual-4th-edition-robert-
f-blitzer/
Operations and Supply Chain Management: The Core, 6th
Edition F. Robert Jacobs
management-the-core-6th-edition-f-robert-jacobs/
Pocket Atlas Of Emergency Ultrasound 2nd Edition

Edition Robert F. Reardon
https://ebookmass.com/product/pocket-atlas-of-emergency-
ultrasound-2nd-edition-edition-robert-f-reardon/
Beyond Audit: Auditing Remotely and Delivering Value

(Wiley Corporate F&A) 1st Edition Robert L. Mainardi
https://ebookmass.com/product/beyond-audit-auditing-remotely-and-
delivering-value-wiley-corporate-fa-1st-edition-robert-l-
mainardi/
Operations and Supply Chain Management, 17e 17th

International Student Edition Edition F. Robert Jacobs
management-17e-17th-international-student-edition-edition-f-
robert-jacobs/
Administración de operaciones: producción y cadena de

suministros 15th Edition F. Robert Jacobs
https://ebookmass.com/product/administracion-de-operaciones-
produccion-y-cadena-de-suministros-15th-edition-f-robert-jacobs/
HANDBOOK OF CLINICAL
NEUROLOGY
Series Editors
MICHAEL J. AMINOFF, FRANÇOIS BOLLER, AND DICK F. SWAAB
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
© 2017 Elsevier B.V. All rights reserved.
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.
British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress
ISBN: 978-0-444-63640-9
For information on all Elsevier publications

visit our website at https://www.elsevier.com/books-and-journals
Publisher: Nikki Levy

Editorial Project Manager: Kristi Anderson
Production Project Manager: Sujatha Thirugnana Sambandam
Cover Designer: Alan Studholme
Typeset by SPi Global, India

Handbook of Clinical Neurology 3rd Series
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
I.J. Abecassis W. Brinjikji

Department of Neurological Surgery, University of Department of Radiology, Mayo Clinic, Rochester, MN,
Washington, Seattle, WA, USA USA
A.A. Abla P.A. Brown

Department of Neurological Surgery, University of Department of Radiology, Duke University, Durham,
Arkansas for Medical Sciences, Little Rock, AR, USA NC, USA
J.R. Adler A. Can

Department of Neurosurgery, Stanford University, Department of Neurosurgery, Brigham and Women’s
Stanford, CA, USA Hospital and Harvard Medical School, Boston, MA,
USA
F.C. Albuquerque
Department of Neurosurgery, Barrow Neurological
P.R. Chen
Institute, St. Joseph’s Hospital and Medical Center,
Department of Neurosurgery, University of Texas Health
Phoenix, AZ, USA
Science Center, Houston, Texas, USA
R.O. Almefty
Department of Neurosurgery, Barrow Neurological A.J. Clark
Institute, St Joseph’s Hospital and Medical Center, Department of Neurological Surgery, University of
Phoenix, AZ, USA California, San Francisco, CA, USA
S. Ambekar M.L. Cohen

Department of Neurosurgery, University of Miami, Department of Pathology, University Hospitals, Case
Miami, FL, USA Western Reserve University, Cleveland, OH, USA
G.S. Atwal E.M. Cox

Department of Neurosurgery, Barrow Neurological Department of Neurosurgery, University Hospitals, Case
Institute, St. Joseph’s Hospital and Medical Center, Western Reserve University, Cleveland, OH, USA
Phoenix, AZ, USA
A.L. Day
N.C. Bambakidis
Department of Neurosurgery, University of Texas
Department of Neurosurgery, University Hospitals, Case
Medical School at Houston, Houston, TX, USA
Western Reserve University, Cleveland, OH, USA
D.L. Barrow D. Ding

Department of Neurosurgery, Emory University, Atlanta, Department of Neurosurgery, University of Virginia,
GA, USA Charlottesville, VA, USA
S.C. Bir R.L. Dodd

Department of Neurosurgery, Louisiana State University Department of Neurosurgery, Stanford University,
Health Sciences Center, Shreveport, LA, USA Stanford, CA, USA
xii CONTRIBUTORS
R. Du B. Jian
Department of Neurosurgery, Brigham and Women’s Department of Neurological Surgery, University of
Hospital and Harvard Medical School, Boston, California, San Francisco, CA, USA
MA, USA
M.Y.S. Kalani
A.F. Ducruet
Department of Neurological Surgery, University of
Pittsburgh Medical Center, Pittsburgh, PA, USA
Phoenix, AZ, USA
M.S. Elhammady
Department of Neurosurgery, University of Miami, A. Kaul
Miami, FL, USA Department of Neurological Surgery, Harborview
Medical Center, University of Washington, Seattle,
J.A. Ellis WA, USA
Department of Neurosurgery, Emory University School
of Medicine, Atlanta, GA, USA L. Kim
Department of Neurological Surgery and Department
C. Ene
of Radiology, University of Washington, Seattle,
Department of Neurological Surgery, Harborview
WA, USA
Medical Center, University of Washington, Seattle,
WA, USA
D. Kondziolka
C. Gesteira Benjamin Department of Neurosurgery and Center for Advanced
Department of Neurosurgery, NYU Langone Medical Radiosurgery, NYU Langone Medical Center,
Center, New York, NY, USA New York, NY, USA
H.E. Goldstein G. Lanzino

Department of Neurosurgery, The Neurological Institute, Department of Neurosurgery, Mayo Clinic, Rochester,
Columbia University Medical Center, New York, NY, USA MN, USA
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
R.C. Heros P. McCormick

Department of Neurosurgery, University of Miami, Department of Neurological Surgery, Columbia
Miami, FL, USA University Medical Center, Neurological Institute of
New York, New York, NY, USA
O.R. Idowu
Division of Neuroradiology, Russell H. Morgan C.G. McDougall
Department of Radiology and Radiological Science, Department of Neurosurgery, Barrow Neurological
Johns Hopkins University School of Medicine, Institute, St. Joseph’s Hospital and Medical Center,
Baltimore, MD, USA Phoenix, AZ, USA
CONTRIBUTORS xiii
G. Mendez R.F. Rudy
Department of General Surgery, Rush University Department of Neurosurgery, Brigham and Women’s
Medical Center, Chicago, IL, USA Hospital and Harvard Medical School, Boston,
MA, USA
K. Moon
Department of Neurosurgery, Barrow Neurological M.J. Rutkowski
Institute, St. Joseph’s Hospital and Medical Center, Department of Neurological Surgery, University of
Phoenix, AZ, USA California, San Francisco, CA, 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, The Neurological Institute,
Columbia University Medical Center, New York, NY, USA
Phoenix, AZ, USA
A. Nanda R.F. Spetzler

Department of Neurosurgery, Louisiana State Department of Neurosurgery, Barrow Neurological
University Health Sciences Center, Shreveport, Institute, St. Joseph’s Hospital and Medical Center,
LA, USA Phoenix, AZ, USA
M. Narayanan R.M. Starke

Department of Neurosurgery, Barrow Neurological Department of Neurosurgery, University of Virginia,
Institute, St. Joseph’s Hospital and Medical Center, Charlottesville, VA, USA
Phoenix, AZ, USA
H. Sun
J.W. Osbun Department of Neurosurgery, Barrow Neurological
Department of Neurosurgery, Emory University, Atlanta, Institute, St. Joseph’s Hospital and Medical Center,
GA, USA Phoenix, AZ, USA
M. Otten
Department of Neurological Surgery, Columbia E.S. Sussman
University Medical Center, Neurological Institute of Department of Neurosurgery, Stanford University,
New York, New York, NY, USA Stanford, CA, USA
A. Ozpinar A.H. Turkmani

Department of Neurological Surgery, University of Department of Neurosurgery, University of Texas
Pittsburgh Medical Center, Pittsburgh, PA, USA Medical School at Houston, Houston, TX, USA
M.R. Reynolds D.D. Wang

Department of Neurosurgery, Emory University, Atlanta, Department of Neurological Surgery, University of
GA, USA California, San Francisco, CA, USA
xiv CONTRIBUTORS
K.Y. Wang D.S. Xu
Division of Neuroradiology, Russell H. Department of Neurosurgery, Barrow Neurological
Morgan Department of Radiology and Radiological Institute, St. Joseph’s Hospital and Medical Center,
Science, Johns Hopkins University School of Phoenix, AZ, USA
Medicine, Baltimore, MD and Department of
J.M. Zabramski
Radiology, Baylor College of Medicine, Houston,
TX, USA
Phoenix, AZ, USA
G.M. Weiner
Department of Neurological Surgery, University A.R. Zomorodi
of Pittsburgh Medical Center, Pittsburgh, Department of Neurosurgery, Duke University, Durham,
PA, USA NC, USA
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/B978-0-444-63640-9.00001-1
© 2017 Elsevier B.V. All rights reserved
Chapter 1
Epidemiology, genetics, pathophysiology, and prognostic

classifications of cerebral arteriovenous malformations
ALP OZPINAR1*, GUSTAVO MENDEZ2, AND ADIB A. ABLA3

1
Department of Neurological Surgery, University of Pittsburgh Medical Center, Pittsburgh, PA, USA
2
Department of General Surgery, Rush University Medical Center, Chicago, IL, USA
3
Department of Neurological Surgery, University of Arkansas for Medical Sciences, Little Rock, AR, USA
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.
INTRODUCTION presentations and symptomatology, as well as a wide var-

iation in severity of the effect of an AVM on a patient’s
Vascular disease entities manifest in various forms, and overall well-being. The literature suggests that most
are particularly dangerous when high-flow arterial sys- brain AVMs (bAVMs) are identified following evalua-
tems erroneously communicate directly with venous sys-
tion of the etiology of acute cerebral hemorrhage, or inci-
tems. Arteriovenous malformations (AVMs) are vascular
dentally on imaging associated with seizure or headache
deformities involving fistula formation of arterial to
workup (Hofmeister et al., 2000; Conger et al., 2015). As
venous structures without an intervening capillary bed.
reinforced by Conger et al. (2015), subsequent to detec-
Such anomalies can prove fatal as the high arterial flow tion of the AVM, history and physical examination
can disrupt the integrity of venous walls, thus leading to coupled with critical imaging, including computed
sequelae such as hemorrhage. The central nervous sys- tomography (CT), magnetic resonance imaging (MRI),
tem (CNS) is particularly sensitive to such insults as and catheter angiogram, allow neurosurgeons to form a
intracranial hemorrhage secondary to AVM rupture can clinical picture of the AVM patient as well as visualize
prove profoundly disabling or deadly. Diagnosis of these the structure and hemodynamics of the vascular malfor-
lesions can prove challenging, as intracranial AVMs rep- mation (Conger et al., 2015).
resent a heterogeneous vascular pathology with various
*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
better stratify patients between surgical and endovascular

50 treatment strategies. The Buffalo score represents a new
tool with potential utility in guiding endovascular man-
40 agement of intracranial AVMs, which will likely undergo
additional external validation moving forward.
30
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).
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.
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.
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.
http://dx.doi.org/10.1016/B978-0-444-63640-9.00002-3
Chapter 2
The natural history of cerebral arteriovenous malformations

ANIL CAN, BRADLEY A. GROSS, AND ROSE DU*
Department of Neurosurgery, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA
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.
(Waltimo, 1973; Pasqualin et al., 1985; Kader et al.,

INTRODUCTION 1996; Lee et al., 2002; Buis et al., 2004).
Cerebral arteriovenous malformations (AVMs) are a Cerebral AVMs are typically located in the cerebral
complex of abnormal arteries and veins consisting of hemispheres, but may be located in any region of the
direct fistulous connections without normal intervening brain, including the cerebellum, brainstem, or spinal
capillary beds or functional neural tissue (Steinheil, cord. Size of AVMs may vary from occult or cryptic,
1895; McCormick, 1966). Although their exact patho- invisible on angiography or during surgery, to giant
genesis and pathophysiology remain poorly understood, AVMs, involving one or several adjacent lobes, the entire
AVMs are generally considered congenital lesions that cerebral hemisphere, or even the whole brain
arise from arrested vascular embryologic development (Al-Rodhan et al., 1986). Due to this heterogeneity in
resulting in atypical differentiation in the capillaries size and location, cerebral AVMs may cause a wide range
and subsequent abnormal communication between arter- of clinical symptoms. Classically, the most frequent clin-
ies and veins (McCormick, 1966; Gross et al., 2015). ical symptoms include hemorrhage resulting from rup-
However, selected case reports of de novo AVM develop- ture, seizures, and focal neurologic deficits.
ment and experimental models in animals showing Treatment paradigms for AVMs continue to change as
de novo AVM formation have challenged congenital the- endovascular, microsurgical, and radiosurgical proce-
ories of AVM development (Gonzalez et al., 2005; dures evolve. However, essential for the management
Kilbourn et al., 2014). In addition, recurrent cerebral of these lesions is a thorough understanding of their nat-
AVMs after complete surgical resection have been ural history as it must be weighed against the risk of any
reported, as well as spontaneously disappearing AVMs anticipated treatment. Natural history not only includes
*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
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
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 arteriovenous 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.
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.
http://dx.doi.org/10.1016/B978-0-444-63640-9.00003-5
Chapter 3
Arteriovenous malformations: epidemiology, clinical

presentation, and diagnostic evaluation
JOSHUA W. OSBUN, MATTHEW R. REYNOLDS, AND DANIEL L. BARROW*

Department of Neurosurgery, Emory University, Atlanta, GA, USA
Abstract
Brain arteriovenous malformations (AVMs) represent an uncommon disease of the central nervous system
characterized by an arteriovenous shunt in which one or multiple arterial pedicles feed into a vascular
nidus, creating early drainage into a venous outflow channel. These lesions are considered to be congenital
and can come to clinical attention in a variety of ways such as seizure, intracranial hemorrhage, chronic
headache or progressive neurological deficit. We focus on the epidemiology, clinical presentation and
diagnostic evaluation in this chapter.
INTRODUCTION can vary greatly from patient to patient in terms of angio-

architectural features related to size, location, and venous
Arteriovenous malformations (AVMs) represent an
drainage pattern. Multiple AVMs in a single patient are
uncommon disease of the central nervous system with
uncommonly observed. No hereditary cause has been esta-
important clinical implications. These lesions can occur
blished and clustering in families is exceedingly infrequent
anywhere in the brain and spinal cord and are a type of
in what is already a rare disease (Berman et al., 2000; van
arteriovenous shunt in which one or multiple arterialized
Beijnum et al., 2014). They are associated with a few syn-
pedicles feed into a vascular nidus and drain in a variety dromes, such as hereditary hemorrhagic telengiectasia
of venous outflow channels. This gives the lesion its clas- (Osler–Weber–Rendu), Wyburn–Mason syndrome, von
sic hallmark of “early venous drainage” on formal Hippel–Lindau disease, and Sturge–Weber syndrome.
catheter-based cerebral angiography. These lesions are Although several mechanisms of AVM formation in the
generally considered to be congenital and can come to developing fetus have been proposed, the heterogeneity
clinical attention in a number of ways, such as seizure, and lack of inheritance of these lesions cause their precise
chronic headache, progressive neurologic deficit, or intra- etiology to remain idiopathic.
cranial hemorrhage. Some remain asymptomatic and are A few population-based studies have been performed to
discovered only incidentally on cerebral imaging for unre- aid in describing both the incidence and prevalence of brain
lated reasons. Because AVMs are associated with a life- AVMs. One of the earliest studies was data retrospectively
long risk of hemorrhage (Ondra et al., 1990), these recorded in Olmsted County, Minnesota, over a 27-year
lesions require proper evaluation, diagnostic workup, period, which captured an incidence of 0.82 AVMs per
and consideration for treatment based upon a variety of
100 000 persons based upon patients presenting with intra-
factors, including age, size, location, and drainage pattern.
cranial hemorrhage (Brown et al., 1996). The New York
Islands AVM study is a registry and surveillance program
EPIDEMIOLOGY
of the population of the New York Islands (Manhattan,
Brain AVMs are most likely congenital lesions that Staten Island, and Long Island) regarding patients with
form for unclear reasons in the developing embryo. They cerebrovascular anomalies (Stapf et al., 2003). The 19
*Correspondence to: Daniel L. Barrow, MD, 1365 Clifton Road Suite B6200, Atlanta GA 30322, USA. Tel: +1-404-778-5770,
E-mail: Daniel_barrow@emoryhealthcare.org
26 J.W. OSBUN ET AL.
Fig. 3.1. A 34-year-old man who presented with a first-time seizure. (A) Anteroposterior and (B) lateral angiogram of the left
internal carotid demonstrates a 4-cm arteriovenous malformation (AVM) nidus in the left frontal lobe fed by branches of the middle
cerebral artery. Drainage is via both a superficial vein into the sagittal sinus (arrowheads) and a deep vein into the internal cerebral
veins (arrows). (C) T2-weighted magnetic resonance demonstrated the appearance of “flow voids” (white arrowhead) in the frontal
lobe signifying the AVM nidus. Due to these high-risk angiographic features, treatment was planned with stereotactic radiosurgery.
Unfortunately, the patient suffered a devastating hemorrhage (D) while awaiting definitive treatment. Images courtesy of Emory
University Hospital Department of Neurosurgery.
major area hospitals in the study captured 95% of the any one race or ethnicity, a few studies have identified
region’s 9 429 541 inhabitants as of the 2000 Census. that Hispanic populations have a two- to threefold
The study prospectively registered new AVM cases in increase in the risk of intracranial hemorrhage from
the population with both hemorrhagic and nonhemorrhagic AVM (Kim et al., 2007; Yang et al., 2015).
presentation. Initial results suggested an incidence of 1.34
per 100 000 and an incidence of first-time AVM hemor-
CLINICAL PRESENTATION AND
rhage of 0.51 per 100 000. A population study in Western
NATURAL HISTORY
Australia reported an incidence of 0.89 per 100 000
(ApSimon et al., 2002) and a similar Scottish study repor- AVMs can present with a variety of clinical symptoms.
ted a rate of 1.12 per 100 000 (Al-Shahi et al., 2003). In Intracranial hemorrhage, seizure, chronic headache,
summary, a collection of population-based studies from and progressive focal neurologic deficit are the most
the literature approximate the incidence of brain AVMs common presenting symptoms. In the modern era of
at 0.69–1.32 per 100 000 (Brown et al., 1996; Berman magnetic resonance imaging (MRI), AVMs are being ini-
et al., 2000; Stapf et al., 2001; ApSimon et al., 2002; tially diagnosed incidentally on imaging performed for
Al-Shahi et al., 2003; Choi and Mohr, 2005; Laakso and potentially unrelated reasons, such as headache, tinnitus,
Hernesniemi, 2012). stroke, and/or transient ischemic attack symptoms, or
AVMs may present at any age and distribution among other vague neurologic complaints.
sexes is equal. Although AVMs are congenital, they most Intracranial hemorrhage is the most common clinical
commonly come to clinical attention in the second presentation (See Fig. 3.1). AVMs are estimated to
through fourth decades of life. While no increase in account for 2–4% of all hemorrhagic strokes and 1–2%
the incidence or prevalence of AVMs is associated with of all total strokes (Choi and Mohr, 2005; Choi et al.,
EPIDEMIOLOGY ARTERIOVENOUS MALFORMATIONS 27
2006; Laakso and Hernesniemi, 2012; Gross and Du, intervention by surgery, embolization, or stereotactic
2013). Population-based studies have reported that the radiosurgery (Mohr et al., 2014). The primary outcome
incidence of hemorrhage from AVM is 0.51–1.11 per measure was death or symptomatic stroke and the
100 000 per year. Natural-history studies report a risk secondary outcome was clinical impairment (modified
of hemorrhage for known AVMs to be 2–5% per year Rankin Scale of 2). During the study period, 1740
(Ondra et al., 1990; Choi et al., 2006; van Beijnum patients were screened, yet only 726 were deemed eli-
et al., 2011; Gross and Du, 2013). In all, 5–25% of all gible to enroll in the study. A total of 323 patients
AVM hemorrhages are fatal. Data from the UCSF Brain refused to enroll and 177 had their management deci-
AVM Study Project have suggested that age, Hispanic sion made by the treating team outside of the random-
race, hypertension, and previous intracranial hemorrhage ization process. The study was halted after randomizing
are associated with increased risk of intracranial hemor- 223 patients with 114 randomized to intervention and
rhage. A recent study from Johns Hopkins reported a 109 to medical management. In the intervention group,
threefold increased risk of hemorrhage in nonwhites. 5 were treated with microsurgical resection, 30 with
Particular angiographic and MR features have not defin- embolization, 31 with radiotherapy, and 28 with multi-
itively been demonstrated to increase the risk of hemor- modality therapy. The primary endpoint was achieved
rhage, but in general there is a trend towards increased in 35 (30.7%) of the 114 patients randomized to treat-
risk of hemorrhage in AVMs with multiple arterial ment and in 11 (10.1%) of the 109 patients in the med-
feeders, deep-seated locations in the brain, exclusively ical management arm. The secondary outcome was
deep drainage, and feeding-artery aneurysms (Stapf achieved in 24 (46.2%) patients receiving intervention
et al., 2006; Lv et al., 2011, 2013; Hetts et al., 2014). and 8 (15.1%) of the medically managed patients at
Smaller AVMS and those with limited venous drainage 30 months. The authors concluded that medical man-
have been reported to have a higher risk of hemorrhage. agement is superior to intervention in preventing death
Patients who present with intracranial hemorrhage or stroke in patients with unruptured AVMs followed
have symptoms that range from sudden severe headache, for 33 months. The design and interpretation of this
decreased level of consciousness, hemiparesis, neglect study have raised a number of legitimate concerns. It
and gaze preference, aphasia, cranial neuropathy, visual is unclear why such a large number of eligible patients,
field deficits, and coma. Approximately 25% of patients who did not refuse enrollment, were not randomized.
expire from their first hemorrhage. Secondly, the study design provides no standardization
Seizures are another common presentation of brain of the treatment arm. Anything less than complete oblit-
AVMs and occur in 20–29% of patients upon initial pre- eration of an AVM does nothing to protect patients from
sentation (van Beijnum et al., 2011; Garcin et al., 2012; future hemorrhage. Many patients in ARUBA were
Laakso and Hernesniemi, 2012; Gross and Du, 2013; Lv treated with suboptimal therapy that did not accomplish
et al., 2013; Galletti et al., 2014; Mohr et al., 2014; Ding the goal of complete obliteration. A recent large
et al., 2015a, b). A recent study from China found that meta-analysis of treated AVMs reported obliteration
20.9% of patients presented with seizure out of a series rates of 96% for surgical resection, 38% for stereotactic
of 3299 patients diagnosed with brain AVMs (Tong radiosurgery, and 13% for embolization (van Beijnum
et al., 2016). Seizures are more common in male patients et al., 2011). In the ARUBA study, only 5 patients
with AVMs in cortical locations, particularly the frontal received surgical resection alone when 76 patients in
and temporal lobe, and increase in frequency of occur- the treatment arm had grade Spetzler–Martin grade
rence with increasing AVM size (Garcin et al., 2012; I or II AVMs. Embolization is generally not considered
Galletti et al., 2014; Ding et al., 2015b; Tong et al., a curative procedure. It is concerning that 30 of the
2016). For AVMs in deep location, up to 32% of insular treatment patients were managed with embolization
AVMs may present with seizure, while other deep loca- alone. Of the 114 patients randomized to treatment,
tions, such as the basal ganglia, thalamus, corpus callo- 53 had not completed therapy and 20 had not even ini-
sum, brainstem, and cerebellum, present with seizure in tiated therapy at the time of analysis. Given the lack of
less than 10% of cases (Ding et al., 2015b). With treat- information and heterogeneity regarding therapies, it is
ment, seizure control rates range from 49% to 85% impossible to understand what the treatment arm
depending on treatment strategy (surgery, radiosurgery, represents.
or embolization), with the highest seizure freedom being Many thoughtful critics of this study have pointed out
obtained in patients with confirmed angiographic obliter- that it is irresponsible to conclude superiority of medical
ation after radiosurgery (Mohr et al., 2014; Ding et al., management for AVMs when the treatment arm con-
2015b; Moon et al., 2015; Nerva et al., 2015, 2016). sisted of therapeutic options that are well below the cur-
The ARUBA trial was a randomized study that rent standard of care for elimination of AVMs. Finally,
sought to compare the natural history of AVMs to the 33-month follow-up for a disease with a lifelong risk
28 J.W. OSBUN ET AL.
of hemorrhage is clearly inadequate to make conclusions hemorrhage adjacent to the AVM nidus. Another classic
regarding the optimal treatment. The fact there was a appearance is a “flame-shaped” hemorrhage emanating
10% incidence of death or stroke over a time interval from the ventricle. In other cases an AVM may have an
of less than 3 years in the medically managed group associated flow-related aneurysm and present with sub-
underscores the need to provide a low-risk cure for these arachnoid hemorrhage from aneurysm rupture. In the
patients and emphasizes the risk of untreated AVMs over setting of acute hemorrhage, the diagnosis of AVM
the lifetime of the patient and warrants consideration of on noninvasive vascular imaging may not be possible
an appropriate treatment strategy. because the hemorrhage will obscure the details of sur-
While intracranial hemorrhage and seizures are the rounding blood vessels. This is particularly true in
most common presenting features of a brain AVM, many smaller AVMs.
present with less ominous symptoms such as chronic Regardless of how an AVM may be demonstrated on
headache or tinnitus. As stated previously, these lesions noninvasive imaging, the gold standard for diagnosis is a
are being discovered more and more frequently as inci- formal catheter-based cerebral angiogram. Since an
dental findings on noninvasive imaging. Regardless of angiogram is a time-resolved dynamic study, an AVM
the clinical presentation or route of discovery, any will demonstrate its hallmark of early venous drainage.
AVM harbors a risk of intracranial hemorrhage over Angiography will demonstrate an AVM’s arterial feeders,
the lifetime of the patient and warrants a full diagnostic nidus configuration, and pattern of venous drainage. It
workup so that an appropriate treatment strategy may will demonstrate any flow-related aneurysms, intranidal
be formed. aneurysms, or venous varices. Three-dimensional rota-
tional angiography is crucial in understanding the precise
angiographic architecture of the lesion, and greatly aids in
DIAGNOSTIC EVALUATION
treatment decisions such as the safety of embolization in
AVMs have signature hallmarks on several noninvasive regard to en passage vessels that simultaneously supply
imaging studies. The appropriate workup will depend on AVM nidus and normal cortex, and the technical feasi-
the patient’s presentation. AVMs are now presenting with bility of embolization in regard to microcatheterization
increased frequency on noninvasive imaging studies. In of feeding arteries. Three-dimensional spatial resolution
such cases, larger AVMs can appear as a hyperdense will also greatly assist in the formation of a surgical resec-
mass that can even be confused for hemorrhage on non- tion plan. Size, location, and drainage pattern are essen-
contrast computed tomography (CT). CT angiography tial for understanding surgical risk as exhibited by the
(CTA) can reveal a nidus or even arterial feeders and Spetzler–Martin grading scale, and angiography will
large tortuous veins. On plain magnetic resonance imag- help define most of these parameters. Angiography will
ing (MRI), T2-weighted sequences often demonstrate a therefore aid in surgical planning and help in determining
classic tangle of “flow voids” around the AVM nidus. whether preoperative embolization can be safe and
Magnetic resonance angiography (MRA) may poten- effective.
tially delineate the nidus, arterial feeders, or draining
veins. Even if an AVM is discovered incidentally on CONCLUSIONS
MRI, T2-weighted imaging is essential for understand-
ing the exact anatomic location and size of the nidus AVMs are an uncommon vascular malformation of the
for determining potential treatment decisions. Functional brain and spinal cord with important clinical implications
MRI testing can help determine the lesion’s relationship in terms of hemorrhagic stroke, epilepsy, and neurologic
to functionally important areas such as motor and lan- disability. An understanding of their epidemiology, clin-
guage areas, and can be extremely valuable for determin- ical presentation, and natural history is imperative when
ing the safety of surgical resection. A fine-cut volumetric forming a plan for full diagnostic workup. A full diagnos-
T2-weighted study or T1-weighted study with contrast tic workup, including both MRI and catheter-based cere-
can be used with neuronavigation systems for intraopera- bral angiography, is paramount in calculating a treatment
tive guidance. modality that will maximize both AVM obliteration and
Other patients will present in a more acute fashion. patient safety.
For patients presenting with seizures, often a noncon-
trast CT is performed to rule out intracranial hemor- REFERENCES
rhage, and occasionally a large AVM may be seen. Al-Shahi R, Bhattacharya JJ, Currie DG et al. (May 2003).
For AVMs with all but a micronidus, MRI will usually Prospective, population-based detection of intracranial vas-
demonstrate the AVM nidus and location. For the cular malformations in adults: the Scottish Intracranial
majority of patients who present with intracranial hem- Vascular Malformation Study (SIVMS). Stroke 34 (5):
orrhage, CT scan will demonstrate an intraparenchymal 1163–1169.
EPIDEMIOLOGY ARTERIOVENOUS MALFORMATIONS 29
ApSimon HT, Reef H, Phadke RV et al. (Dec 2002). with nonhemorrhagic neurologic deficits. World Neurosurg
A population-based study of brain arteriovenous malfor- 79 (3-4): 484–488.
mation: long-term treatment outcomes. Stroke 33 (12): Mohr JP, Parides MK, Stapf C et al. (2014). Medical manage-
2794–2800. ment with or without interventional therapy for unruptured
Berman MF, Sciacca RR, Pile-Spellman J et al. (Aug 2000). brain arteriovenous malformations (ARUBA): a multicen-
The epidemiology of brain arteriovenous malformations. tre, non-blinded, randomised trial. Lancet 383 (9917):
Neurosurgery 47 (2): 389–396; discussion 397. 614–621.
Brown Jr RD, Wiebers DO, Torner JC et al. (Apr 1996). Moon K, Levitt MR, Almefty RO et al. (Dec 2015). Safety and
Incidence and prevalence of intracranial vascular malfor- efficacy of surgical resection of unruptured low-grade arte-
mations in Olmsted County, Minnesota, 1965 to 1992. riovenous malformations from the modern decade.
Neurology 46 (4): 949–952. Neurosurgery 77 (6): 948–953.
Choi JH, Mohr JP (May 2005). Brain arteriovenous malforma- Nerva JD, Mantovani A, Barber J et al. (May 2015). Treatment
tions in adults. Lancet Neurol 4 (5): 299–308. outcomes of unruptured arteriovenous malformations with a
Choi JH, Mast H, Sciacca RR et al. (May 2006). Clinical out- subgroup analysis of ARUBA (A Randomized Trial of
come after first and recurrent hemorrhage in patients with Unruptured Brain Arteriovenous Malformations)-eligible
untreated brain arteriovenous malformation. Stroke 37 (5): patients. Neurosurgery 76 (5): 563–570; discussion 570;
1243–1247. quiz 570.
Ding D, Quigg M, Starke RM et al. (Sep 2015a). Cerebral arte- Nerva JD, Kim LJ, Barber J et al. (2016). Outcomes of multimod-
riovenous malformations and epilepsy, part 2: predictors of ality therapy in pediatric patients with ruptured and unrup-
seizure outcomes following radiosurgery. World Neurosurg tured brain arteriovenous malformations. Neurosurgery 78:
84 (3): 653–662. 695–707.
Ding D, Starke RM, Quigg M et al. (Sep 2015b). Cerebral arte- Ondra SL, Troupp H, George ED et al. (Sep 1990). The natural
riovenous malformations and epilepsy, part 1: predictors of history of symptomatic arteriovenous malformations of the
seizure presentation. World Neurosurg 84 (3): 645–652. brain: a 24-year follow-up assessment. J Neurosurg 73 (3):
Galletti F, Costa C, Cupini LM et al. (Mar 2014). Brain arte- 387–391.
riovenous malformations and seizures: an Italian study. Stapf C, Mohr JP, Pile-Spellman J et al. (2001). Epidemiology
J Neurol Neurosurg Psychiatry 85 (3): 284–288. and natural history of arteriovenous malformations.
Garcin B, Houdart E, Porcher R et al. (2012). Epileptic seizures Neurosurg Focus 11 (5): e1.
at initial presentation in patients with brain arteriovenous Stapf C, Mast H, Sciacca RR et al. (May 2003). The New York
malformation. Neurology 78 (9): 626–631. Islands AVM study: design, study progress, and initial
Gross BA, Du R (Feb 2013). Natural history of cerebral arte- results. Stroke 34 (5): e29–e33.
riovenous malformations: a meta-analysis. J Neurosurg 118 Stapf C, Mast H, Sciacca RR et al. (2006). Predictors of hem-
(2): 437–443. orrhage in patients with untreated brain arteriovenous mal-
Hetts SW, Cooke DL, Nelson J et al. (Jul 2014). Influence of formation. Neurology 66 (9): 1350–1355.
patient age on angioarchitecture of brain arteriovenous Tong X, Wu J, Lin F et al. (2016). The effect of age, sex and
malformations. AJNR Am J Neuroradiol 35 (7): lesion location on initial presentation in patients with brain
1376–1380. arteriovenous malformations. World Neurosurg 87:
Kim H, Sidney S, McCulloch CE et al. (Sep 2007). Racial/ 598–606.
ethnic differences in longitudinal risk of intracranial hem- van Beijnum J, van der Worp HB, Buis DR et al. (2011).
orrhage in brain arteriovenous malformation patients. Treatment of brain arteriovenous malformations: a sys-
Stroke 38 (9): 2430–2437. tematic review and meta-analysis. JAMA 306 (18):
Laakso A, Hernesniemi J (Jan 2012). Arteriovenous malfor- 2011–2019.
mations: epidemiology and clinical presentation. van Beijnum J, van der Worp HB, Algra A et al. (Nov 2014).
Neurosurg Clin N Am 23 (1): 1–6. Prevalence of brain arteriovenous malformations in
Lv X, Wu Z, Jiang C et al. (Jul–Aug 2011). Angioarchitectural first-degree relatives of patients with a brain arteriovenous
characteristics of brain arteriovenous malformations with malformation. Stroke 45 (11): 3231–3235.
and without hemorrhage. World Neurosurg 76 (1–2): 95–99. Yang W, Caplan JM, Ye X et al. (Aug 2015). Racial associ-
Lv X, Li Y, Yang X et al. (Mar–Apr 2013). Characteristics ations with hemorrhagic presentation in cerebral arterio-
of brain arteriovenous malformations in patients presenting venous malformations. World Neurosurg 84 (2): 461–469.
http://dx.doi.org/10.1016/B978-0-444-63640-9.00004-7
Chapter 4
Seizures associated with cerebral arteriovenous malformations

JOHANNES SCHRAMM*
Department of Neurosurgery, University of Bonn, Bonn, Germany
Abstract
Various types of seizures and epilepsy are associated with 20–45% of cerebral arteriovenous malforma-
tions (AVMs). The necessity to differentiate between occasional seizures, epilepsy with repetitive seizures,
and the much rarer drug-resistant epilepsy (DRE) is underlined. It is clear that where there is frequent
seizures or DRE, vascular surgeons should take epilepsy surgery aspects into account. The epidemiology
of AVM-associated seizures, assumed pathophysiologic mechanisms, most frequent seizures types, and
medical treatment are described. Depending on the severity of the epilepsy, the diagnostic workup, includ-
ing electroencephalogram (EEG), video-EEG, and, rarely, invasive evaluation, is explained. An invasive
presurgical workup is only necessary in rare cases of DRE. The indication to extend the resection to more
than just removal of the AVM is defined and the various specific resection techniques for this rare form are
outlined. In the vast majority of AVM cases removal of the AVM with some adjoining gliotic or hemo-
siderotic rim of cortex will be sufficient, however. In the majority of cases with preoperative epilepsy,
patients will be seizure-free after surgery. Patients who never had a seizure before AVM removal may
develop de novo epilepsy postoperatively (5–15%). Rates of seizure freedom after different treatments
(microsurgery, radiosurgery, endovascular) vary.
INTRODUCTION DEFINITIONS
Seizures are the second most frequent presentation of Seizures in cerebral AVM may be induced by a first hem-
cerebral arteriovenous malformations (AVMs). Seizures orrhage, but may be present in unruptured AVMs even
are found in 20–45% of AVM cases seen for treatment, for years. The World Health Organization (WHO) has
yet the major problem in cerebral AVM is the risk of hem- defined epilepsy as a chronic neurologic disorder charac-
orrhage which is around 2–3% annually. In large series terized by recurrent seizures, i.e., brief episodes of invol-
hemorrhage is the most frequent presentation in up to untary movement of a part or the entire body which may
50% of cases and the true significance of this is the result- or may not be accompanied by loss of consciousness and
ing morbidity and mortality. Thus, the major concern loss of control of bowel and bladder function (World
with cerebral AVM is prevention of bleeding or rebleed- Health Organization, 2016). The International League
ing with increased morbidity. However, having seizures Against Epilepsy (ILAE) prefers to call epilepsy a dis-
or epilepsy brings about a set of problems for the patient ease instead of disorder (International League Against
which may be considerable – many professions off Epilepsy, 2014). The WHO uses the term epilepsy if
limits, more accidents, higher mortality, social stigma, two or more unprovoked seizures are found, whereas a
no swimming, no car driving, no climbing, plus the need single seizure does not constitute epilepsy.
for medication with its side-effects. These factors make Seizures are caused by sudden synchronous electric
AVM-related epilepsy a significant factor to consider discharges in large groups of neurons. They may be
in case management to achieve a fully satisfactory out- triggered by various stimuli (e.g., trauma, hemorrhage,
come. This chapter deals with seizures and epilepsy types blood–brain barrier disruption). The patient with
found in patients with cerebral AVM. AVM-associated seizures carries a double burden. Not
*Correspondence to: Johannes Schramm, M.D., Medical Faculty, Bonn University, Sigmund-Freud Str. 25, 53127 Bonn, Germany.
E-mail: johannes.schramm@ukb.uni-bonn.de
32 J. SCHRAMM
only is there the fear of hemorrhage; epilepsy is also seen in seizures may be due to potentiation of excitatory
associated with many psychosocial problems and may transmissions or to the failure of inhibitory transmission.
need lifelong medication (with potential side-effects). The precise mechanisms are as yet unknown, but a
Patients with chronic epilepsy are also endangered by number of mechanisms have been described in experi-
seizure-associated injuries. Epilepsy patients have a mental setups, including gliosis, blood–brain barrier dis-
higher mortality rate and patients with drug-resistant epi- ruption, localized ischemia, hemosiderotic deposits, and
lepsy (DRE) have in fact a shortened life expectancy. The scars following microbleeds (LeBlanc and Rasmussen,
older classification into symptomatic and kryptogenic 1974; Stefan et al., 1987; Wolf et al., 1993; Englot
epilepsies has been replaced by a new classification of et al., 2012). It is known that not one single cellular
epilepsies or syndromes: localization-related, generalized mechanism is involved in the generation of seizures.
(idiopathic, special and other), and “unspecified” (World Numerous models of experimental epilepsy exist and sei-
Health Organization, 2016). Single or occasional seizures zures can be induced by topical application of chemical
are observed also in healthy brains associated with condi- substances (kainate, penicillin, aluminum oxide, or com-
tions such as seizure-provoking drugs or sleep depriva- pounds with metallic ions, e.g., hemosiderin). The pri-
tion: seizures in patients with epilepsy are frequently mary inhibitory neurotransmitter gamma-aminobutyric
associated with structural lesions, cortical malformations, acid (GABA) is presumed to play an important role in
and biochemical alterations in cortical tissue. epileptogenesis. For temporal-lobe epilepsy GABA
It may be practical to look at AVM-associated seizures increases have been demonstrated by microdialysis tech-
from the perspective of a neurosurgeon who sees the nique in the extracellular space. GABA homeostasis is
patient for the first time, and classify them into three complicated and may be influenced by many factors,
groups: such as quantity of GABA transporter proteins, and by
complicated astrocyte–pericyte interactions. Drugs
1. Sporadic seizures group: patients who have had one
effective in convulsive seizures can augment GABAer-
or two seizures only.
gic neurotransmission (Macdonald and McLean, 1986)
2. Chronic epilepsy group: all patients who have had
or can reduce the capability of cells to generate
more than two seizures, but do not fulfill the criteria
high-frequency discharges. Focal or generalized convul-
for DRE.
sions are associated with prolonged membrane deep
3. DRE: if seizures persist for over 2 years after failure
polarizations.
of adequate trials of two tolerated and appropriately
A more recent discovery was the significant role of
used antiepileptic drug (AED) schedules.
astroglia in epileptogenesis. Although astrocytes do
The mechanisms whereby epilepsy turns into DRE are not produce ictal discharges, they play a role in K+
unclear. DRE may include many or few seizures, e.g., buffering and metabotropic glutamate receptor signaling
four seizures per year, four seizures per month, or more. (Kovács et al., 2012). Blood–brain barrier disruption has
The definition of seizure freedom must be based been found to play a role in epileptogenesis, especially
somehow on the typical interval between seizures. The the leaking of serum proteins, leading to inflammatory
ILAE has therefore decided that sustained seizure free- response, and impaired homeostasis of the extracellular
dom is achieved when the seizure-free period after a space, leading to increased excitability of neurons, which
new drug regime or after surgery was longer than at least results in an altered balance of excitatory and inhibitory
three times the longest interseizure interval (Kwan elements (Friedman and Heinemann, 2012).
et al., 2010).
MECHANISMS IN AVM-RELATED
EPILEPSY
PATHOPHYSIOLOGY OF
EPILEPTOGENESIS Lesional epilepsies, such as AVM-related epilepsy, affect
brain areas differently; they are more common in tempo-
AVMs consist of atypical and malformed blood vessels ral and frontal lobe (Friedman and Heinemann, 2012).
and do not contain neurons. Seizures are not generated Several factors have been found to be associated with
by the AVM proper but they are induced by the effects AVM-related seizures: hemorrhage from the AVM
the AVM exerts on the adjacent cortex. The basic rule (Thorpe et al., 2000), size of the AVM, feeders from
is that only neurons can generate seizures and thus cor- the middle cerebral artery territory, associated venous
tical lesions are most likely to produce seizures. In prin- varices, superficial large AVMs (Morello and Borghi,
ciple, seizures can develop if the balance between 1973; Crawford et al., 1986b), and superficial venous
excitation and inhibition at the cellular level or synaptic drainage (Turjman et al., 1995; Garcin et al., 2012). Of
level is disturbed. The synchronized neuronal discharges 27 cases, histology of excised seizure foci contained
SEIZURES ASSOCIATED WITH CEREBRAL ARTERIOVENOUS MALFORMATIONS 33
gliosis in 26, hemosiderin deposits in 10, and focal hem- may stop generating seizures after the primary lesion
orrhages in 4 (Yeh et al., 1990). In a recent series of 103 with its surrounding epileptogenic cortex has been
patients (von der Brelie et al., 2015) temporal and frontal removed. It may be warranted to opt for removal of
localization were most frequent (37.9% and 33%). In the primary lesion and secondary focus in one procedure,
that series unruptured AVMs were associated with sei- or for a stepwise approach with invasive evaluation and
zures in 61%. A seizure as a first symptom was reported secondary focus resection only done when seizures per-
in that series in 43%; other authors report 15–35% of sist after lesionectomy.
patients having a seizure as a first symptom (Perret The more frequently the seizures occur, the larger the
and Nishioka, 1966; Morello and Borghi, 1973; lesion, the more cortex is affected, and the longer the
Michelsen, 1979; Graf et al., 1983; Fults and Kelly, duration of the epilepsy, the greater the chance that epi-
1984; Wilkins, 1985; Brown et al., 1988; da Costa leptogenic cortex is not only present in the immediate
et al., 2009). vicinity of the AVM. A more extensive cortical resection
The storage of albumin in astroglial cells around may be necessary, as demonstrated earlier (Yeh et al.,
vascular lesions has recently been implicated in 1990; Yeh and Privitera, 1991; von der Brelie et al.,
AVM-associated epileptogenesis (Raabe et al., 2012). 2015). Seizure semiology, ictal electroencephalogram
Since seizures are induced by the effects of AVM on (EEG) findings, and the occasional invasive evaluation
the surrounding cortex via ischemia, microhemorrhages, with grids or strips may pinpoint the area of a secondary
gliosis, inflammatory processes, and albumin extravasa- focus (Yeh and Privitera, 1991). The big question is when
tion, it is no surprise that epilepsy may cease after to go for a more extensive epilepsy investigation, even
removal of the AVM. On the other hand, there must be with invasive recordings. DRE as opposed to chronic
mechanisms that cause persistence of seizures despite epilepsy or sporadic seizures is a definitive candidate
removal of the AVM, most likely because these changes for extensive presurgical evaluation (von der Brelie
induced permanent epileptogenic areas in the neighbor- et al., 2015).
ing cortex. One such factor may be the duration of the
epilepsy until the removal of the primary lesion (e.g.,
NATURAL HISTORY
AVM). The longer the epileptogenic process, brought
AND EPIDEMIOLOGY
on and maintained by the presence of an AVM, has
existed, the more likely that the AVM-related epilepsy Seizures are the presenting symptom in 20–40% of cases
may persist after removal of the vascular malformation. in larger series (Kida et al., 2000; Hoh et al., 2002;
Epilepsy occurs more frequently in larger AVMs Josephson et al., 2011; von der Brelie et al., 2015).
(McKissock and Paterson, 1956; Hyun et al., 2012). The most frequent seizure type in AVM patients is gen-
eralized tonic-clonic seizure (Kida et al., 2000; Hoh et al.,
2002; Josephson et al., 2011; von der Brelie et al., 2015),
EXTENDED EPILEPTOGENIC ZONE
although in the DRE subgroup initial focal seizures were
AND SECONDARY FOCI
more frequent (von der Brelie et al., 2015). Generalized
Epileptogenesis may take place in cortical areas distant tonic-clonic seizures were most frequent in a large series
from the immediate border of the AVM, where it nor- (Piepgras et al., 1993) and of the focally starting seizures
mally has its primary location. This can occur in two half were associated with secondary generalization.
different forms: (1) the primary ictal onset zone enlarges Larger AVMs were found to be more likely to present
over time to form a larger ictogenic area in direct with seizures, whereas AVMs < 3 cm were more likely
proximity to the AVM; and (2) a second ictogenic area to present with hemorrhage (Piepgras et al., 1993). Tem-
develops away from the cortex surrounding the porally located AVMs have a tendency to present with
AVM – a process that usually takes time and happens complex partial seizures, many of which, however, later
only in some cases. These distant secondary foci are generalize.
thought to be induced by exposure to repetitive seizure The risk of developing seizures in an incidentally dis-
patterns from cortex close to the epileptogenic lesion. covered AVM, where no seizures had occurred so far, is
This kindling phenomenon is particularly prominent in relatively low (8% within 5 years: Josephson et al., 2011,
the mesial temporal lobe but may also occur in the neo- or 1.1% per person-year: Crawford et al., 1986a, b),
cortex, especially in the central area. If kindling in a tem- so prophylactic AED therapy may not be indicated. If
porally located AVM occurs, it may lead to secondary the patient had initially presented with hemorrhage the
hippocampal sclerosis, which may become an indepen- risk of a first seizure within the next 5 years increases
dent focus, then called dual pathology, with a higher risk to 23%. AVM patients who presented with a single
for persisting seizures despite removal of the AVM. seizure had a 5-year risk of developing epilepsy of 58%
However, some secondary epileptogenic cortical areas (Josephson et al., 2011). In a study of 153 nonoperated
34 J. SCHRAMM
patients who were followed up for a 20-year period, 18% first-line treatment is medical, unless AVM removal is
developed epilepsy (Crawford et al., 1986a, b). When diag- indicated. Patients with primary generalized seizures
nosed, younger patients aged 10–19 years were found to need other AEDs than do patients with focal seizures.
have a 44% risk compared to 6% risk in those who were Since recommended drug regimes change over time it
over the age of 30, during a 20-year follow-up period is wise to involve an epileptologist. The chance to
(Crawford et al., 1986a). The same study also confirmed achieve seizure freedom with AED is about 70% in
that temporal-lobe location carries the greatest risk of any epilepsy, but was only 60% in a large observational
developing epilepsy (Crawford et al., 1986b). population-based AVM study (Josephson et al., 2011).
The same study described that immediate AED prescrip-
Drug-resistant epilepsy tion increased the time to seizure recurrence and reduced
the time to 2-year seizure freedom. After successful
Whereas about 30% of all epilepsy patients suffer from
AVM removal in cases with previous chronic epilepsy
DRE, the percentage of DRE in AVM-related seizures
or DRE it is recommended to continue the AED for at
is much lower. A few studies have clearly described
least 1 year, and preferably 2 years, as is standard practice
DRE cases (Yeh and Privitera, 1991; Yeh et al., 1993;
in epilepsy surgery.
Lim et al., 2006); however their percentage of DRE in
the whole series was not given. It was 24 of 103 in the
own AVM series (von der Brelie et al., 2015), and less PRESURGICAL DIAGNOSTICS
than 1% AVM cases out of over 2000 resective epilepsy
The neuroradiologic routine examination is the same for
surgery cases in the Bonn epilepsy surgery program. In a
AVM cases with and without seizures: four-vessel digital
very large series 18% of epilepsy cases had DRE (Englot
subtraction angiography and magnetic resonance imag-
et al., 2012).
ing (MRI). The MRI examination may give hints with
In our own DRE group preoperative duration of epi-
regard to gliotic areas, edematous zones, and microhe-
lepsy was significantly longer (156 months vs. 37 months
morrhages at the AVM border, but also of edema in the
for chronic epilepsy and 6 months for sporadic seizures).
area of a more distended or stenotic draining vein as
Initial intracerebral hemorrhage was less frequent and
potential sources of distant epileptogenic cortex. The
frontal localization of the AVM was most common;
MRI in temporal AVM cases with chronic epilepsy or
patients more frequently had multimodal treatment
DRE should ideally be performed with sequences used
before surgery and they were twice as frequently associ-
for presurgical epilepsy workup. In order to obtain good
ated with high Spetzler–Martin grades compared to the
visualization of the hippocampus and other mesiotem-
total cohort (von der Brelie et al., 2015). More extensive
poral structures the axial slices should be inclined paral-
resections led to better results: 8 of 10 patients with
lel to the length axis of the hippocampus.
extended lesionectomy were seizure-free, but only 6 of
No specific electrophysiologic techniques need be
14 with standard resections (no statistical significance
applied for AVMs with only sporadic seizures.
because of low numbers). In another series factors pre-
For chronic epilepsy, as defined above, preoperative
disposing to develop DRE were hemorrhage, frontotem-
video-EEG is advisable if seizure frequency is high
poral lesion, male gender, and larger AVM size (3.1 cm
and seizure types or frequency are very disturbing, for
vs. 2.4 cm in nonseizure cases) (Englot et al., 2012).
example, affecting professional activity.
Few data on management of DRE in AVM cases are
Workup in DRE cases focuses on localization of the
available (Yeh et al., 1993; Cao et al., 2011; von der
ictogenic area, since AVM removal is done in a standard
Brelie et al., 2015). In another series with mostly
fashion. DRE cases need to have video-EEG with the aim
DRE cases, 25 of 54 patients had additional cortical
of recording at least two seizures. Video-EEG may not be
excisions, i.e., extended lesionectomies (Yeh et al.,
available everywhere and should be done in collabora-
1993); 12 of these were in a remote location, i.e., foci
tion with an epilepsy service. The seizure semiology
of secondary epileptogenesis. In our own series 13 of
and possible postictal neurologic deficits may give
24 DRE cases underwent epileptologic evaluation lead-
important hints regarding the presence and location of
ing to an epilepsy surgery approach in 11 of those
a so-called remote ictogenic focus.
13 cases (von der Brelie et al., 2015); in another series
The use of invasive evaluation with strip, depth, or grid
10 of 17 temporal AVMs were associated with a remote
electrodes will rarely be necessary. It will be indicated if
seizure focus (Yeh et al., 1993).
there is a discrepancy between the MRI and the findings
from the noninvasive evaluation, i.e., video-EEG record-
DRUG TREATMENT
ings of two seizures, and the localizing signs obtained from
For patients presenting with a first ever seizure but no careful history taking and the semiology of the seizures.
hemorrhage leading to the diagnosis of AVM, the A typical example would be a small AVM in the frontal
SEIZURES ASSOCIATED WITH CEREBRAL ARTERIOVENOUS MALFORMATIONS 35
lobe and a seizure semiology pointing to the temporal lobe. freedom, i.e., potentially an invasive evaluation to detect
DRE from temporal AVM must be particularly carefully extended areas of epileptogenicity. The indication for
evaluated because of the well-known propensity of the surgery may also be given for patients who cannot toler-
temporal lobe to develop secondary epileptogenic areas ate side-effects of anticonvulsants (Yeh et al., 1993).
(Yeh and Privitera, 1991; Cascino et al., 1992; Jooma
et al., 1995; Moore et al., 1999). Intraoperative electrocor-
Epilepsy surgery concepts applicable in
ticography does not reliably indicate the epileptogenic
AVM removals
zone or remote seizure foci because no seizures are
recorded. The resection of an AVM which presented with one or a
The implantation of strip and grid electrodes can be few seizures does not have to consider principles of epi-
used for extraoperative mapping of both, epileptogenic lepsy surgery, since the term epilepsy surgery is reserved
and functional cortex. It requires a first craniotomy for for cases with DRE. AVM removal for chronic seizures
the implantation and a second one for the resection, and DRE may be considered lesional epilepsy surgery.
but offers the huge advantage that mapping of cortical Thus, the vascular surgeon should be aware of the defi-
function can be performed outside the operating room nition of DRE (see section on definitions, above).
for lengthy periods if necessary. Video-EEG with AVM removal alone (in epilepsy surgery terminology
implanted electrodes is the most useful EEG recording a “pure lesionectomy”) may be supplemented by addi-
type for detection of seizure origin. It can help to plan tional removal of a cortical rim, or adjacent scarred/glio-
the resection of AVM close to eloquent cortex and the tic cortical areas. For lesional epilepsy several options
localization of secondary epileptogenic areas and their exist: simple lesionectomy (i.e., only the AVM is
topographic relationship to functional cortex. Again, col- removed), lesionectomy with rim of cortex, and extended
laboration with an epilepsy surgery group is very impor- lesionectomy, i.e., lesionectomy with rim plus resection
tant. Awake craniotomy can only be used for functional of secondary epileptogenic focus or dual pathology. The
mapping, not for precise localization of ictogenic areas, extended cortical removal may concern a defined cortical
since intraoperative seizures are normally not present, area in the vicinity of the lesion beyond the small cortical
and rather are to be avoided. rim or in temporal-lobe epilepsy it may include amygda-
lohippocampectomy, or, rarely, even temporal lobec-
tomy. Rasmussen already defined in 1983, for the
SURGICAL STRATEGIES
purpose of resective surgery in epilepsy, primary and sec-
Considering surgery, prevention of hemorrhage is the ondary localization. In cases of cerebral AVM the pri-
classic indication, not treatment of seizures. Thus, with mary localization would be the irritated or disturbed
an unruptured AVM discovered because of one or a cortex immediately bordering on the AVM, correspond-
few seizures, the seizures do not constitute an indication ing to the term lesionectomy with rim. The secondary
for surgery. The primary treatment for rare seizures is localization would be the adjacent cortex that must be
pharmacotherapy. The main aim is complete removal recruited to generate a clinical seizure. In a typical case
of the AVM using the same microsurgical techniques these two regions would be lying side by side and the sec-
as in cases without epilepsy. In patients with only a ondary location would correspond in some cases to a
few seizures, the chance to obtain seizure freedom with larger adjacent cortical area needing an extended lesio-
simple AVM removal is good, so no specific resection nectomy. The rarer variant of extended lesionectomy is
strategy appears necessary, apart from including small necessary if the recruited epileptogenic cortex lies distant
areas of gliosis and hemosiderotic cortex immediately from the immediate cortex around the AVM, such as, for
adjacent to the AVM border, provided eloquent cortex example, the hippocampus in a case of lateral superficial
is not affected. temporal AVM.
In patients with AVMs associated with chronic epi- During the very long debate on what procedure is best,
lepsy where DRE is present, it is worth being aware of some facts have evolved for lesion-related epilepsy in
some concepts of epilepsy surgery. Ideally the vascular general, not necessarily restricted to AVM-related sei-
surgeon confronted with these two conditions should zures. Although resection of the AVM alone may be asso-
know about and include epilepsy surgery aspects in the ciated with good postoperative control of seizure activity,
planning of the AVM removal. simple lesionectomy for temporal-lobe lesions has
Rarely, patients who do not initially wish to have their poorer results than lesionectomy combined with resec-
AVM removed but suffer from uncontrolled seizures are tion of mesial structures (Cascino et al., 1992; Jooma
candidates for surgery with the primary aim of achieving et al., 1995; Moore et al., 1999). Findings that support
seizure freedom. For these cases planning should include a greater likelihood of being seizure-free after a simple
considerations to optimize chances to achieve seizure lesionectomy were shorter duration of the seizure
36 J. SCHRAMM
disorder, lower seizure frequency, or good response to various patient groups or for different durations of the
antiepileptic treatment (Rassi-Neto et al., 1999). For sev- epilepsy disorder, two factors known to influence seizure
eral lesion types (benign tumors and vascular malforma- outcome.
tions) combined lesionectomy plus corticectomy were In principle, one would expect best outcomes after the
more successful (Weber et al., 1993). removal of the AVM with its surrounding damaged cor-
For chronic epilepsy with disturbing seizure types or tical rim. The AVM-derived factors known to induce epi-
affecting professional activity it may be advisable in leptogenesis should ideally be removed in the course of
those cases where the AVM is distant from eloquent cor- AVM resection (e.g., hemosiderin deposits, gliosis) or
tex to perform a lesionectomy with a wider rim, i.e., disappear shortly after (e.g., edema, localized ischemia).
AVM removal, including areas of gliotic and hemosi- Since some of these factors disappear only if they are
derotic cortex in the immediate vicinity plus 0.5 cm of mechanically removed, and others when the hemody-
healthy-appearing cortex. namic abnormalities around the AVM disappear, it
Although the superiority of an epilepsy surgery con- becomes clear that complete obliteration of the AVM
cept used in AVM removal with DRE or chronic epilepsy alone can lead to cessation of seizures. With microsur-
has been seen (Yeh et al., 1993; Jafar and Huaeng, 1999), gery the hemodynamic effects will be gone but also
but not proven so far, success in non-AVM cases supports the mechanical removal of structural changes has
its use. occurred, so a wider spectrum of ictogenic factors is dealt
with when microsurgery is used. If edema, pulsatile com-
pression, and relative ischemia are leading factors in sei-
POSTOPERATIVE SEIZURES AND
zure generation, total occlusion by embolization will
OUTCOME ASSESSMENT
significantly reduce the ictogenic potential of the AVM.
There are a variety of postoperative seizures. They may In microsurgical removal of the AVM by bipolar
be a continuation of preoperative seizures, they may be coagulation and use of sucker a small corridor of work
de novo seizures in previously seizure-free patients, around this lesion is created and a proportion of the struc-
and they may constitute a deterioration of pre-existent tural changes very close to the AVM will automatically
seizures because of higher frequency or worse seizure be included in the resected tissue volume. During micro-
type. To make assessment more complicated some surgery, hematomas and cortical zone overlying hemato-
patients have a single early postoperative seizure; for mas will additionally be removed. It is thus no surprise
example, in the San Francisco series 16% had a single that experience has demonstrated that simple AVM
postoperative seizure (Englot et al., 2012). removal can lead to quite satisfactory seizure-free rates.
It is difficult to interpret the available literature, since One has to assume that, once the AVM is gone, the icto-
outcome scales common in epilepsy surgery are not used genic factors originating from or produced by the AVM
by many authors; only modern series have used the Engel are so greatly reduced or have disappeared that they no
outcome scale, sometimes considerably modified longer irritate the surrounding cortex. Any treatment
(Przybylowski et al., 2015). Many authors have used modality that does not obliterate or remove the AVM
individualized outcome classifications. Unfortunately it completely would therefore be expected to lead to
is not always clear whether seizure freedom is meant less satisfactory seizure outcomes. Some results support
according to the Engel classification or to the ILAE clas- the concept that seizure-free rates likely improve when
sification. In Engel I seizure-free patients can still have the AVM has been completely occluded (Ghossoub
auras, whereas ILAE outcome class I is seizure free with- et al., 2001; Hoh et al., 2002; Lim et al., 2006; Hyun
out aura. Outcome scores using different cutoffs vary et al., 2012; Yang et al., 2012).
widely. Sometimes a difference is made between one
seizure versus more than two (Thorpe et al., 2000;
Early postoperative seizure versus de novo
Ghossoub et al., 2001; Sch€auble et al., 2004; Cao
epilepsy
et al., 2011) and others choose one to three seizures ver-
sus more than four (Piepgras et al., 1993; Yeh et al., 1993; Unfortunately patients who never experienced seizures
Lim et al., 2006). In older references one may find before AVM removal may have de novo seizures postop-
terms like “improved” or combined outcome descrip- eratively. One or two early postoperative seizures do not
tions like “seizure-free or significantly improved.” In constitute epilepsy. Occasional single seizures in the
one meta-analysis (Chen et al., 2014), the terms early postoperative period are not rare (Englot et al.,
“seizure-free status” and “seizure control” were used, 2012) and are not synonymous with persistent epilepsy.
with the latter including “seizure improvement.” Com- De novo seizures were seen in 6% of a large series
paring seizure outcomes is difficult to impossible. It is (Piepgras et al., 1993); additionally in 2%, seizure pat-
also not easy to extract different outcome data for the terns deteriorated, and there was a high rate of de novo
Another random document with
no related content on Scribd:
mitä perusteellisemmin hän tutki Schopenhaueria, intialaista
viisaustiedettä ja nykyaikaisen tieteen teorioja, joka pyrkii yhä
enemmän palauttamaan kaikkia ilmiöitä voimaan ja liikuntoon. Tämä
vakaumus sai viimeisen, täydellisen ja suuremmaisen ilmauksensa
Jälkisanassa teokseen Das Lebensabend einer Idealistin (Idelistin
elämänilta). Katsellessaan verratonta taulua, jonka tähtitaivas ja meri
levittivät hänen ikkunainsa eteen Nettunossa, seuratessaan hurjia
syysmyrskyjä, jotka näyttivät muuttavan maailmankaikkeuden
takaisin alkuperäiseksi kaaokseksi, hän oli kuulevinaan »Alkusyyn
äänen, Alkusyyn, jonka suuruuden edessä sielumme heittäytyy
maahan syvimmän hartauden vallassa ja autuaana riemuiten
saadessaan sanoa itselleen: se vaikuttaa minussakin.» Tämä
ikuisesti ainoa, luova alkusyy voi olla vain henki, ja meidän sielumme
on »erittynyt säde ikuista henkeä.» Ja harrasmielisen innostuksen
vallassa hän ylistää Jumalallista, Ikuista, »joka ei sinussa koskaan
pääse täydelliseksi», tuona hetkenä, jolloin hän huutaa maailmalle
kiitollisen jäähyväissanan, tuolle maailmalle, joka on ollut hänelle
»tutkivan hengen maailma».
Nämä eivät olleet sellaisia sanoja, joita voi kirjoittaa ilon ja

terveyden päivinä unohtaakseen ne myöhemmin kärsimyksen ja
tuskan hetkinä, kuoleman edessä; ne ilmaisivat syvintä vakaumusta,
ja kokemus oli osoittava, että Malwida von Meysenbug tästä
vakaumuksesta ammensi harvinaiset sielunvoimansa, jotka
elämässä pitivät häntä pystyssä ja antoivat hänelle sittemmin kyvyn
uhmata sairauden ja pitkällisen riutumuksen tuskia ei ainoastaan
alistuvana, vaan eräänlaisella iloisella sankaruudella.
Aina syksyyn 1901 saakka oli hänen terveytensä, joskin

ruumiinvoimat vähenivät, pysynyt ihmeen tyydyttävänä hänen
korkeaan ikäänsä nähden, ja hänen työkykynsä oli ennallaan. Siihen
aikaan julkaisi hän viimeisen teoksensa Individualitäten
(Yksilöllisyydet) Heinäkuusta lokakuuhun, jonka ajan hän vietti
Sorrentossa ja Amalfissa, nautti hän vielä suuresti Napolin lahden
kauneudesta. Mutta juuri siihen aikaan ilmaantuivat ensimäiset sen
taudin oireet, joka oli vievä hänet pois, syöväntapainen paise
sisälmyksissä. Onneksi hän ei koskaan saanut tarkempaa tietoa
kipunsa luonteesta. Palatessaan Roomaan tunsi hän olevansa liian
uupunut voidakseen aloittaa uusia töitä, ja hän tyytyi kokoamaan
ottotytärtään varten merkillisimmät runonsa ja muodostelemaan
uudestaan pitempää novellia Die Improvisatrice, jota hän piti
suuressa arvossa. Hän tunsi siksi heikentyneensä että pelkäsi
elonpäiviensä olevan luetut, ja hän lausui toivomuksen, että joku
jäsen perhettä, jonka hän oli ottanut lapsenasemaan, aina olisi
hänen luonaan. Muuten oli hänellä Roomassakin antautuvia ystäviä,
joiden läheisyys oli hänelle lohdutus ja jokapäiväinen ilo ja piti vireillä
hänen mielenkiintoaan elämää kohtaan. Niinpä eräs tuli usein
keskustelemaan hänen kanssaan opinnoistaan ja uskonnollisen
tunteensa muuttumisesta, sairas otti vilkkaasti osaa hänen
teoksensa valmistukseen ja keskusteli hänen kanssaan siitä vielä
muutamia päiviä ennen kuolemaansa. Eräs toinen ystävätär antoi
hänelle herkeämätöntä tukea kaikissa hänen taloutensa
yksityiskohdissa ja vapautti hänet kaikista arkihuolista. Kolmas auttoi
lääkäriä, joka kävi neiti von Meysenbugin luona, ja otti huolekseen
koko hoidon, jota hänen terveydentilansa vaati. Eräs ystävätär
vihdoin, joka osoitti suorastaan lapsellista kiintymystä häneen,
omistautui hänelle mitä uhrautuvimmalla antaumuksella, niin että
hänen viimeiset elinvuotensa eivät ainoastaan keventyneet ja
kirkastuneet sen kautta, vaan varmaan itse elinaika piteni tämän
jokapäiväisen, uskollisen hoidon tähden.
Niin kauan kuin mahdollista oli salasi hän kuitenkin mielellään
tilansa vakavuuden ympäristöltään. Koskaan hän ei ollut eloisampi ja
hilpeämpi puheissaan kuin keväällä 1902. Tahdon sitkeydellä, josta
hän oli antanut niin monta todistusta elämänsä kuluessa, jatkoi hän
toimintaansa totutulla tavalla kunnes tuli se päivä, jolloin voimat
jättivät hänet. Vaikkakin hänen alkaen tammikuusta 1902 täytyi
luopua rakkaista ajeluistaan, hoiti hän vielä kuten ennenkin ilman
vierasta apua aina erittäin huolitellun pukeutumisensa ja puuhasi
taloudessaan, seurasi tarkoin kaikkea, mikä koski politiikkaa,
kirjallisuutta ja taidetta, oli laajassa kirjevaihdossa ei ainoastaan
ystäviensä kanssa vaan myöskin tuntemattomien, jotka tahtoivat
häneltä neuvoa tai lohdutusta. Joka päivä omisti hän hetkisen aikaa
henkiselle työlle, toisen keskustelulle uskottujen ystävien kanssa.
Kesällä ei hänen enää ollut mahdollista palata rakkaimpaan

kesäpaikkaansa Sorrentoon. Hänen täytyi tyytyä merenrantaan
Nettunossa lähellä Roomaa, ja siellä oli hänen vielä kerran suotu
nauttia Italian taivaan ihanuudesta ja noiden hilpeitten ja
luonnonraikkaitten ihmislasten kauneudesta, jotka hänen ikkunainsa
edessä ilakoivat Välimeren sinisissä laineissa. Mutta taudin
edistyminen aiheutti jo jokapäiväisiä kipukohtauksia ja pakotti hänet
jatkuvaan ja tuskalliseen hoitoon. Niinpä olivat hänen ajatuksensa
enimmäkseen vakavia luonteeltaan, sillä hän tunsi voimien
vähenevän nopeasti, ja joka päivä askarteli hän filosoofisissa ja
moraalisissa omantunnon tutkimuksissaan ja valmistui valoisan, voi
sanoa iloisen alttiisti erkanemaan tästä näennäisyyksien elämästä ja
siirtymään ikuisen todellisuuden helmaan. Hänen ajatuksensa saivat
usein kuin itsestään runollisen muodon, ja hän on toimittanut ne
julkaistaviksi sarjassa enemmän tai vähemmän viimeisteltyjä runoja.
Erään useimmin toistuvista ajatuksista on hän selvimmin kiteyttänyt
näihin neljään riviin:
»So versöhnt nun mit den vielen Fehle
Der Erscheinungswelt, mit dem Geschick
Lass uns heimwärts ziehen, liebe Seele,
Nur das alle Kleid bleib' hier zurück!»
[Sovitettuina näennäisyyksien maailman monien virheiden,

sovitettuina kohtalon kanssa käykäämme kotia kohden, rakas
sielu, vanha vaate vain jääköön tänne!]
Lähtiessään Nettunosta hän kirjoitti toisen runon, jota hän vielä

tahtoi muodostella, mutta joka ensimäisessä alkuperäisessä
muodossaan kuvaa ihmeen kauniisti ja syvällisesti hänen
sieluntilaansa:
»Der Tag ist hin. In tiefem Frieden ruht

Die weite Wasserfläche unbewegt;
In Westen färbt den Himmel milde Glut
Und rosa Wölkchen ziehen um den Mond.
Verstummt sind all die hohen Stimmen just,

Die aus dem Wasser erst heraufgeschallt,
Von froher Jugend übermütiger Lust
Die da im Wasser Scherz und Frische fand.
Nur in dem Mondlicht schwebt noch geistergleich —

Ein weisses Segel, eine Seele scheint's,
Die losgebunden von dem irdischen Reich
In ihre Heimat hinzieht, die Unendlichkeit.
Es ist mein Bild. Ich liege hier allein,

Verstummt isi auch für mich der laute Tag
Ich fühl's, ich scheide bald!
Und bin ich fern, das Leben setzt sich for
Gang wie es war, in buntem Vielerlei.
Man war einmal nur.
Blieb nichts von mir auf dieser Erdenwelt

Doch in den Herzen, die mich treu geliebt,
Wie der Mondglanz jetzt im Wasser
So lebt mein Bild Erinnerung.
[Päivä on päättynyt. Syvässä rauhassa lepää

avara veden kalvo liikkumattomana;
lännen taivaan värittää vieno hehku ja
ruusuiset pilvet liitelevät kuun ympärillä.
Vaienneet ovat kaikki äänekkäät huudot,

jotka äsken vedestä kajahuttivat
iloisten nuorten vallatonta kisailua,
kun he aalloista ammensivat hilpeyttä ja raikkautta.
Vain kuutamossa välkkyy vielä aavemaisena

valkoinen purje, se on kuin sielu,
joka irtaantuneena maan valtakunnasta
liitää kohti kotiansa, äärettömyyttä.
Se on minun kuvani. Lepään täällä yksin,

vaiennut minullekin on äänekäs päivä.
Minä tunnen että pian erkanen!
Ja kun sitten olen kaukana, jatkuu elämä aivan kuin ennen,

kirjavan moninaisena. Ihminen oli vain kerran.
Mitään ei jäänyt minusta tähän maailmaan, mutta
sydämissä, jotka rakastivat minua uskollisesti, elää kuvani
muisto kuin kuun välke nyt vedessä.]
Tuskin oli hän palannut takaisin Roomaan, kun tauti alkoi

ratkaisevasti edistyä. Vain muutamia kuukausia voisi hän enää elää,
arvelivat lääkärit; hän itse tosin tunsi käyvänsä yhä heikommaksi ja
heikommaksi, mutta luuli olevansa tekemisissä vain äkillisen taudin
kanssa, joka kiirehti lopputaistelua, mutta mahdollisesti oli
voitettavissa tahdonlujuudella. Tämä harvinainen henkinen voima se
varmaan antoi uupuneelle ruumiille kykyä tehdä vastarintaa pitkälle
ohi lääketieteen asettaman määräajan. En voi tässä kajota hänen
viimeisten elinkuukausiensa yksityiskohtiin, ajan, jolloin hänen
kasvattityttärelleen ja tämän puolisolle ja tyttärelle oli onneksi sallittu
viipyä alituisesti hänen vierellään. Sairaus oli yltynyt pelottavassa
määrässä, tuskat kävivät hyvinkin kiivaiksi kohtausten sattuessa yhä
tiheämmin ja saivat lievikettä vain morfiiniruiskutuksista,
ravinnonsaanti vaikeutui ja jatkuva huolehtiminen puhtaanapidosta,
joka mahtoi olla kaksinverroin tuskallista niin hienotunteiselle ja
aikaisemmin niin kokonaan itsenäiselle sairaalle, vaati päivittäin yhä
kasvavan määrän tunteja. Ja kuitenkaan hän ei tahtonut vielä itse
pitää tilaansa toivottomana ja oli toimessa mikäli suinkin voi; joka
päivä otti hän vastaan ainakin yhden ystävän ja kirjoitti jonkun
kirjeen, seurasipa vielä tarkoin valtiollisia uutisia. Hän luki vielä aika
ajoin, joko nuoren ystävänsä Romain Rollandin Beethovenia tai de
Wetin Sotaretkiä, sankarillisen buuripäällikön, jonka urotöitä hän oli
seurannut innostuneen myötätuntoisesti, hän pukeutui edelleen omin
neuvoin, kunnes useita kertoja sattuneet kaatumiset estivät
heikentynyttä vanhusta tästäkin. Tammikuussa 1903 onnistui hänen,
vielä aamuisin kirjoittaa jälkisanaa teokseen Lebensabend vapisevin
käsin, jo niin voimatonna, että hän päivittäin sai piirretyksi vain viisi
tai kuusi riviä. Kaiken tämän ohella ei hän valittanut koskaan, hän
selitti olevansa onnellinen saadessaan, kiitos sairautensa, nähdä
meidät kokoontuneina ja tunnusti ettei hänellä ollut mitään oikeutta
nurkua kohtaloa vastaan, joka oli sallinut hänen suorittaa loppuun
hyvän osan kaikesta siitä mitä hän kerran oli uneksinut
suorittavansa. Hän moitti meitä ystävällisesti siitä että muka liiaksi
antauduimme hänen ruumiilliseen hoitoonsa emmekä kyllin usein
seurustelleet hänen kanssaan hengen korkeammissa ilmakehissä.
Huhtikuun alussa sanoi hän minulle: »Minä uskon olevani hyvä

profeetta; tämän kuun loppuessa loppuu minunkin sairauteni, ja jos
silloin voimani riittävät, voin ehkä taas nousta, mutta ne eivät riitä!»
Hänen oli täytynyt luopua melkein kaikesta lukemisesta kirjeiden
kirjoittamisesta, ja hänen koko henkinen elämänsä rajoittui pariin
tuttavallisen keskustelun hetkeen, jotka hän vielä joka päivä saattoi
myöntää eräälle läheiselle ystävälleen. Kuitenkin oli hänelle näinä
viimeisinä aikoina suotu vielä yksi suuri ilo; hän oli antanut
ottotyttärensä pojanpojan äiteineen tulla Roomaan, tuon viehättävän
nelivuotiaan lapsen läheisyys virkisti suuresti hänen sydäntään. Ei
voi kuvitella mitään ihastuttavampaa kuin nähdä vanhuksen
lepertelevän, leikkivän ja nauravan hänen kanssaan ja opettavan
häntä vapisevin käsin muodostamaan pieniä esineitä ja eläimiä
savesta.
Vasta kahdeksan päivää ennen kuolemaansa suostui hän

alituisesti pysymään vuoteessa ja salli sairaanhoitajattaren valvoa
vierellään öisin sekä luopui kokonaan taloushuolista. Siitä päivästä
alkaen ei hän enää nauttinut mitään ravintoa ja puhui alituisesti
lähestyvästä kuolemastaan. Tiistaina 21 p. huhtikuuta, niin sanottuna
Nascita di Roma päivänä, ikuisen kaupunginperustamisen
vuosipäivänä perimätiedon mukaan, oli ystävätär, josta ylempänä
olen puhunut, antanut erään etevän taitelijan tulla soittamaan hänelle
viululla hänen lempilaulujaan. Hän ei kuitenkaan
hermokiihtymykseltä voinut kestää tätä soittoa ja pyysi meitä kaikkia,
myöskin hoitajatarta ja palvelijoita, kokoontumaan vuoteensa
ympärille. Sitten hän käski poikasen istahtaa vuoteelleen, ja laskien
kätensä hänen päänsä päälle hän sanoi. »Minä siunaan sinua,
enkelini, sinä viet kaikille muille minun siunaukseni.» Ja senjälkeen
hän puhui kanssamme kolmen tunnin ajan, kello yhdeksästä
puolipäivään, iloisen haltioitumisen vallassa rakkaudestaan meihin,
suurista filosoofisista toiveista, joiden elähyttämänä hän oli astuva
kuoleman tuntemattomaan maahan. Hän uskoi ja toivoi että hänen
viimeinen hetkensä oli lyönyt ja odotti saavansa keskellä tätä
puhelua eritä elämästä meidän käsivarsillamme. Nyt oli, sanoi hän,
Rooman syntymähetki ja päivä, se oli myös oleva hänen uuden
syntymänsä päivä, sillä elämä ja kuolema ovat yhtä. Jokaiselle
meistä hän puhui hellän rakkauden sanoja, kullekin luontonsa
mukaan, ja kiitti saamastaan ystävyydestä ja moninaisesta
huolenpidosta ympärillään sanoen olevansa onnellinen tässä
asemassa, keskellä kaikkia rakkaitansa. Hän ei voinut kenellekään
toivottaa ihanampaa loppua kuin hänen oli. Ja kuitenkin keskeyttivät
tuskalliset suonenvetokohtaukset häntä melkein joka hetki, mutta
hän sanoi vain: »Älkää pelästykö, eihän se mitään ole, se on vain
suonenvetoa ja minä en kärsi, ruumisparkani vain. — Aine ei ole
mitään, henki yksin pysyy.»
Ja tämä sana henki tuli yhä uudelleen hänen huulilleen samoin

kuin sanat Amore ja Pace, joita hän alati toisteli. Sitten hän, vielä
puhui'. »Te seuraatte minua Paratiisin portille saakka, mutta portti
suljetaan edessänne ja teille sanotaan. Palatkaa takaisin maan
päälle, te ette ole vielä valmiit, teidän on vielä vaikutettava ja tehtävä
työtä! Mutta minä olen valmis, minä en voi enää tehdä työtä», lisäsi
hän huoaten. Hän sanoi vielä: »L'amore ju il principio di tutto. L'ho
sempre pensato, ma oggi lo so meglio che mai.» Omituista oli että
keskellä näitä jäähyväiskohtauksia ja huolimaita kyynelistä, joita
ympärilläolijat vuodattivat, hän itse ei itkenyt, vaan hänen silmissään
loisti kuin mystillinen ilo. Vain kerran kihosi kyynel hänen silmäänsä,
kun hän puhui päivästä kolmekymmentä vuotta sitten, jolloin hän oli
sanonut kasvattityttärelleen hyvästit tämän viettäessä häitään.
Tämä puhe oli oleva hänen viimeisensä. Seuraavina päivinä salli

yhä lisääntyvä heikkous hänen vain silloin tällöin lausua kiitoksen tai
rakkauden sanan ympäröiville henkilöille. Hän tahtoi sanoa hyvästit
kaikille ystävilleen Roomassa. Hänen henkensä pysyi
häiriintymättömänä, hänen mielensä oli suuntautunut elämän suuriin
kysymyksiin, vielä torstaina hän sanoi eräälle lääkärilleen, jonka näki
viimeistä kertaa: »Te olette todellinen Italian poika, tehkää työtä
Italian hyväksi!» Viimeinen sana, jonka kuulin hänen huuliltaan iltaa
ennen hänen kuolemaansa, oli sana Amore.
Lauantai-iltapäivällä hän ei enää kyennyt puhumaan, lauantain ja

sunnuntain välisenä yönä oli hänen heikkoutensa niin suuri että me
kokoonnuimme hänen vuoteensa ympärille odottaen hänen viimeistä
hengenvetoaan. Ei ollut enää kärsimyksiä, ei viimeistä kamppailua,
ei kuoleman korinaa. Kello kaksi iltapäivällä huhtikuun 26 p. hän
sammui hiljaa, ilman että mikään muu ulkonainen merkki kuin
kasvojen kalmankalpeus ilmaisi elämän loppuneen ja sen toiveen
vihdoinkin täyttyneen, jonka hän lausui eräässä viimeisiä
kirjoittamiaan rivejä:
»Sie ruht im Vollbesitze ewigen Lebens»
[Hän lepää iäisen elämän täydessä omistuksessa.]

Hänen toivomuksensa mukaan poltettiin Malwida von
Meysenbugin maalliset jäännökset Campo Veranolla. Hänen
tuhkansa säilytetään uurnassa, joka on laadittu hänen
ottolapsenlapsensa Edouard Monod'n muovaileman maljan mukaan,
Monte Testaccion hautausmaalla aivan lähellä Goethen pojan
hautaa.
Hän ei ollut tahtonut mitään kirkollisia menoja eikä puheita

haudallaan. »Minä en enää kuulu», on hän kirjoittanut, »mihinkään
oikeauskoiseen kirkkoon, vaan niiden ihmisten suureen
seurakuntaan, jotka rakastavat hyvää, korkeata ja kaunista ja
pyrkivät toteuttamaan sitä itsessään ja ympärillään.»
1. AIKAISIMMAT MUISTOT.
Suuren kaupungin keskellä olisi vaikea löytää paremmalle paikalle

rakennettua taloa kuin se, jossa synnyin ja vietin lapsuuteni
ensimäiset päivät. Talo sijaitsi Casselin kaupungissa, rinnan muiden
rakennusten kanssa ja rajoitti katua, jolle oikeudenmukaisesti oli
annettu nimi Bellevue, sillä vastakkaisella sivulla ei ollut ainoatakaan
rakennusta, vaan ihana näköala kauniita puisto- ja puutarha-
istutuksia päin. Nämä kohosivat terrassinmuotoisesti hedelmällisellä
tasankomaalla, jonka läpi virtaa suurehko Fulda-joki. Olin lähinnä
nuorin kymmenestä lapsesta — kaikki terveitä ja lahjakkaita.
Vanhempani olivat vielä nuoria minun tullessani maailmaan. Heidän
elämänsä oli onnellista keskitietä ylellisyyden ja puutteen välillä —
asema, joka takaa suurimmat koti-onnen edellytykset. Olen
ensimäisistä lapsuuspäivistäni säilyttänyt loppumattoman hilpeyden
valoisan hohteen. Tuosta valoisasta taustasta eroittuu vain kolme
määrätympää muistoa.
Ensimäinen muisto on äitini huone maalattuine seinäpapereineen,

joille oli kuvattu palmumaisemia, kaislametsiä ja oudonmuotoisia
rakennuksia. Lapsellinen mielikuvitukseni iloitsi tuosta
tarunomaisesta maailmasta. Sitäpaitsi kertoi muuan talon ystävä
minulle siitä satuja; niinpä oli eräs noista pikku taloista Blumenbach
nimisen velhon asunto, jolle koko luonto oli kuuliainen. Talon
vieressä seisoi suuri haikara pitkillä, jäykillä jaloillaan,
pitkänokkainen pää rinnalle vaipuneena. Se oli Blumenbachin
palvelija, kuten ystäväni sanoi. Hän seisoi muka siinä odottamassa
herransa käskyjä.
Toinen muisto liittyy iltaan, jolloin hoitajattareni kertoi, että pieni,

äskensyntynyt siskoni, oli kuollut. Hän antoi minun vasten äidin
kieltoa katsoa lasioven läpi huoneeseen, jossa seisoi musta arkku;
arkussa lepäsi pieni sisareni, valkeana kuin lumi ja kukkien peitossa.
Kolmas muisto vihdoin punoutuu vanhan ruhtinaan persoonaan,

joka oli hallitsijana Hessenin vaaliruhtinaskunnassa, pienessä
saksalaisessa kotivaltiossani. Hänen vaununsa ajoivat joka päivä
talomme ohi. Vaunuissa istui vanha ukko Fredrik Suuren aikuisessa
univormussa kolmikolkkainen hattu päässä. Hänen valkoiset
hiuksensa olivat takaa palmikossa ja toista poskea rumensi hirveä
paise. Siihen sairauteen hän kuolikin. Minä en nähnyt hänen
hautajaisiaan, mutta vanha hoitajattareni kuvasi niitä minulle
lukemattomat kerrat. Häntä ei haudattu esi-isien kammioon suuressa
kirkossa. Hänet haudattiin toivomuksensa mukaan erään huvilinnan
kirkkoon. Linnan oli hän rakennuttanut ja se oli ollut hänen
mielipaikkansa. Hautaus tapahtui yöllä tulisoihtujen valossa, vanhan
tavan mukaan. Mustapukuisen ritarin oli mustalla hevosella
ratsastettava heti ruumisvaunujen jäljessä. Tämä ritari valittiin
korkeasta ylimystöstä, mutta hänen täytyi aina, tarun mukaan,
maksaa tämä kunnia hengellään. Tässäkään tapauksessa ei kansa
pettynyt luulossaan. Tuon öisen näytelmän silloinen esittäjä, nuori
aatelismies, terveyttä ja voimaa uhkuva, joutui kolme viikkoa
hautajaisten jälkeen kuumetaudin uhriksi. Johtuiko tuo kuumetauti
yksinkertaisesti vilustumisesta kylmässä rautapuvussa pitkän
yöllisen kulkueen aikana? Kansa ei ottanut huomioon tuollaista
mahdollisuutta, ja minun lapsellinen mielikuvitukseni piti kansan
uskoa oikeana. Joka kerta tunsin salaista väristystä, kun
vanhempieni kanssa kävin huvilinnassa ja näin asehuoneessa
mustan puuhevosen selässä mustan varustuksen, jota onneton ritari
oli tuona yönä kantanut.
2. YLEISTÄ JA YKSITYISTÄ.
Vanhan herran kuolema ei aiheuttanut suuria muutoksia ainoastaan

minun perheessäni, vaan lopetti, niin sanoakseni, kokonaisen
ajanjakson pienen isänmaani historiassa. Hallitsijahuone, johon hän
oli kuulunut, oli hyvin vanha ja laski esi-isiensä joukkoon jäseniä,
jotka olivat kuuluja urhoollisuudestaan ja jalomielisyydestään. Mutta
viimeiset sukupolvet olivat huonontuneet. He olivat häpeällisillä
keinoilla kartuttaneet yksityisomaisuuttaan, myymällä alamaisiaan
vieraille valloille, jotka käyttivät niitä kaukaisissa sodissa.
Rakastajattaret olivat jo pitkät ajat hallinneet maata. Viimeinen

hallitus oli ollut merkillisen vaihteleva. Vanhan ruhtinaan luonteessa
oli kaikista virheistä huolimatta ollut kunniallisuutta. Vuosisatamme
kaikkivoivan valloittajan lähestyessä jätti hän vapaaehtoisesti
ruhtinaskuntansa, tietäen hyvin olevansa kykenemätön sitä
puolustamaan ja tahtoen säästää alamaisiltaan turhaa
verenvuodatusta. Hän piti parempana maanpakoa kuin häpeällistä
antautumista muiden saksalaisten ruhtinasten tapaan. Hänen pienen
valtakuntansa pääkaupunki joutui taaskin samaan asemaan
suuressa kuningaskunnassa, jonka valloittaja lahjoitti nuorimmalle
veljelleen Jérômelle. Ranskalainen hienostus, siekailemattomuus ja
herkkäuskoisuus pääsi valtaan kankipalmikkoväen autioissa
asunnoissa. Nuori kuningas loi pienen Pariisin saksalaiselle
maaperälle.
Isäni, joka ei voinut viedä perhettään maanpakoon, jäi paikoilleen

astuen uuden valtion palvelukseen. Äitini oli aivan nuori ja hyvin
kaunis; oli siis luonnollista, että molemmat ottivat osaa nuoren hovin
vilkkaaseen, hilpeään elämään. Kuinka usein kiusasinkaan äitiäni
yhä uudelleen kertomaan minulle tarinoita noilta päiviltä, jotka olivat
päättyneet aikoja ennen syntymääni! Kuinka ahnaasti kuuntelinkaan
kuvauksia loistavista juhlista ja viehättävistä naisista, jotka olivat
perheineen saapuneet Pariisista kaunistamaan suloillaan ritarillisen
kuninkaan hovia! Millä mielenkiinnolla tutkinkaan äitini vaatesäiliötä,
jossa vielä oli monta muistoa noilta ajoilta! Miten minua
viehättävätkään nuo paimenpuvut, turkkilaiset viitat ja antiikit
poimukkeet! Monet mietteet heräsivät minussa, kun ajattelin kaiken
tämän ihanuuden ennenaikaista katoamista. Kuin uni oli koko loisto
haihtunut. Venäläiset olivat ilmestyneet kaupungin porteille, heidän
luotinsa olivat vinkuneet kaduilla. Vanhempani olivat panneet säilöön
parhaat tavaransa ja jättäneet talon, joka oli liian suojaton vihollisen
hyökkäyksille. Vanha, luonamme asuva täti, oli kätkenyt useita
esineitä aitan jauho-astioihin. Kanuunankuula osui taloon ja jäi
muuriin.
Kuullessani näitä kertomuksia, pahoittelin koko sydämestäni, että

en ollut jo silloin elänyt ja saanut jakaa vaaraa vanhempieni kanssa.
Olisin myöskin hartaasti halunnut tietää, olisivatko kasakat löytäneet
jauhoihin kätketyt esineet, jos kaupunki olisi joutunut ryöstölle
alttiiksi.
Mutta kaupunki, kuten vanhempanikin, varjeltui perikadosta.

Vanha ruhtinas palasi takaisin maahansa. Kankipalmikot ja
korpraalinsauvat, jotka ranskalaiset sulottaret olivat karkoittaneet,
pääsivät jälleen käytäntöön. Maata alkoi jälleen hallita epämiellyttävä
rakastajatar, kivulloisen ukon hoitaja. Suosikit, jotka jakaessaan
maanpaon ruhtinaansa kanssa, olivat rikastuneet, saivat valtion
ensimäiset virat. Kansa oli uskollisena ruhtinashuoneelleen, kuten eri
Saksan heimot yleensä, alussa tervehtinyt iloisella innostuksella
palaavaa ruhtinastaan; pian kuitenkin kävi ilmi, että nykyisyyden ja
menneisyyden välinen side oli katkennut. Ruhtinaat ja kansat olivat
käsittäneet aivan eri tavalla kansallisen nousun ja vapaustaistelut.
Monen jalon sydämen haltioituneet unelmat häipyivät tyhjiin ja
vapauden aamuruskon asemasta, jota saksalainen nuoriso oli
toivonut, nousi uusi, synkkä, sumuinen päivä. Vanhan ajan ihmiset
pitivät välinäytöstä täydellisen yksinvallan suuressa huvinäytelmässä
päättyneenä ja ottivat uudelleen haltuunsa vanhat valtaistuimet ja
entisen herruuden. Kansojen veri oli vuotanut turhaan. Historia oli
jälleen pysähtynyt kulussaan.
Mutta kuolema ei ollut pysähtynyt; se kävi noutamassa vanhuksen

kankipalmikkoineen, ja siitä silmänräpäyksestä alkaen vapautuivat
kaikkien hiukset maassa tuosta menneisyyden kahleesta.
Ulkonaisesti tapahtui paljon muutoksia uuden hallituksen aikana.

Isäni, joka oli lapsena ollut kruununprinssin leikkitoveri, kutsuttiin
nykyisen ruhtinaan läheisyyteen ja sai huomattavan aseman
valtiossa. Me jätimme, talon, josta olen maininnut kertomukseni
alussa ja siirryimme suurempaan ja loistavampaan asuntoon aivan
ruhtinaallisen linnan vieressä.
3. PERHE-ELÄMÄMME.
Kunnioituksen tunne oli minussa varmaan synnynnäinen, sillä minua

ei koskaan pakotettu minkäänlaiseen palvontaan ja kuitenkin
jumaloin vanhempiani ja vanhinta sisartani. Hän oli jo täysikasvanut,
kun itse vielä olin pieni tyttö. Vertaan häntä muistissani
muinaissaksalaisten maalarien madonniin, noihin naisellisen
kauneuden perikuviin — kauneuden, joka enemmän ilmenee
taivaallisena puhtautena ja nöyryytenä kuin piirteiden
säännöllisyytenä. Ihailin tätä sisartani siinä määrin, että äitini oli
jonkun verran mustasukkainen. Se oli lapsuuteni ensimäinen
ristiriita; se päättyi kun sisareni meni naimisiin ja lähti puolisonsa
mukana tämän kotiseudulle. Tämä ensimäinen ero maksoi minulle
monta kyyneltä.
Rakkauteni äitiin heräsi kuitenkin sen jälkeen täydessä

voimassaan. Muistan vielä, millä ihastuksella häntä katselin, kun
hän, yhä vieläkin kauniina, koristi itseään jotakin juhlaa varten,
varsinkin hovitanssiaisiin. Asetuin sitten jonkin pimeän huoneen
ikkunaan, josta saatoin nähdä ruhtinaallisen linnan valoa säteilevät
ikkunat. Odotin, kunnes hän oli astunut valaistuihin loistosaleihin ja
hiukan raottanut joitakin ikkunaverhoja, niin että voin katsella sisään.
Näin komeasti puettuja naisia ja salin kummallakin seinustalla
herroja kullalla kirjailluissa univormuissa. Näin ruhtinaan perheineen
astuvan sisään ja kulkevan pitkin ihmisrivejä lausuakseen jokaiselle
pari sanaa. Näitä sanoja näyttiin pitävän hyvin tärkeinä, ja taas
päinvastoin, jos jokin henkilö sivuutettiin puhuttelematta, oli se
nöyryytyksen huippu. Kuinka ylpeä olinkaan nähdessäni, että äitini
kanssa keskusteltiin kauemmin kuin muiden! Uskoin varmasti, että
se oli suuri kunnianosoitus. Eikö ruhtinas ollutkin muita ylevämpi
olento? Minkätähden hän muuten olisi ollut ruhtinas? Olin Tuhannen
yhden yön tarinoissa lukenut niin paljon Harun al Raschidin jalosta,
suuremmoisesta luonteesta ja pääni oli niin täpötäynnä kertomuksia
keisari Friedrich Punaparran ritarillisista hyveistä, hänen, joka istuu
Kyffhäuserissa odottaen silmänräpäystä, jolloin uudelleen saa
perustaa ihanan saksalaisen valtakunnan, etten epäillyt ruhtinaiden
ylevämielisyyttä.
Isäni, jolla oli ylenpalttisesti työtä, ei voinut uhrata paljon aikaa

lapsilleen; jos hän joskus ehti olla parissamme, oli se oikea
juhlahetki, sillä rehellisempää, rakastettavampaa, kultaisempaa
luonnetta kuin hänen olisi ollut vaikea löytää.
Äitini puolestaan yritti herättää meissä eloon taiteellisia

taipumuksia. Hän kuului ajatussuuntansa puolesta tuohon ajan
henkiseen keskukseen, johon lukeutuivat Humboldtit, Rahel,
Schleiermacher, Schlegelit ja muut senaikuiset kuuluisuudet. Tässä
samalla kertaa vapaamielisessä, isänmaallisessa ja filosofisessa
suunnassa oli myöskin omituinen sekoitus mystisismiä, jota silloin
parhaillaan kukoistava romanttinen koulu osaltaan lisäsi. Äitini
riippumattoman luonteen johti tämä, suunta usein vastustamaan sen
seurapiirin sopivaisuuskäsitteitä, johon hänen isäni aseman vuoksi
oli pakko kuulua. Tämä koski erityisesti niiden henkilöiden
valitsemista, jotka muodostivat hänen läheisimmän piirinsä. Sen
sijaan, että yksinomaan olisi seurustellut ylimystön piireissä, valitsi
hän läheisimmät ystävänsä paljon useammin hengen ja sydämen
ominaisuuksien mukaan, välittämättä heidän yhteiskunnallisesta
arvo-asteestaan. Erittäin mielellään etsi hän etevimpien
näyttelijöiden seuraa, joiden esitykset ihastuttivat häntä ja tuottivat
hänelle nautintoa ja jotka hän asetti täysin samalle tasolle kuin muut
vieraansa. Tämä oli siihen aikaan vielä hyvin rohkeata, sillä
teatterinjäseniä pidettiin halpa-arvoisena kastina, korkeintaan kyllin
hyvänä varjelemaan muita kuolevaisia ikävystymiseltä, mutta ei
suinkaan oikeutettuna pyrkimään heidän rinnalleen. Äitini saikin
tämän vuoksi moitteita, eikä edes isäni ollut tässä asiassa yhtä
mieltä hänen kanssaan. Hän saapui harvoin näihin äitini pieniin
seuroihin. Mutta vanhemmat veljeni ja sisareni, jotka kaikki
harrastivat taidetta, mikä mitäkin, olivat niissä mukana, ja erittäin
usein sai talossamme kuulla oivallisia soitto-esityksiä. Tällaisessa
henkisessä ja taiteellisessa keskuksessa vietin lapsuuteni. Meidän
perhe-elämässämme ei lapsia pidetty niin kokonaan erillään
aikuisten elämästä kuin esim. Englannissa. Äitini oli sitä mieltä, että
seurustelu etevien ihmisten kanssa voi ainoastaan vaikuttaa hyvää
lasten henkiseen kehitykseen ja kehittää vähitellen heidän
arvostelukykyään ja makuaan. Uskon, että hän oli täysin oikeassa ja
että tällainen huolenpito, mikäli se on mahdollista, on kasvatuksen
tärkeimpiä tekijöitä. Kreikkalaiset tiesivät sen hyvin, ja lyseet, joissa
heidän filosofinsa ja viisaansa seurustelivat nuorison kanssa, eivät
todennäköisesti vähäisessä määrässä vaikuttaneet siihen, että
heistä tuli kansa, joka hakee vielä tänään vertaistaan.
Me emme saaneet minkäänlaista n.s. uskonnollista kasvatusta. En

muista, kuka minulle ensin puhui Jumalasta ja opetti pienen
rukouksen. Meidän ei koskaan tarvinnut tuoda esille hurskauttamme
palvelijoiden tai vieraiden ihmisten aikana, kuten Englannissa on
tapana. Minä puolestani seurasin tietämättäni Kristuksen määräystä,
jonka mukaan on oltava yksin, jos tahdotaan rukoilla oikein. Joka ilta,
kun makasin vuoteessani eikä minulla ollut muuta uskottua kuin
tyynyni, toistin hiljaa itsekseni pienen rukoukseni totisen uskon
hartaudella. Ei kukaan tietänyt siitä mitään. Rukoilin näin: »Rakas
Jumala, olen pieni, tee sinä sydämeni puhtaaksi, ettei siellä asu
kukaan muu kuin rakas Jumala yksin.»
En kuitenkaan voinut sille mitään, että sydämessäni oli muitakin

rakkaita asukkaita. Keksin itse toisen rukouksen, jonka rukoilin joka
ilta salavihkaa ensimäisen jälkeen ja jossa pyysin Jumalaa
siunaamaan vanhempiani ja siskojani, opettajaani ja tämän perhettä
sekä lopuksi kaikkia hyviä ihmisiä. Kun olin täyttänyt tämän sydämen
velvollisuuden, oli omatuntoni rauhallinen ja nukuin vanhurskaan
unta.
Eräänä aamuna ennen päivänkoittoa heräsin outoon hälinään

äitini makuuhuoneessa, jossa nuorin sisareni ja minä myöskin
nukuimme. Oli talvi ja uuniin sytytettiin juuri tulta. Kuulin äitini itkevän
ja vanhan tädin, joka seisoi hänen vuoteensa vieressä, sanovan:
»Älä sure, lapsesi on nyt Jumalan luona.» — Ymmärsin heidän
puhuvan pienestä veljestäni, joka oli vasta muutaman kuukauden
vanha ja sairastanut useampia päiviä. Minä itkin myöskin hiljaa
tyynyyni antamatta heidän huomata, että olin herännyt ja kuullut
sanat, joissa näytti piilevän ylevä, minun käsityskantani yli menevä
salaisuus. Kun minun oli aika nousta, kerrottiin, että pieni veljeni oli
kuollut. Olisin mielelläni tahtonut tietää enemmän kuolemaa ja
yhtymistä Jumalaan koskevasta salaisuudesta, mutta en uskaltanut
kysyä, kun pelkäsin lisääväni toisten murhetta.

Full Download Arteriovenous and Cavernous Malformations 1st Edition Edition Robert F. Spetzler PDF

Uploaded by

Copyright:

Available Formats

Full Download Arteriovenous and Cavernous Malformations 1st Edition Edition Robert F. Spetzler PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Full Download Arteriovenous and Cavernous Malformations 1st Edition Edition Robert F. Spetzler PDF

Uploaded by

Copyright:

Available Formats

Download More ebooks [PDF]. Format PDF ebook download PDF KINDLE.

Full download ebooks at ebookmass.com

Arteriovenous and Cavernous

Download More ebooks from https://ebookmass.com

Revision Lumbar Spine Surgery 1st Edition Robert F.

Algebra and Trigonometry, 7th Global Edition Robert F.

Operations and Supply Chain Management: The Core 6th

Algebra and Trigonometry, 4e - Instructor’s Solutions

Pocket Atlas Of Emergency Ultrasound 2nd Edition

Beyond Audit: Auditing Remotely and Delivering Value

Operations and Supply Chain Management, 17e 17th

Administración de operaciones: producción y cadena de

MICHAEL J. AMINOFF, FRANÇOIS BOLLER, AND DICK F. SWAAB

© 2017 Elsevier B.V. All rights reserved.

British Library Cataloguing-in-Publication Data

Library of Congress Cataloging-in-Publication Data

For information on all Elsevier publications

Publisher: Nikki Levy

Typeset by SPi Global, India

I.J. Abecassis W. Brinjikji

A.A. Abla P.A. Brown

J.R. Adler A. Can

S. Ambekar M.L. Cohen

G.S. Atwal E.M. Cox

D.L. Barrow D. Ding

S.C. Bir R.L. Dodd

H.E. Goldstein G. Lanzino

R.C. Heros P. McCormick

A. Nanda R.F. Spetzler

M. Narayanan R.M. Starke

A. Ozpinar A.H. Turkmani

M.R. Reynolds D.D. Wang

Epidemiology, genetics, pathophysiology, and prognostic

ALP OZPINAR1*, GUSTAVO MENDEZ2, AND ADIB A. ABLA3

INTRODUCTION presentations and symptomatology, as well as a wide var-

better stratify patients between surgical and endovascular

The natural history of cerebral arteriovenous malformations

(Waltimo, 1973; Pasqualin et al., 1985; Kader et al.,

Arteriovenous malformations: epidemiology, clinical

JOSHUA W. OSBUN, MATTHEW R. REYNOLDS, AND DANIEL L. BARROW*

INTRODUCTION can vary greatly from patient to patient in terms of angio-

Seizures associated with cerebral arteriovenous malformations

Nämä eivät olleet sellaisia sanoja, joita voi kirjoittaa ilon ja

Aina syksyyn 1901 saakka oli hänen terveytensä, joskin

Kesällä ei hänen enää ollut mahdollista palata rakkaimpaan

[Sovitettuina näennäisyyksien maailman monien virheiden,

Lähtiessään Nettunosta hän kirjoitti toisen runon, jota hän vielä

»Der Tag ist hin. In tiefem Frieden ruht

Verstummt sind all die hohen Stimmen just,

Nur in dem Mondlicht schwebt noch geistergleich —

Es ist mein Bild. Ich liege hier allein,

Blieb nichts von mir auf dieser Erdenwelt

[Päivä on päättynyt. Syvässä rauhassa lepää

Vaienneet ovat kaikki äänekkäät huudot,

Vain kuutamossa välkkyy vielä aavemaisena

Se on minun kuvani. Lepään täällä yksin,

Ja kun sitten olen kaukana, jatkuu elämä aivan kuin ennen,

Tuskin oli hän palannut takaisin Roomaan, kun tauti alkoi

Huhtikuun alussa sanoi hän minulle: »Minä uskon olevani hyvä