Sei sulla pagina 1di 353

Neurosurgical Operative Atlas

Second Edition

Pediatric Neurosurgery

American Association of Neurosurgeons Rolling Meadows, Illinois

Neurosurgical Operative Atlas

Second Edition

Pediatric Neurosurgery

James Tait Goodrich, MD, PhD, DSci (Honoris Causa)

Professor of Clinical Neurological Surgery, Pediatrics, Plastic and Reconstructive Surgery
Albert Einstein College of Medicine
Director, Division of Pediatric Neurosurgery
Center for Craniofacial Disorders
Childrens Hospital at Monteore
Bronx, New York

New York Stuttgart
American Association of Neurosurgeons
Rolling Meadows, Illinois

Thieme Medical Publishers, Inc.

333 Seventh Ave.
New York, NY 10001

American Association of Neurosurgeons (AANS)*

5550 Meadowbrook Drive
Rolling Meadows, Illinois 60008-3852

*The acronym AANS refers to both the American Association of Neurological Surgeons and the American Association of Neurosurgeons.
Associate Editor: Birgitta Brandenburg
Assistant Editor: Ivy Ip
Vice President, Production and Electronic Publishing: Anne T. Vinnicombe
Production Editor: Print Matters, Inc.
Vice President, International Marketing and Sales: Cornelia Schulze
Chief Financial Ofcer: Peter van Woerden
President: Brian D. Scanlan
Cover illustration: Anita Impagliazzo
Compositor: Compset, Inc.
Printer: Everbest Printing Company
Library of Congress Cataloging-in-Publication Data
Neurosurgical operative atlas. Pediatric neurosurgery / [edited by] James Tait Goodrich.
p. ; cm.
Includes bibliographical references and index.
ISBN 978-1-58890-510-9 (alk. paper)
1. Nervous systemSurgeryAtlases. 2. ChildrenSurgeryAtlases. 3. Pediatric neurologyAtlases. I. Goodrich, James T.
[DNLM: 1. Nervous System DiseasessurgeryAtlases. 2. Child. 3. Infant. 4. Neurosurgical ProceduresmethodsAtlases.
WL 17 P371 2008]
RD593.P3822 2008
Copyright 2008 by Thieme Medical Publishers, Inc., and the American Association of Neurosurgeons (AANS). This book, including all
parts thereof, is legally protected by copyright. Any use, exploitation, or commercialization outside the narrow limits set by copyright
legislation without the publishers consent is illegal and liable to prosecution. This applies in particular to photostat reproduction,
copying, mimeographing or duplication of any kind, translating, preparation of microlms, and electronic data processing and storage.
Important note: Medical knowledge is ever-changing. As new research and clinical experience broaden our knowledge, changes in
treatment and drug therapy may be required. The authors and editors of the material herein have consulted sources believed to be
reliable in their efforts to provide information that is complete and in accord with the standards accepted at the time of publication.
However, in view of the possibility of human error by the authors, editors, or publisher of the work herein or changes in medical
knowledge, neither the authors, editors, or publisher, nor any other party who has been involved in the preparation of this work,
warrants that the information contained herein is in every respect accurate or complete, and they are not responsible for any errors or
omissions or for the results obtained from use of such information. Readers are encouraged to conrm the information contained herein
with other sources. For example, readers are advised to check the product information sheet included in the package of each drug they
plan to administer to be certain that the information contained in this publication is accurate and that changes have not been made in
the recommended dose or in the contraindications for administration. This recommendation is of particular importance in connection
with new or infrequently used drugs.
Some of the product names, patents, and registered designs referred to in this book are in fact registered trademarks or proprietary
names even though specic reference to this fact is not always made in the text. Therefore, the appearance of a name without designation
as proprietary is not to be construed as a representation by the publisher that it is in the public domain.
Printed in China
ISBN 978-1-58890-510-9

To Setti S. Rengachary, MD, and Robert H. Wilkins, MD

When the rst edition of this remarkable atlas came out in early 1990s, it was an instant and powerful success in the
neurosurgical literature. Drs. Rengachary and Wilkins recognized the power of the illustrated text and in this case designed
a work that was clearly visual with the text being secondary. As neurosurgeons are visual animals, this was a successful
design. Drs. Rengachary and Wilkins contributions to neurosurgery have been enormous. It is a true pleasure to help bring
this remarkable atlas back in a second edition. However, it must be remembered that their original editorship led to this most
elegant and inuential series of operative chapters. In acknowledgment of their insights, their educational leadership, and,
most importantly, their longstanding contributions to neurosurgery, I dedicate this work to these two scholars and superb


Continuing Medical Education Credit Information and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

Continuing Medical Education Disclosure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
Series Foreword Robert Maciunas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xiii
Foreword Richard D. Hayward . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xix
Chapter 1

Fibrous Dysplasia Involving the Craniofacial Skeleton. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

James Tait Goodrich

Chapter 2

Chiari Malformations and Syringohydromyelia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Elizabeth C. Tyler-Kabara and W. Jerry Oakes

Chapter 3

Unilateral and Bicoronal Craniosynostosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Kant Y. K. Lin, John A. Jane Jr., and John A. Jane Sr.

Chapter 4

Transoral Surgery for Craniovertebral Junction Abnormalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Arnold H. Menezes

Chapter 5

Malposition of the Orbits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

John A. Persing and Bianca I. Knoll

Chapter 6

Ventriculoatrial Shunting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
John Drygas and Stephen J. Haines

Chapter 7

Repair of Growing Skull Fracture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

Tadanori Tomita

Chapter 8

Occipital Encephaloceles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
William O. Bell

Chapter 9

Surgical Management of Pansynostosis (Craniosynostosis) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

James Tait Goodrich and David L. Staffenberg

Chapter 10

Tethered Spinal Cord, Intramedullary Spinal Lipomas, and Lipomyelomeningoceles. . . . . . . . . . . . . . . . . . . . . . . . . 62

Elizabeth C. Tyler-Kabara and W. Jerry Oakes

Chapter 11

Encephaloceles of the Anterior Cranial Base. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

Jonathan P. Miller and Alan R. Cohen

Chapter 12

Exorbitism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Constance M. Barone, David F. Jimenez, and James Tait Goodrich

Chapter 13

Depressed Skull Fracture in Infants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

Marion L. Walker

Chapter 14

Orbital Hypertelorism and Orbital Dystopia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Constance M. Barone, David F. Jimenez, and James Tait Goodrich


viii Contents
Chapter 15

Closure of the Myelomeningocele . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

David G. McLone

Chapter 16

Dandy-Walker Malformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

Arthur E. Marlin and Sarah J. Gaskill

Chapter 17

Surgical Management of Chiari I Malformations and Syringomyelia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

Elizabeth C. Tyler-Kabara, Richard B. Morawetz, and W. Jerry Oakes

Chapter 18

Split Cord Malformations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

Dachling Pang

Chapter 19

Tethered Cord Syndrome Secondary to Previous Repair of a Myelomeningocele . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

Timothy A. Strait

Chapter 20

Sectioning of the Filum Terminale. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

Frederick B. Harris, Naina L. Gross, and Frederick A. Boop

Chapter 21

Diastematomyelia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Frederick B. Harris, Naina L. Gross, and Frederick A. Boop

Chapter 22

Lipomyelomeningoceles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Frederick B. Harris, Naina L. Gross, and Frederick A. Boop

Chapter 23

Untethering of the Spinal Cord after a Previous Myelomeningocele Repair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

Frederick B. Harris, Naina L. Gross, and Frederick A. Boop

Chapter 24

Brain Abscesses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

Darric E. Baty, Eli M. Baron, and Christopher M. Loftus

Chapter 25

Unilateral Coronal Synostosis (Plagiocephaly) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

James Tait Goodrich and David L. Staffenberg

Chapter 26

Moyamoya Syndrome in Children with Pial Synangiosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

R. Michael Scott and Edward R. Smith

Chapter 27

Selective Dorsal Rhizotomy for Spastic Cerebral Palsy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

Tae Sung Park and James M. Johnston

Chapter 28

Treatment of Lambdoidal Synostosis with Calvarial Reconstructive Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

David F. Jimenez, Constance M. Barone, and James Tait Goodrich

Chapter 29

Early Treatment of Lambdoid Synostosis with Endoscopic-Assisted Craniectomy . . . . . . . . . . . . . . . . . . . . . . . . . . 190

David F. Jimenez and Constance M. Barone

Chapter 30

Posterior Plagiocephaly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194

Richard G. Ellenbogen, Sudesh J. Ebenezer, and Richard Hopper

Chapter 31

Sagittal Synostosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

Larry A. Sargent and Timothy A. Strait

Chapter 32

The Separation of Craniopagus Twins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

Sami Khoshyomn and James T. Rutka

Chapter 33

Endoscopic Approaches to the Ventricular System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215

David F. Jimenez

Chapter 34

Intraventricular Endoscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

Jonathan P. Miller and Alan R. Cohen

Chapter 35

Infected Ventriculoperitoneal Shunts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233

Keyne K. Thomas, Sohaib A. Kureshi, and Timothy M. George

Chapter 36

Combined Fronto-Orbital and Occipital Advancement for Total Calvarial Reconstruction . . . . . . . . . . . . . . . . . . . 241
Ian F. Pollack

Chapter 37

Lumbosacral Meningoceles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

Ciaran J. Powers, Eric M. Gabriel, and Timothy M. George



Chapter 38

Surgical Correction of Unilateral and Bilateral Coronal Synostoses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

Jack Chungkai Yu and Ann Marie Flannery

Chapter 39

Myelomeningoceles, Split Cord Malformations, and Filum Terminale Dysgenesis . . . . . . . . . . . . . . . . . . . . . . . . . . 262

Robert F. Keating

Chapter 40

Lipomyelomeningoceles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
James Tait Goodrich

Chapter 41

Brain Stem Gliomas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275

Darlene A. Lobel and Mark R. Lee

Chapter 42

Posterior Fossa Decompression without Dural Opening for the Treatment of Chiari I Malformation . . . . . . . . . 281
Jonathan D. Sherman, Jeffery J. Larson, and Kerry R. Crone

Chapter 43

Metopic Synostosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286

Kant Y. K. Lin, John A. Jane Jr., and John A. Jane Sr.

Chapter 44

Total Cranial Vault Repair for Sagittal Craniosynostosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

John A. Jane Jr., Kant Y. K. Lin, Tord D. Alden, and John A. Jane Sr.

Chapter 45

Metopic Craniosynostosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296

Paul C. Francel and Jayesh Panchal

Chapter 46

Unicoronal Synostosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304

Jayesh Panchal and Paul C. Francel

Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

Continuing Medical Education Credit

Information and Objectives

Upon completion of this activity, the learner should be able to:
1. Describe the treatment and surgical management of pediatric neurosurgical disorders.
2. Demonstrate a full understanding of current neurosurgical operative techniques in pediatric neurosurgical disorders.
3. Discuss the operative management of complex pediatric neurosurgical disorders.

This activity has been planned and implemented in accordance with the Essentials and Standards of the Accreditation Council
for Continuing Medical Education through the American Association of Neurological Surgeons (AANS*). The AANS is accredited
by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians.

The AANS designates this educational activity for a maximum of 15 AMA PRA Category 1 credits. Physicians should only
claim credit commensurate with the extent of their participation in the activity.
The Home Study Examination is online on the AANS Web site at:
Estimated time to complete this activity varies by learner; activity equaled up to 15 Category 1 credits of CME.

Release/Termination Dates
Original Release Date: August 1, 2008
The CME termination date is: August 1, 2011

*The acronym AANS refers to both the American Association of Neurological Surgeons and the American Association of


Continuing Medical Education Disclosure

The AANS controls the content and production of this CME activity and attempts to ensure the presentation of balanced,
objective information. In accordance with the Standards for Commercial Support established by the Accreditation Council for
Continuing Medical Education, speakers, paper presenters/authors, and staff (and the signicant others of those mentioned)
are asked to disclose any relationship they or their co-authors have with commercial companies which may be related to the
content of their presentation.
Speakers, paper presenters/authors, and staff (and the signicant others of those mentioned) who have disclosed a relationship* with commercial companies whose products may have a relevance to their presentation are listed below.

Author Name


Type of Relationship

Frederick A. Boop
Richard G. Ellenbogen
Paul C. Francel
Arthur E. Marlin
John A. Persing
Jonathan D. Sherman

Medical Energy
W. Lorenz Group

Paid for speech given at Grand Rounds 2/2007

Clinical research study grant
Consultant fee
Grant (no personal funds paid)
Consultant fee (signed agreement)
Honorariumgave speech

Speakers, paper presenters/authors, and staff (and the signicant others of those mentioned) who have reported they do not
have any relationships with commercial companies:
Author Name
Tord D. Alden
Eli M. Baron
Constance M. Barone
Darric Baty
William O. Bell
Alan R. Cohen
Kerry R. Crone
John Drygas
Sudesh J. Ebenezer
Ann Marie Flannery
Eric M. Gabriel
Sarah J. Gaskill
Timothy M. George

James Tait Goodrich

Naina L. Gross
Stephen J. Haines
Frederick B. Harris
Richard Hopper
John A. Jane Jr.
John A. Jane Sr.
David F. Jimenez
James M. Johnston
Robert F. Keating
Sami Khoshyomn
Bianca Knoll
Sohaib A. Kureshi

Jeffery J. Larson
Mark R. Lee
Kant Y. K. Lin
Darlene A. Lobel
Christopher M. Loftus
David G. McLone
Arnold H. Menezes
Jonathan P. Miller
Richard B. Morawetz
W. Jerry Oakes
Jayesh Panchal
Dachling Pang
Tae Sung Park

Ian F. Pollack
Ciaran J. Powers
James T. Rutka
Larry A. Sargent
R. Michael Scott
Edward R. Smith
David L. Staffenberg
Timothy A. Strait
Keyne K. Thomas
Tadanori Tomita
Elizabeth C. Tyler-Kabara
Marion L. Walker
Jack C. Yu

*Relationship refers to receipt of royalties, consultantship, funding by research grant, receiving honoraria for educational services
elsewhere, or any other relationship to a commercial company that provides sufcient reason for disclosure.


Series Foreword

The Publications Committee of the American Association

of Neurological Surgeons began publishing the rst edition
of the Neurosurgical Operative Atlas in 1991. To allow for
timely publication, coverage of six operations was published
at bimonthly intervals in looseleaf format in the order nished manuscripts were received. The completed series had
nine volumes and covered the entire spectrum of neurosurgery.
The goal was to publish a comprehensive reference that
included well-established neurosurgical procedures as
practiced in the United States and Canada by authors who
are respected in the eld.
Working together, the AANS Publications Committee and
Thieme New York have organized the second edition of this
atlas series. The atlass main purpose remains the same, to
be a ready reference for well-established neurosurgical procedures for trainees and practitioners of neurosurgery worldwide. The new edition contains ve volumes, covering neurooncology; spine and peripheral nerves; functional; pediatric;
and vascular neurosurgery. For each volume, one or more

lead editors with known expertise in the subject area were

selected. Each volume editor had complete freedom to add,
revise, or delete chapters. The number of chapters per volume
is approximately the same as the number of chapters in that
particular subject area found in the rst edition.
Each chapter is designed to teach a specic surgical technique or approach. The illustrations of the techniques are a
vital part of the work, and the authors commissioned most
of the drawings in color. The text in each chapter covers the
case selection, the operative indications and contraindications, special points in the anesthetic technique, a step-bystep detailed description of the operation, and postoperative
complications. Detailed discussion of diagnostic techniques,
pathology, mechanisms of disease, histology, and medical
management are not included since they are logically outside the scope of a surgical atlas. Detailed tables, reference
lists, and statistical analysis of results are also not included
because they are readily available in standard texts.
We hope you nd this reference of value in your practice
of neurosurgery.
Robert Maciunas, MD
Past Chair, AANS Publications Committee
Professor of Neurosurgery
University Hospitals of Cleveland
Cleveland, Ohio



I came to pediatric neurosurgery after a nearly exclusively

adult neurosurgical training, and I confess to being a late
convert to pediatric neurosurgery as an almost autonomous
specialty. The surgical skills required to remove a medulloblastoma were no different, I thought, to those needed to excise a cerebellar metastasis (a far more frequently occurring
pathology), and as for all those shunts . . . !
And the conditions whose management fell within the
province of the pediatric neurosurgeon were all so rare . . .
I was wrong for several reasons, not least of which was to
regard pediatric neurosurgery too narrowlyas no more than
the deployment of particular operations. But I was right about
how uncommon (fortunately) in terms of their overall incidence
the conditions we deal with are. Pediatric neurosurgery is a specialty greedy for the resources it devours to treat comparatively
few patients, hydrocephalus aside. How should the expense of
managing a newborn with, say, Apert syndrome through infancy, childhood, and adolescence be balanced against returning
so many more adults to their former lifestyle after the replacement of a painful hip or knee? Or screening for such occult harbingers of morbidity as hypertension or glaucoma?
It does not take long for the fascination, indeed the privilege,
of working with children and their families to suppress such
thoughts, but nevertheless these remain questions to which
every pediatric neurosurgeon must have their own answer.
It has been said, and rightly so, that a society should be
judged by the way it treats its young and its oldthose on
the vulnerable edges of life. James Goodrich, in assembling
the impressive list of internationally renowned experts who
make up the list of his contributors, has demonstrated only
too clearly the commitment of the North American pediatric
neurosurgical community to the care of its sick children.
Since the rst edition of this book there has been further
acceptance and consolidation of pediatric neurosurgery as a
discipline that requires its own specialty training.
The conditions that are encountered in a pediatric neurosurgical practice vary markedly from those encountered in
an adult practice, as a glance at the titles of these chapters
with their emphasis upon a variety of congenital disorders
immediately reveals.

The days are long past, I hope, when a surgeon, as ambitious

as foolish, would wish to embark upon his or her rst transoral
procedure with no more technical experience than having read
Arnold Menezess excellent chapter on this subject.
Then whoand whatis an Operative Atlas such as this
for? It is rst and foremost a most valuable educational tool,
one that I would submit should be an essential companion for
pediatric neurosurgeons at all stages of their career.
Neurosurgical residents can read up on the procedures on
which they are about to assist; the increased understanding
of how and why these procedures are done will make them
not only more knowledgeable but all the more useful. Furthermore, when the surgery is over they can come back to
the atlas to nd out exactly why and how some maneuver,
whose signicance may have escaped them at the time, was
And not only the residents. Just as there exists no operation that cannot be improved (the nal improvement being
to do away with the need for surgery altogether) so there is
no neurosurgeon whose training is nished; indeed, anyone
who makes such a claim about themselves has, whether
they recognize it or not, already retired.
What pediatric neurosurgeon, however long ago their residency ended, would not wish to discover how the experts
who have contributed to this volume deal with conditions as
frustrating as a lipomyelomeningocele or a brainstem glioma?
Or the management of a common nuisance such as a shunt
infection? Or something more critical, such as a total calvarial
remodelling for the older child with sagittal synostosissurely
the limit to which cosmetic neurosurgery can be stretched?
I cannot imagine a pediatric neurosurgeon (in or postresidency) who will not be stimulated, intrigued, and (hopefully) sometimes provoked by what has been contributed
by the heady selection of experts James Goodrich has assembled here.
He is to be congratulated not only for bringing them together to produce this book but also for demonstrating so
clearly the commitment of the pediatric neurosurgical community of North America to its own continuing education
for the benet of sick and disabled children worldwide.
Richard D. Hayward, FRCS
Professor, Pediatric Neurosurgery
Great Ormond Street Hospital for Children
London, United Kingdom



O! Author, with what words will describe

with such perfection the whole conguration,
such as the sketch does here?
Leonardo da Vinci Quaderni dAnatomia
Volume II, fol. 2r.
Christiania: Dybwad, 19111916.
O! reader, the same feeling that inspired Leonardo to restate the Vitruvian man and provide the perfect proportion
inspired the authors to provide this palimpsest atlas, now
in the second edition with pages now re-etched and redone. Surgical techniques have continued to evolve with
both new technologies and also newly designed surgical approaches. In the early part of the 16th century, Leonardo da Vinci planned a 120-volume text on anatomy
with mostly anatomic illustrations. Early on, Leonardo
recognized that the visual images were more important
in educating the individual; he has clearly remained the
master at that concept. In producing a second edition of
this work, we clearly wanted to keep the original design
with the illustrations being paramount and with text being
added clarification. In an effort to disseminate these surgical techniques, this atlas was formulated on the principle
that the visual image is most important; it must be clear,
precise, and bring forth the surgical design with precision.
In addition, the text design and presentation needs to be
straightforward and practical. The atlas-style format using
an operative narrative was selected in the belief that, like
Leonardo da Vinci, surgeons are more comfortable with
the visual image than the written wordthough both
are provided! Starting with the first edition of this work,
we selected the chapters that dealt with pediatric subjects.
The original authors were asked to update and revise their
presentations as they felt necessary. In some cases extensive changes were made, and in others only some editing
was done. A number of the chapters are edited by several
surgical subspecialists, a common trend these days. Many
surgical procedures have clearly benefited from the skills
of several surgical subspecialties.
In recent years new materials, techniques, and equipment
have been offered to neurosurgeons for their operations. When
relevant, those contributions have been added. We specically
asked the authors to avoid lengthy bibliographies and to be

practical in their presentations; we think the readership will

see that this is clearly evident throughout the volume. Where
appropriate, the authors have been asked to include imaging,
whether computed tomography, magnetic resonance imaging, or other radiologic studies. The format of the chapters has
been kept the same throughout in order that the presentation
remains clear and follows a consistent structure.
For our surgical colleagues outside of neurosurgery, a
careful review of this book will reveal useful surgical ideas
and techniques. The format is designed so that not only is
the concept provided and discussed, but the surgical technique is discussed step by step in an atlas format, detailing
each member of the surgical teams contributions. Standards, techniques, and styles are continuing to change so that
hopefully a third edition of this book will be necessary in
the not so distant future. Our founding neurosurgical father
Harvey Cushing put it best when he stated:
The knowledge which a man can use is the only real knowledge,
the only knowledge which has life and growth in it
and converts itself into practical power.
The rest hangs like dust about the brain
and dries like raindrops off the stones.
Harvey Cushing
Laboratories: Then and Now, 1922, p. 9

No book comes about without the efforts of a number of
people, and this book is certainly no exception.
We would like to start by thanking the editorial and
production teams at Thieme New Yorkas always an extremely talented and most pleasant group to work withit
is clearly through their efforts that the published result is
so outstanding. At the top of the list of people are Brian
Scanlan, Tim Hiscock, Birgitta Brandenburg, Richard Rothschild, Dominik Pucek, and Ivy Ip. A special thanks to you
all for your individual contributions, which included skilled
editing, production, and design.
To Helen Lopez and Daniel Jimenez, a sincere thanks for
handling all the calls, mailing the bulky manuscripts, and
reminding us to be nice and to be on time.


xviii Preface
This book is about neurosurgery, and a fundamental part of
that team is our neurosurgical operating room nurses and technicians. To my operating room nurses and technicians, such an
essential part of my surgical team, special thanks for watchful
vigilance and your helpful advice and insight offered in the care
of patients. Hopefully this volume will be helpful to operating
room teams around in the world. Thanks to Mary Speranza,

Esther Ko-Uy, Katie Thompson, Charles Price, Loretta Caldwell,

Sylvan Fowles, Gabriel Ofurhie, Noel Greene, Danielle Dunne,
Cecille Edwards, Bindu Peter, and, nally, Charisse Terrya most
remarkable and skilled team with which to work!
Finally, to all the authors who contributed to this book,
thanks for doing such a wonderful job in presenting some
very complex subjects.


Tord D. Alden, MD
Assistant Professor
Department of Pediatric Neurosurgery
Feinberg School of Medicine
Northwestern University
Division of Neurosurgery
Childrens Memorial Hospital
Chicago, Illinois
Eli M. Baron, MD
Attending Neurosurgeon
Cedars-Sinai Institute for Spinal Disorders
Los Angeles, California
Constance M. Barone, MD, FACS
Division Head and Professor
University of Texas Health Science Center at San Antonio
Division of Plastic and Reconstructive Surgery
San Antonio, Texas
Darric E. Baty, MD
Department of Neurosurgery
Temple University School of Medicine
Temple University Hospital
Philadelphia, Pennsylvania
William O. Bell, MD, FACS
Neurosurgical Associates of the Carolinas
Winston-Salem, North Carolina
Frederick A. Boop, MD
Associate Professor of Neurosurgery
Chief, Pediatric Neurosurgery
Semmes-Murphey Clinic
LeBonheur Childrens Hospital
Memphis, Tennessee

Alan R. Cohen, MD
Division of Pediatric Neurosurgery
Rainbow Babies and Children Hospital
Cleveland, Ohio
Kerry R. Crone, MD
Professor of Neurosurgery and Pediatrics
Director, Pediatric Neurosurgery
Cincinnati Childrens Hospital Medical Center
Cincinnati, Ohio
John Drygas, MD, MS
Neuroscience and Spine Associates
Naples, Florida
Sudesh J. Ebenezer, MD, EdM
Department of Neurological Surgery
The University of Washington
Childrens Hospital and Regional Medical Center
Seattle, Washington
Richard G. Ellenbogen, MD, FACS
Professor and Chairman
Department of Neurological Surgery
University of Washington School of Medicine
Neurological Surgeon
Childrens Hospital and Regional Medical Center
Seattle, Washington
Ann Marie Flannery, MD, FACS, FAAP
Reinert Chair in Pediatric Neurosurgery
Department of Surgery
Saint Louis University School of Medicine
Cardinal Glennon Childrens Hospital
St. Louis, Missouri
Paul C. Francel, MD, PhD
Oklahoma Sports Science and Orthopedics
Oklahoma City, Oklahoma


xx Contributors
Eric M. Gabriel, MD
St. Vincents Medical Center
Jacksonville, FLorida
Sarah J. Gaskill, MD, FAAP, FACS
Associate Professor
Department of Neurosurgery
University of South Florida
Tampa, Florida
Center for Pediatric Neurosurgery and
All Childrens Hospital
St. Petersburg, Florida
Timothy M. George, MD, FAAP, FACS
Chief of Service
Dell Childrens Medical Center of Texas
Austin, Texas
James Tait Goodrich, MD, PhD, DSci
(Honoris Causa)
Professor of Clinical Neurological Surgery,
Pediatrics, Plastic and Reconstructive Surgery
Albert Einstein College of Medicine
Director, Division of Pediatric Neurosurgery
Center for Craniofacial Disorders
Childrens Hospital at Monteore
Bronx, New York
Naina L. Gross, MD
Department of Neurosurgery
Oklahoma University
Oklahoma City, Oklahoma
Stephen J. Haines, MD
Professor and Head
Department of Neurosurgery
University of Minnesota
Minneapolis, Minnesota
Frederick B. Harris, MD, MS
Division of Neurosurgery
St. Marys Duluth Clinic
Duluth, Minnesota
Richard Hopper, MD
Surgical Director
Craniofacial Center
Childrens Hospital & Regional Medical Center
Seattle, Washington

John A. Jane Jr., MD

Assistant Professor of Neurosurgery and
Department of Neurosurgery
University of Virginia Health System
Charlottesville, Virginia
John A. Jane Sr., MD, PhD
Professor and Chairman
Department of Neurosurgery
University of Virginia
Charlottesville, Virginia
David F. Jimenez, MD
Professor and Chairman
Department of Neurosurgery
University of Texas Health Science Center at San Antonio
San Antonio, Texas
James M. Johnston, MD
Department of Neurosurgery
Washington University School of Medicine
St. Louis Childrens Hospital
St. Louis, Missouri
Robert F. Keating, MD
Associate Professor of Neurological Department of Surgery
and Pediatrics
George Washington University School of Medicine
Division of Pediatric Neurosurgery
Childrens National Medical Center
Washington, DC
Sami Khoshyomn, MD
Pediatric Neurosurgery Fellow
Hospital for Sick Children
Toronto, Ontario
Bianca I. Knoll
Chief Resident
Department of Plastic Surgery
Yale University School of Medicine
Division of Plastic Surgery
Yale-New Haven Hospital
New Haven, Connecticut
Sohaib A. Kureshi, MD
Neurosurgical Medical Clinic
Chula Vista, California
Jeffrey J. Larson, MD
Coeur d Alene Spine and Brain, PLLC
Coeur dAlene, Idaho

Mark R. Lee, MD
Chairman, Department of Neurosurgery
Medical College of Georgia
Augusta, Georgia
Kant Y. K. Lin, MD
Department of Plastic Surgery
University of Virginia School of Medicine
Charlottesville, Virginia
Darlene A. Lobel, MD
Chief Resident
Department of Neurosurgery
Medical College of Georgia
Augusta, Georiga
Christopher M. Loftus, MD, DHC (Hon.), FACS
Professor and Chairman
Department of Neurosurgery
Assistant Dean for International Afliations
Temple University School of Medicine
Philadelphia, Pennsylvania
Arthur E. Marlin, MD, MHA
Department of Neurosurgery
University of South Florida
Division of Pediatric Neurosurgery
All Childrens Hospital
Tampa, Florida
David G. McLone, MD
Department of Pediatric Neurosurgery
Feinberg School of Medicine
Northwestern University
Childrens Memorial Hospital
Chicago, Illinois
Arnold H. Menezes, MD, FACS, FAAP
Professor and Vice Chairman
Department of Neurosurgery
University of Iowa Carver College of Medicine Department
of Neurosurgery
University of Iowa Hospitals and Clinics
Iowa City, Iowa
Jonathan P. Miller, MD
Department of Neurosurgery
University Hospitals and Health System
Cleveland, Ohio
Richard B. Morawetz, MD
Division of Neurosurgery
UAB Hospital
Birmingham, Alabama


W. Jerry Oakes, MD
Dan L. Hendly, Professor
Department Surgery
Division of Neurosurgery
University of Alabama at Birmingham School of Medicine
Pediatric Neurosurgery
Childrens Hospital
Birmingham, Alabama
Jayesh Panchal, MD, MBA, FRCS
Genesis Plastic Surgery and Medical Spa
Edmond, Oklahoma
Dachling Pang, MD, FRCS(C), FACS
Professor of Pediatric Neurosurgery
University of California
Davis Chief, Regional Centre of Pediatric Neurosurgery
Kaiser Permanente Hospital, Northern California
Oakland, California
Tae Sung Park, MD
Shi H. Huang Professor of Neurosurgery
Department of Neurosurgery
Washington University in St. LouisSchool of Medicine
Division of Neurosurgery
St. Louis Childrens Hospital
St. Louis, Missouri
John A. Persing, MD
Professor and Chief
Department of Plastic Surgery
Yale University School of Medicine
Division of Plastic Surgery
YaleNew Haven Hospital
New Haven, Connecticut
Ian F. Pollack, MD, FACS, FAAP
Department of Neurosurgery
University of Pittsburgh School of Medicine
Chief of Pediatric Neurosurgery
Childrens Hospital of Pittsburgh
Pittsburgh, Pennsylvania
Ciaran J. Powers, MD, PhD
Department of Surgery, Division of
Duke University
Division of Neurosurgery
Durham, North Carolina
James T. Rutka, MD, PhD, FRCS(C), FACS, FAAP
Professor and Chairman
Department of Neurosurgery
University of Toronto
Division of Neurosurgery
The Hospital for Sick Children
Toronto, Ontario

xxii Contributors
Larry A. Sargent, MD
Professor and Chair
Department of Plastic Surgery
UT College of MedicineChattanooga Unit
Chattanooga, Tennessee

Keyne K. Thomas, MD
Pediatric Neurosurgery Service
Division of Neurosurgery
Duke University Medical Center
Durham, North Carolina

R. Michael Scott, MD
Department of Neurosurgery
The Childrens Hospital
Boston, Massachusetts

Tadanori Tomita, MD
The Yaeger Professor of Pediatric Neurosurgery
Professor of Neurosurgery
Feinberg School of Medicine
Northwestern University
Chairman, Division of Pediatric Neurosurgery
Childrens Memorial Hospital
Chicago, Illinois

Jonathan D. Sherman, MD
Mountain Neurosurgical and Spine Center, PA
Asheville, North Carolina
Edward R. Smith, MD
Assistant Professor
Department of Neurosurgery
Harvard Medical School
Attending Neurosurgeon
The Childrens Hospital, Boston
Boston, Massachusetts
David L. Staffenberg, MD
Associate Professor
Department of Clinical Plastic Surgery, Neurological
Surgery, and Pediatrics
Albert Einstein College of Medicine, Yeshiva University
Monteore Medical Center, Childrens Hospital at
Bronx, New York
Timothy A. Strait, MD
Clinical Assistant Professor of Surgery
Department of Surgery
University of Tennessee College of Medicine
Chattanooga Unit
The Neurosurgical Group of Chattanooga
Chattanooga, Tennessee

Elizabeth C. Tyler-Kabara, MD, PhD

Assistant Professor
Department of Neurological Surgery
University of Pittsburgh School of Medicine
Division of Pediatric Neurosurgery
Childrens Hospital of Pittsburgh
Pittsburgh, Pennsylvania
Marion L. Walker, MD
Department of Neurosurgery
University of Utah School of Medicine
Division of Pediatric Neurosurgery
Primary Childrens Medical Center
Salt Lake City, Utah
Jack Chungkai Yu, MD, DMD, MS Ed
Department of Surgery
Section of Plastic and Reconstructive Surgery
Medical College of Georgia
Augusta, Georgia

Fibrous Dysplasia Involving the Craniofacial
James Tait Goodrich
This chapter will deal with brous dysplasia of the craniofacial complex, in particular those regions involving the
forehead, orbital rim, lateral and medial orbital walls, the
orbital roof, and the optic foramen. The discussion will involve the worst case scenario, assuming that, if the surgeon can handle this type of case, the simpler cases will be
easier to treat.
Fibrous dysplasia can involve the calvaria and any of the
upper facial bones. Its etiology is unknown, but the pathology involves a replacement of normal bone with a broosseous matrix. The surgical principle involves removing
all of the dysplastic bone (or as much as possible) and replacing it with normal calvarial bone harvested from other
parts of the head. Fibrous dysplasia can be of a simple type
called monostotic, where only one bone unit is involved,
or polyostotic, where two or more bones are involved. In
this chapter we will deal with the more complicated polyostotic type.
The most common presenting complaints in brous dysplasia of the craniofacial complex are proptosis (Fig. 11),
diplopia and headaches, and in severe cases, progressive

blindness due to optic nerve compression; fortunately,

blindness is an extremely rare outcome.
An x-ray lm of the skull will show a sclerotic mass expanding the calvarial and orbital bones. The radiologist typically describes a ground glass appearance. There will also be
sclerosis or even a cystic appearance to the bone. It is not uncommon to see complete obliteration of the frontal and nasal
sinuses. The proptosis is secondary to the orbital brous dysplasia compressing the globe and forcing the eye forward. As a
result of this, an early presenting complaint can be diplopia.
The principle behind the surgical treatment of brous
dysplasia of the craniofacial complex is threefold: (1) relief
of optic nerve compression (decompression of the nerve
should be considered though is not always essential); (2)
removal of all dysplastic bone if possible, as any residual can
form a new dysplastic center; and (3) use of the patients
own bone for grafts to achieve a satisfactory cosmetic result
is preferred. We now avoid the use of any foreign implantable materials such as methylmethacrylate, wire mesh, or
metal xation plates, particularly in children, due to high
risk of migration and infection.


Figure 11 (A) Frontal view and (B) superior view of a patient with
orbital proptosis secondary to brous dysplasia. Typical proptosis is
evident and brous dysplasia involving the right orbital unit including

rim, lateral, and medial walls. As a result, the eye is pushed forward
and downward. Interestingly, the only visual symptom was double vision: The visual acuity was normal.

14535_C01.indd 1

4/23/08 3:13:33 PM

2 Pediatric Neurosurgery
At the Craniofacial Center of the Childrens Hospital at
Monteore we now do as much as possible of the reconstruction with normal calvarial bone, that is, bone not involved with brous dysplasia. We have found this considerably lessens the risk of resorption, which occasionally occurs
with rib grafts placed in the craniofacial region. The use of
ribs, particularly in the forehead region, can sometimes lead
to an unacceptable washboard appearance. Another advantage of using calvarial bone is the reduction in operative
exposure. This technique also avoids the complications that
can occur with rib harvesting, such as pneumothorax and
chest wall pain.

Preoperative Preparation
All patients should have radiographic studies of the skull
in the routine views to document the extent of dysplastic
involvement of the skull and surrounding orbital and nasal structures. Computed tomography scanning with bone
windows in the axial and coronal views along with threedimensional reformatting are obtained for the operative
planning. We have not found magnetic resonance imaging
to be helpful, so we do not use it routinely.
If the optic nerve is compressed, we routinely do visual
acuity and visual eld testing to have baseline values. Damage to the optic apparatus and to the nerves supplying the
extraocular muscles are the most signicant complications
to be avoided. Subtle damage may already have occurred
preoperatively, and it is essential to document this prior to
any surgical intervention. In recent years our ophthalmologic colleagues now feel it is no longer always necessary to
decompress the optic nerve, even in cases of severe radiological compression. Some surgical teams now feel the risk
of removing the dysplastic bone is too great in causing direct
injury to the nerve. As a result of these recent discussions
we no longer just routinely decompress the optic nerve. The
exception is a rapid and clear progression of visual loss due
to an overgrowth of dysplastic bone.
We routinely start an anti-staphylococcal antibiotic at the
time of anesthetic induction in the operating room. Because
the surgical manipulations are extradural, we do not routinely use anticonvulsant medications.

pressure line. An arterial line is highly recommended for

monitoring blood gases, hematocrit, electrolytes, etc., during the procedures. We request an osmotic diuresis, usually
with mannitol (0.5 mg/kg) at the time of anesthetic induction. A spinal drainage system for cerebrospinal uid (CSF)
removal can sometimes also be helpful for brain relaxation.
These relaxation measures can be key in getting a relaxed
frontal lobe for retraction when working back toward the
optic foramens.
We do not routinely use steroids in these types of cases;
the exception is if there is evidence of brain or optic nerve
edema during the case.

Operative Procedure
The patient is placed in the supine position with the head
resting on a cerebellar (horseshoe) headrest (Fig. 12). The
head is placed in a slightly extended, brow-up position.
Rigid xation devices such as a Mayeld clamp are specically avoided, as the surgical team will need to move

Fibrous dysplastic bone can be, and usually is, highly vascular. As a result, the blood loss in these procedures can be
signicant. We routinely plan for a blood loss of 2 to 3 units.
If the family wishes, we arrange for pedigree blood donations from family members 1 week in advance. The patient
can also donate his/her own blood prior to surgery. If available, a cell saver unit can rescue up to 50% of the patients
lost blood volume. Because of the risk of extensive blood
loss, all patients require at least two large-bore intravenous
lines of 16 gauge or larger. If there is any history of cardiac or
pulmonary problems, we routinely put in a central venous

14535_C01.indd 2

Figure 12 Schematic showing the location of the surgical and anesthesia teams.

4/23/08 3:13:34 PM

1 Fibrous Dysplasia Involving the Craniofacial Skeleton

the head (usually never more than 10 to 15 degrees): This
exibility can prove to be very useful.
We also like to reverse the operating room table so the
head of the patient is at the foot end of the operating room
table. This positioning allows the surgeon and assistant to
sit with their knees comfortably under the table and not
obstructed by the table pedestal or foot unit.
Anesthesia equipment is placed on the side opposite the
lesion and parallel to the table. Routine orotracheal intubation is performed. All lines are run off to the side of the
anesthesia unit. The operating surgeon is placed at the head
of the patient with the assistant to the side. The scrub nurse
comes in over the patients abdomen but is positioned no
higher than the mid-thoracic region. This allows the surgeon
to be able to move around to see the patients face fully for
cosmetic evaluation. For this reason we also avoid the use of
bulky overhead tables, such as the Fallon table.
If the patient has signicant proptosis or the threat of injury to the globe is signicant, then we place tarsorrhaphies
just prior to formal surgical draping. This maneuver can be
quite helpful in preventing any unintentional injury to the
globe and cornea.

Surgical Draping
The head is draped for a bicoronal incision. The hair is not
shaved but is either parted for the incision or a small 1.5
cm width of hair is taken for the incision line. We also nd
it helpful to braid the hair if it is long to keep it out of the
eld. In children a zig-zag type of incision is made. We have
found these incisions reduce the keloid formation over the
temporalis muscle and also prevent the hair from parting
directly over the incision when it is wet. The draping is done
in such a fashion that both eyes are visible. The facial drape
is placed over the nose and nares but well below the lower
orbital rim. This allows the eyes to be visualized during the
reconstruction. The rest of the draping can be done according to the surgeons preference. An important additional
point is to keep the drapes reasonably loose, so that the
head can be moved.
We routinely run all our suction lines, cautery cords, etc.,
past the foot of the patient. As both surgeons are sitting, this
allows easy mobility of the chair; that is, they are not rolling
over the cords and tubes.
Because the operative site is usually copiously irrigated
during the procedure, it is important to have waterproof
outer drapes. Some of the recent drape designs provide a
large plastic bag for uid collection: We have found these
to be quite useful.

Skin Incision
Over the years we have used several different incisions but
now almost always routinely use a bicoronal incision carried from ear to ear. The incision is started behind the ear
helix, not in front (as is typically the case with most surgeries), to reduce scarring. The incision, particularly in children, is done in a zig-zag fashion until reaching the vertex,
where it is straightened. The incision is made well behind

14535_C01.indd 3

the patients hairline, not at the hair edge, a common error in placement. This incision type allows for an extensive
exposure of the calvaria for tumor removal and additional
bone harvesting as necessary. In addition a large sheet of
pericranium is available for any repairs of dura or frontal

Flap Elevation
A full-thickness ap is turned following the standard subgaleal plane. It is important to leave the pericranium intact.
The pericranium is then elevated as a second separate layer.
The aps are carried down to the orbital rim to the level of
the supraorbital nerve and artery. These structures are frequently encased in a small notch of bone. This notch can be
opened with a small Kerrison rongeur or osteotome. It is easier to elevate the artery and nerve with the pericranial layer.
It is important to preserve these structures or there will be
anesthesia, or even worse dysesthesia, in the forehead postoperatively. The ap must also expose the entire belly of the
temporalis muscle and the zygomatic arch. In the midportion of the face the nasion suture should be fully exposed.
Using the small periosteal dissector or a Peneld dissector it
is possible to come under the orbital rim and dissect it safely
back ~1 to 2 cm. The temporalis muscle has to be elevated as
a unit. Starting at its squamosal insertion, it is elevated using
a Bovie electrocautery with a ne needle tip. The dissection
is performed in such a fashion that the temporalis muscle
will be elevated from the frontozygomatic suture back to the
ear, fully exposing the pterional keyhole.

The craniotomy is performed to incorporate all of the dysplastic
bone in the removal. It is easiest to do the frontal craniotomy
by rst taking out a forehead bone ap that encompasses
as much of the forehead dysplasia as possible (labeled A in
Fig. 13). This provides the window that will allow exposure
to the orbital roof and walls. We prefer to use a high-speed
drill system with a craniotome (e.g., Midas Rex with a B-1;
Medtronic Inc. footplate [Fort Worth, Texas]) as this gives
a speedy bone removal, thereby decreasing blood loss. In
some cases the dysplastic bone can be extremely thick and
we will use an S-1 attachment. We next elevate the frontal
lobe with gentle retraction to see how far into the orbital
roof the dysplastic bone extends. Then, by further dissecting under the orbital roof, the dysplastic portion can be
completely visualized (Fig. 14). There is usually extensive
blood supply crossing these planes, so the bleeding can be
quite copious. Keep plenty of Avitine and Gelfoam available
for packing in these spaces to control the oozing. Once the
limits of the dysplastic bone have been determined and the
brain is adequately relaxed and retracted, we proceed with
the bone resection. Using a combination of osteotomes and
a small cutting bur, such as the Midas Rex C-1 attachment,
the roof is removed as a unit (Fig. 14). It is helpful to have
the assistant place a malleable retractor under the orbital
roof. This will prevent the drill or osteotome from damaging
the perioribita. On occasion, the dysplasia can go back to the

4/23/08 3:13:36 PM

4 Pediatric Neurosurgery

Figure 13 Frontal view showing four-piece bone removal. (A) Frontal

bone maximally involved with brous dysplasia; (B) lateral orbital wall;
(C) medial orbital wall. The orbital wall roof, which is also removed, is
not shown in this drawing. (D) Graft site from the opposite calvaria.

clinoids and orbital foramen. In these cases, an operative decision has to be made in regards to the orbital nerveleave
the foramen alone or unroof it (Fig. 14) A small diamond
bur on a high-speed drill unit is the best method for removing this part of the bone. Copious irrigation is applied to
prevent any unnecessary thermal injury to the bone and
nerve. Once this is completed, attention is turned to the
lateral orbital wall and zygoma (labeled B in Fig. 13). This
portion of the procedure can be done quite easily. The only
important points are to have adequate exposure of the zygomatic arch and a good dissection of the orbit. The lateral
canthal ligament must be sectioned and then reattached at
the end of the procedure. We often place a suture through
the canthal ligament for later identication. Doing this prior
to the medial part will allow easy mobilization of the eye
and surrounding structures with minimal trauma.
Next, attention is turned to the most difcult phase
resecting the medial nasal structures (labeled C in Fig. 13).
By removing the orbital roof and lateral orbital wall, the surgeon now has some mobility and freedom in the moving the
globe. If the dysplastic bone involves the nasal bone and medial orbital wall, the medial canthal ligament is identied
with a ligature and then cut. The assistant then retracts the
eye laterally, and the bone is removed with an osteotome and

14535_C01.indd 4

Figure 14 Schematic drawing showing the frontal fossa after removal of the dysplastic orbital roof and decompression of the optic
nerve at the foramen.

ne cutting bur. The frontal sinuses are often obliterated with

dysplastic bone, which can complicate matters. If the sinuses
are not occluded, the frontal sinus can be entered and used as
an operating space within which to work. Once all the dysplastic bone is removed, the reconstruction is started.

Some Helpful Hints

On occasion, the dysplastic bone can be extremely thick, in
one case more than 8 cm in thickness, which exceeds the
cutting width of any known craniotome. In these cases one
can just shave the bone down with a large bur. The other
option is to put a long, ne cutting tip on a short attachment and then gently cut the bone. The disadvantage is one
cannot see deep into the craniotomy so the craniotomy is
all done by feedback to the ngers. Careful attention has
to be focused on the midline structures such as the sagittal
sinus. We have also found on occasion the dura can also be
very dysplastic and densely adherent to the overlying bone.
It can be extremely tedious to remove the dura from the
dysplastic bone and care must be taken to avoid disrupting large veins that commonly run through the dura. On a
couple of occasions serious air embolisms have occurred,
and the anesthesia and surgical teams need to prepare to ag-

4/23/08 3:13:36 PM

1 Fibrous Dysplasia Involving the Craniofacial Skeleton

gressively deal with an air embolism as these are potentially

life-threatening situations if not treated rapidly.

Calvarial Bone Harvesting

By using a bicoronal skin ap, typically a large amount of normal calvarial bone can be exposed. Once the surgeon has resected the dysplastic bone and determined how much bone is
needed to reconstruct the defect, a craniotomy is performed
on the opposite calvaria (labeled D in Fig. 13). Remember
that the most useful bone is over the convexity, where the
diplo is well formed. In the squamosal area, the bone thins
out and is harder to split. The bone is taken to a sterile table
set up next to the operating eld. Using a combination of
small osteotomes, a ne cutting tip such as a Midas Rex C-1,
and a reciprocating saw, the bone is split along the diploic
space. Copious irrigation is essential to prevent thermal injury to the bone; dead or necrotic bone does not heal well.
Once the bone has been split, the inner table of the calvaria is
placed back in the harvest site. The outer table, because of its
smooth contours, is used as the reconstructing bone.

Craniofacial Reconstruction
The reconstruction is done in the reverse order from the
resection. The medial orbital wall is constructed rst and
plated into position (labeled C in Fig. 15). The nasal bone
and cribriform plate are usually the most solid structures to
work with. The medial canthal ligament also has to be reattached, which can be done easily through a small drill hole.
Next, a piece of bone is fashioned to form the orbital roof.
This is an important structure that must be solidly placed
(Fig. 16). If it is not, subsequent proptosis (sometimes enophthalmos, too) of the eye can occur due to downward
pressure of the frontal lobe. The bone used to reconstruct
the lateral orbital wall is attached to the roof with either
wires or miniplates (labeled A in Fig. 15). The squamosal
portion of the temporal bone can also act as an excellent
place to anchor this bone. The orbital rim is then fashioned

Figure 16 Schematic drawing showing the split-thickness calvarial

bone graft in position in the orbital roof region. This bone unit is key
to preventing an unacceptable eye migration.

and attached medially to the nasal unit and opposite orbital

rim (labeled B in Fig. 15). This is the key cosmetic unit and
must be perfectly placed to avoid facial asymmetry. The rest
of the craniotomy is then closed in a mosaic fashion using
the remaining pieces of bone. Miniplates have proved to be
extremely useful in stabilizing these various bone units. In
children and adolescents we now routinely use the absorbable plating systems (Figs. 17 and 18).

Repair of Frontal Sinus

One of the most devastating postoperative complications
is infection arising from the sinus. If the frontal sinus is
not obliterated by dysplastic bone, it must be cleaned and
exenterated of mucosal lining. We routinely cover the sinus
with the pericranial ap to isolate it from the epidural space.
The same principle applies to the other paranasal sinuses if
they are violated.

Pericranial Tissue
Figure 15 Schematic drawing showing the harvested bone grafts in

14535_C01.indd 5

The pericranial tissue is a most useful repair structure. It

not only provides additional vascularity to the bone, but it
also helps smooth out the rough contours of the bone that

4/23/08 3:13:37 PM

6 Pediatric Neurosurgery
Use of Dysplastic Bone in Repair
In some cases, if there is a shortage of bone for the repair,
the dysplastic bone can be reused. The dysplastic bone must
be rm, if not hard, with no soft or mushy spots present. In
some cases we have found the dysplastic bone to be nearly
as rm as normal bone and have reused it in the reconstruction. In some cases, where there is extensive brous
dysplasia, one has no choice but to reuse the original diseased bone. However, the caveat remains: If normal bone is
available, this is the better option to use.


Figure 17 Intraoperative view showing the harvested bone grafts in


The closure is done in a routine fashion. Hemostasis must

be meticulous because of the amount of dead space that
can form. In some, but not all, cases a subgaleal drain to
light suction is placed for 24 to 48 hours. A uid collection
next to sinus spaces can lead to a devastating postoperative
infection. Scalp closure is done in a routine fashion closing
both the subgaleal and skin layers.


has been harvested and used as grafts. Therefore, we make

every effort to use this structure and place it back into its
natural anatomical position.

Temporalis Muscle
To prevent a postoperative depressed concavity over the
temporal unit, the temporalis muscle is laid back into position. Sometimes a relaxing incision must be made posteriorly to allow the muscle to be advanced forward to cover the
keyhole and to be reattached to the zygoma. This is critical
or there will be a signicant hourglass deformity over this
region postoperatively.

Figure 18 Postoperative photograph taken 6 weeks after surgery

showing aesthetic reconstruction obtained.

Postoperative Management Including

Possible Complications
We routinely place the patient on antibiotics to cover skin
organisms and possible nasal contaminants for at least 72
hours. The risk of osteomyelitis is high and can be quite
devastating to the patient, so every attempt must be made
to avoid it. There may be signicant periorbital swelling
postoperatively; ice packs are applied to the eye and periorbital region for symptomatic relief. If there is signicant
swelling at the end of the operation, we ordinarily leave the
tarsorrhaphy in place for about 2 days. Intensive care for at
least 48 hours is mandatory with close monitoring for hemodynamic changes from excessive blood loss and for the
development of an epidural hematoma.
The neurosurgeon must always be attentive to postoperative CSF leaks. If dural tears have occurred, they must be repaired meticulously. Should a postoperative CSF leak occur,
then placement of a lumbar CSF drain may be necessary to
divert the uid. These drains usually need to be left in place
for 5 to 7 days. However, close attention to dural tears and
verifying dural integrity by asking the anesthesiologist to
perform a sustained Valsalva maneuver at the end of the
case should prevent this problem from occurring.
This chapter is a revision of the chapter, Fibrous Dysplasia
Involving the Craniofacial Skeleton by James T. Goodrich,
M.D. and Craig D. Hall, M.D. The chapter appeared in the
Neurosurgical Operative Atlas, Volume # 1, edited by Setti S.
Rengachary and Robert H. Wilkins. The Neurosurgical Operative Atlas was published by the American Association of
Neurological Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank Craig Hall, M.D.,
for his help and efforts on the original chapter published in
the rst edition of this work.


14535_C01.indd 6

4/23/08 3:13:39 PM

Chiari Malformations and
Elizabeth C. Tyler-Kabara and W. Jerry Oakes
Chiari malformations, or hindbrain hernias, are being diagnosed and operated upon with increasing frequency. For the
purposes of this chapter, two separate entities will be discussed. The Chiari I malformation is characterized by caudal
descent of the cerebellar tonsils. The brain stem and neocortex
are typically not involved and the patient does not suffer from
a myelomeningocele. Syringomyelia is commonly but not invariably present. The Chiari II malformation is almost always
seen in conjunction with spina bida and is a more severe
form of hindbrain herniation. The neocortex and brain stem
are dysmorphic and the cerebellar vermis (not the tonsils) is
displaced into the cervical spine. Accompanying the vermis are
dysmorphic and elongated aspects of the medulla and lower
pons as well as the lower aspect of the fourth ventricle. Again,
syringomyelia is commonly associated with this lesion. Not
discussed in this chapter is the rare Chiari III malformation.

Chiari I Malformations
Patient Selection
With the advent of magnetic resonance imaging (MRI) the
detection of caudal displacement of the cerebellar tonsils
and the presence of an associated syrinx has become safe
and accurate. Typically the tonsils are at least 3 mm below
the plane of the foramen magnum. They lose the rounded
appearance of their caudal pole and become pointed or
peg-like. This is associated with obliteration of the subarachnoid space at the craniocervical junction with the impaction of tissues into this conned region. When all of the
above criteria are not met, the situation should be judged
in conjunction with the clinical symptomatology of the patient. The presence of syringomyelia or other developmental
anomalies will further assist in the interpretation of the
intradural ndings at the craniocervical junction.
Patients with a symptomatic Chiari I malformation are
generally offered operative intervention. The more severe
the neurological decit, the stronger the case for intervention. When occipital pain is the only symptom and no neurological signs are present, the degree of disability from the
discomfort should be carefully weighed against the risks
of the procedure, prior to the implementation of surgical
intervention. With advances in cine MRI, some patients with
occipital headaches and mild hindbrain hernias have been

found to have abnormal cerebrospinal uid (CSF) ow at

the craniocervical junction. When syringomyelia is present,
we generally favor intervention even with minimal symptoms. The absence of caudal displacement of the cerebellar
tonsils but with a compressed look to the tissues at the
foramen magnum associated with a signicant syrinx has
been termed a Chiari 0. When the syrinx is signicant, the
Chiari 0 patient should be considered for decompression. Intracranial pressure should be normalized prior to consideration of craniocervical decompression. Approximately 10% of
patients with Chiari I malformation will have hypertensive
hydrocephalus, and ventriculoperitoneal shunt insertion
should precede other considerations. Flexion and extension
views of the cervical spine are also important to resolve
questions of spinal stability and other bony anomalies. If
signicant basilar invagination or retroexion of the dens
is present, the need for an anterior decompression should
be discussed prior to proceeding with a posterior procedure.
If the posterior procedure is performed rst, it should be
recognized that acute neurological decompensation postoperatively warrants emergent anterior decompression.

Preoperative Preparation
Once a candidate for surgery has been appropriately chosen,
the patient is prepared with preoperative antibiotics. The
patient is positioned prone (Fig. 21) in a pin-type head
holder with the neck exed. The head of the table is elevated
somewhat, but no central venous access is mandatory because lowering the head will eliminate the gradient for air
embolization. A chest Doppler monitor may be used for the
detection of air embolization and to monitor slight changes
in the patients pulse. Patients are paralyzed and are not
allowed to breathe spontaneously. This signicantly lowers
the likelihood of serious pulmonary complications postoperatively. Muscle relaxants are allowed to become fully
effective during the induction of anesthesia to avoid the
Valsalva maneuver during placement of the endotracheal
tube. A severe Valsalva maneuver has been associated with
progression of symptoms in some patients.

Operative Procedure
The skin incision is made from a point 2 cm below the
external occipital protuberance to the midportion of the

14535_C02.indd 7

4/23/08 3:16:05 PM

8 Pediatric Neurosurgery

Figure 21 Optimal positioning of a patient for exposing a Chiari malformation.


spinous process of C-2 (Fig. 22A). It is quite unusual for

the tonsilar tissue to descend below the level of the upper portion of C-2 (Fig. 22B), and by preserving the important muscular attachments at C-2 postoperative pain
is signicantly decreased. The likelihood of postoperative
spinal deformity seen in conjunction with syringomyelia
is also substantially lessened. The avascular midline plane
of the occipital musculature is divided with monopolar
current. No incision transecting muscle is necessary in this
procedure because the plane dividing the left and right
muscular bundles completely separates these two groups.
A small amount of fat will mark this natural cleavage plane.
Again using the monopolar current, the muscle insertion
immediately above the foramen magnum is separated
from the occipital bone and the posterior arch of C-1. First
the occipital bone is removed followed by the arch of C-1
(Fig. 22C). There is no need for lateral exposure, and bone
laterally situated is left intact. This minimizes the risk of
injury to the vertebral veins and arteries. The bone edges
are waxed, and the dura is opened in the midline. Initially,
the dura over C-1 is opened. Care is taken as the incision is
extended across the circular sinus near the foramen magnum. This sinus can be formidable and should not be ap-

14535_C02.indd 8

proached nonchalantly. With the dura retracted laterally,

the arachnoid is opened in the midline. The subarachnoid
adhesions are lysed with a sharp instrument and are not
simply torn. The arachnoid edge is then clipped to the dural
edge with metal clips.
The tonsils, which can be recognized by their vertical folia, are separated in the midline to expose the oor of the
fourth ventricle (Fig. 22C). Care is taken to free the caudal
loop of the posterior inferior cerebellar artery and to avoid
damage to this vessel or its branches. Again adhesions are
cut rather than torn. On separating the cerebellar tonsils, a
veil of arachnoid is sometimes encountered. This veil should
be opened widely. Obstruction to CSF ow can also occur
from the posterior inferior cerebellar arteries. These vessels
may approximate in the midline and be adherent to one
another. They should be separated and mobilized laterally
with great care. Closure is then accomplished with a generous dural graft. We prefer to use the patients own pericranium, but either cadaveric or articial dural substitutes may
be used. The dural closure is with absorbable suture, which
is known to react minimally in the subarachnoid space. A
Valsalva maneuver is performed to check for CSF leaks prior
to closing the wound in multiple layers.

4/23/08 3:16:06 PM

Chiari Malformations and Syringohydromyelia


Postoperative Management Including Possible

Following the operation, patients may experience some
nausea and vomiting as well as hiccups. These are almost
always self-limited. Neurological decits that are wellestablished prior to the operation are unlikely to reverse
following manipulation. Long-standing pain and temperature loss is very unlikely to return. Hand and arm weakness
with fasciculations and loss of muscle bulk may improve
functionally but may not normalize. A particular problem
exists when pain is a major component of the presentation.
Children and adolescents infrequently have a major problem
with pain. Adults, however, may be quite discouraged by
the persistence of discomfort in the neck, shoulders, and/or
arms. Pain may very well persist despite a physiologically

14535_C02.indd 9

Figure 22 (A) Schematic of incision placement and bone work for

operative exposure of a Chiari I malformation. (B) Chiari I malformation associated with syringomyelia, midline sagitttal section. (C) Operative exposure of a Chiari I malformation with the oor of the fourth
ventricle exposed.

successful operation with obliteration of the syrinx cavity.

This limitation of surgical intervention should be carefully
explained to the patient prior to surgery. Mild scoliosis (less
than 35 degrees) may improve or simply stabilize, whereas
more severe spinal deformity may well progress despite
adequate treatment. With the advent of MRI scanning the
status of syringohydromyelia can be assessed easily. If a
sizable syrinx persists months to years after craniocervical
decompression, a second decompression should be considered. It is our experience that >85% of children will have
resolution of their syringes following a simple craniocervical decompression. If the syrinx persists for 9 to 12 months
following the initial decompression or if there is a return
of symptoms and a syrinx that had once shown evidence
of decompression, then consideration should be given to a
secondary decompression. During this secondary procedure

4/23/08 3:16:13 PM

10 Pediatric Neurosurgery
one or both cerebellar tonsillar tips are resected to allow an
unimpeded expansion of the foramen of Magendie. With
this second intervention the vast majority of the recurrent
and persistent syringes will resolve. If after two decompressions a larger syrinx is persistent and symptoms attributable
to this lesion are serious or progressive, consideration can
be given to a laminectomy over the lower aspect of the syrinx and the placement of a syrinx to subarachnoid shunt
or a syrinx to peritoneal shunt. If a syrinx to subarachnoid
shunt is chosen, placement of the distal catheter in the free
subarachnoid space is an important technical maneuver.
Catheters can easily be mistakenly placed in the subdural
space without benet to the patient.

Chiari II Malformations
Patient Selection
Children with myelomeningocele may develop symptoms
referable to their hindbrain hernias. Symptoms and signs are
generally age-specic, with infants developing lower cranial nerve disturbances (difculty with swallowing, weak
cry, inspiratory wheeze, aspiration pneumonia, absent gag,
and opisthotonos) and older children more commonly developing progressive upper extremity spasticity. Ataxia of
the trunk or appendages is recognized much less often. Because some degree of hindbrain herniation is present in
the vast majority of spina bida patients, MRI evidence of
hindbrain herniation must be accompanied by progressive
or signicant symptomatology to warrant operative intervention. Many patients will remain clinically stable for long
periods despite signicant anatomical deformity. As many
as one-third of patients will develop difculty with phonation, swallowing, or apnea by age 3 years. If the asymptomatic remainder were followed for a longer period or
if less serious symptoms were considered signicant, this
one in three gure would undoubtedly be higher. Because
the symptoms of the Chiari II malformation are frequently
life-threatening, symptomatic Chiari II malformation is the
leading cause of death in the treated myelomeningocele
population today. When treated conservatively, as many as
5 to 10% of all patients will die from the malformation by
the age of 3 years.
The decision for surgical intervention is controversial.
Because there is a signicant likelihood of stabilization or
actual improvement with conservative care, some would
argue against operative intervention. This is supported to
some degree by autopsy material that demonstrates hypoplasia or aplasia of vital lower cranial nerve nuclei. Against
this, however, is the experience of numerous surgeons who
have seen dramatic improvement in many patients following decompression. In addition, objective evidence of physiological functioning has been reported to improve with both
brain stem evoked responses and CO2 curve following operation. With these conicting pieces of evidence one can
quickly appreciate the surgeons dilemma.
With increasing experience, the senior authors willingness to operatively intervene is increasing. This is due to the

14535_C02.indd 10

relatively low incidence of operative complications and the

clear improvement demonstrated by some patients. Poor
results are more commonly due to a delay in offering operative intervention. Once serious difculties are clinically
evident with breathing, swallowing, or phonation, the situation may very well be irreversible. In that case, the best
outcome that surgery can be expected to yield is maintenance of the poor level of lower cranial nerve function seen
immediately prior to operation. Problems with aspiration
pneumonia, apnea, and other life-threatening difculties
may very well persist.
The solution to this problem does not seem to be a continuation of a conservative approach, accepting 10 to 15%
mortality. Rather, an earlier identication of patients at high
risk for serious problems, and offering this group intervention, seems to be a more logical option. Being able to detect
this high-risk group prior to the development of irreversible
life-threatening problems is a key provision.
If serious problems with phonation, swallowing, or breathing are detected and normal intracranial pressure is present,
urgent intervention is appropriate when full support of the
child is proposed. It is also important to emphasize that
normalization of intracranial pressure is a prerequisite to
consideration of craniocervical decompression. Patients
with questionable shunt function are well served to rst
have their shunts revised. If progressive or serious symptoms persist after adequate shunt revision, decompression
of the craniocervical junction can be contemplated. Again,
the MRI has made the diagnostic evaluation of this group of
patients almost risk-free and quite precise.

Preoperative Preparation
As with the Chiari I patients, preoperative antibiotics are
given. The anesthetic management and positioning of the
patient are similar to those for the Chiari I patient. Of some
difference, however, is the fact that decompression should
extend to the level of the caudally displaced posterior fossa
tissue. This is frequently below the level of C-4. By removing this additional bone and displacing the musculature,
the risks of cervical deformity are substantially increased
even if the laminectomy is kept quite medial, preserving
the facets. Because the lower portion of the fourth ventricle is usually not within the posterior fossa, the occiput
may need to be removed minimally if at all (Figs. 23A and
23B). If it is elected to open the dura over the posterior
fossa, great care is necessary. The transverse sinus in the
patient with spina bida is frequently placed near if not at
the level of the foramen magnum (Fig. 23B). An unknowing opening of the dura and sinus in this area may well lead
to an operative disaster. The elasticity of the tissues of the
cervical spine is pronounced. In removing the laminal arch
of small infants, each bite with the rongeur needs to be crisp
and clean. Undue distortion of the spinal cord may occur if
this principle is not followed. It is important to study the
preoperative MRI for the position of the fourth ventricle,
the cerebellar vermis, and the possibility of a medullary
kink. The position of all these structures is critical to the
intradural exploration.

4/23/08 3:16:15 PM

Chiari Malformations and Syringohydromyelia


Figure 23 (A) Schematic of incision placement and bone work for operative exposure of a Chiari II malformation. (B) Chiari II malformation,
midline sagittal section. (C) Operative exposure of a Chiari II malformation before and after the oor of the fourth ventricle is exposed.

Operative Procedure
Once the dura is opened, nding the caudal extent of the
fourth ventricle can be difcult (Fig. 23C). Intraoperative
ultrasound may be of help in localizing this structure. The

14535_C02.indd 11

choroid plexus usually maintains its embryonic extraventricular position, marking the caudal end of the fourth ventricle. When present, this is a reliable intraoperative marker.
Unfortunately, dense adhesions and neovascularity at points
of compression or traction may be found, especially near

4/23/08 3:16:15 PM

12 Pediatric Neurosurgery
C-1, and this may make dissection treacherous. The fourth
ventricle may be covered by vermis with its horizontal folia,
or the choroid plexus may simply lie within the displaced
The purpose of the intradural manipulation is to open
the fourth ventricle and provide free egress of CSF from the
fourth ventricle. It is necessary to nd and open the tissue
widely over the caudal aspect of the fourth ventricle. It may
happen that several planes of dissection are developed before the oor of the fourth can be adequately appreciated.
It is important during the exploration of each of these avenues that vascular and neural tissues be preserved and that
natural planes are developed so that no irreparable damage
to the delicate tissues of the lower brain stem occurs. The
caudal aspect of a medullary kink can easily be mistaken
for the appropriate target. This dissection is one of the most
difcult in pediatric neurosurgery. Errors or simple tissue
manipulation may convert a tenuous portion of the medulla
or lower pons to permanently damaged tissue. The surgeon
should always bear in mind the risk-benet ratio for each
of his or her actions, as this particular area is unforgiving of
even small excesses of manipulation. Grafting of the dura
and closure are similar to the previous description.

Postoperative Management Including Possible

Following the operation, patients may experience some
nausea and vomiting as well as hiccups. These are almost
always self-limited. Neurological decits that are well-

14535_C02.indd 12

established prior to the operation are unlikely to reverse

following manipulation. With the advent of MRI scanning
the status of syringohydromyelia can be assessed easily. If a
sizable syrinx persists months to years after craniocervical
decompression, a second decompression should be considered. If after two decompression a larger syrinx is persistent and symptoms attributable to this lesion are serious or
progressive, consideration can be given to a laminectomy
over the lower aspect of the syrinx and the placement of
a syrinx to subarachnoid shunt or a syrinx to peritoneal
shunt. If a syrinx to subarachnoid shunt is chosen, placement of the distal catheter in the free subarachnoid space
is an important technical maneuver. Catheters can easily be
mistakenly placed in the subdural space without benet to
the patient.

In addition to the avoidance of problems with infection,
hemorrhage, and increased neurological decit, patient selection and the timing of intervention are critical to the successful outcome of decompressing a patient with a Chiari
malformation. Despite what was thought to be appropriate and timely intervention, an alarmingly high percentage
of patients with lower cranial nerve abnormalities treated
surgically eventually progress. This raises the question of
whether the current strict selection criteria are too restrictive and whether less symptomatic infants should be considered for decompression. This area of speculation remains
in dispute.

4/23/08 3:16:18 PM

Unilateral and Bicoronal Craniosynostosis
Kant Y. K. Lin, John A. Jane Jr., and John A. Jane Sr.

Coronal craniosynostosis is dened as the premature fusion

of the coronal suture(s) of the skull; sutural involvement
may be either unilateral or bilateral. Because of the position
of the coronal suture, the consequences of premature fusion
are manifested in the calvaria as well as in the face. As with
all forms of craniosynostosis, compensatory growth occurs
at the adjacent nonaffected suture sites. This is evident in
cases of unilateral involvement at the contralateral coronal
suture, where the contralateral side of the metopic and sagittal sutures, and at the ipsilateral squamosal suture growth
leading to a unilateral anterior plagiocephaly (Fig. 31). In
bilateral involvement, compensation is noted at both squamosal sutures, as well as at the sagittal suture leading to an
overall turribrachycephalic or tower-shaped appearance
(Fig. 32). With the resulting skull shapes being so disparate, operative treatments must be geared toward different
issues, and the two types will be discussed separately.
Coronal craniosynostosis may be associated with elevated
generalized intracranial pressure. The likelihood of this occurrence increases when more than one suture is involved.
Bilateral coronal craniosynostosis in often associated with
craniosynostosis syndromes, such as Crouzons or Aperts
syndrome. A distinction must be made between the syndromic and nonsyndromic varieties as management and
expectations of outcome differ between the two.

Patient Selection
Diagnosis is based on the characteristic medical history and
physical examination. Conrmation and more precise delineation of the dysmorphology are obtained from computed
tomography of the skull. In particular, three-dimensional
reconstruction of the images is useful for presurgical planning. A thorough ophthalmologic examination is indicated
both for purposes of detecting intracranial hypertension, as
well as to document any orbital axis issues related to the
changes in the bony orbit secondary to the stenotic adjacent
coronal suture. Often an eyelid ptosis or extraocular muscle
imbalance is seen and must be addressed, usually after the
bone deformities are corrected. Increasingly sophisticated
DNA mapping techniques have resulted in an additional
method of diagnosis that is especially useful with inherited
forms of coronal craniosynostosis, or when a craniosynosotosis syndrome is involved.

Indications and Timing of Surgery

Indications for surgical correction of a unilateral deformity
are improvement of overall skull shape with advancement
of a recessed forehead and correction of the bossed contralateral forehead, correction of the orbital dysmorphology,
which can subsequently allow for correction of the orbital
adnexal structures, and possible relief of either generalized
or localized intracranial hypertension. Indications for surgical correction of the bilateral deformity are similar but also
include the need to correct the overly high or tower-shaped
skull and the overall brachycephaly.
Diagnosis of intracranial hypertension can be difcult
and is based on soft ndings such as cerebral markings
seen on the inner calvarial table (copper-beaten appearance) on plain x-rays, or by late fundoscopic changes seen
by slit lamp evaluation. Earlier signs suggestive of increased
pressure may be seen with subtle behavioral changes in
the child, or with a bulging anterior fontanelle. In the nal
analysis, a monitoring bolt is needed to accurately record
pressures intracranially. Any evidence of elevated pressures
is an indication for a more urgent need for surgery.
Although controversial, most surgeons would agree that
surgery is best performed before the child has reached the
age of 1 year. Our tendency is to perform surgery closer to 6
months of age. Because the volume of the brain almost triples in the rst year of life, it would seem prudent to allow
the intracranial cavity to accommodate this rapid growth
through earlier surgery. The correction should certainly be
performed before brain damage has occurred; often, subsequent brain growth can be utilized to help direct future
growth and maintain the newly corrected skull and orbit
shape following suture release and bony recontouring. Earlier correction also spares the child emotional or psychological trauma over his/her appearance, before the age of
self-awareness (5 years or younger).

Preoperative Preparation
Once the decision has been made to proceed with surgery,
a preoperative workup consisting of routine blood tests, including a complete blood cell count, electrolyte panel, and
a pro-time and prothrombin time, are performed. Because
of the potential for signicant blood loss, a type and screen


14535_C03.indd 13

4/11/08 11:31:20 AM

14 Pediatric Neurosurgery
are obtained and compatible donors among relatives are
encouraged to donate for donor-directed intra- and perioperative transfusions.
The child is brought to the hospital on the day of surgery
having been kept NPO for 4 hours prior to the anticipated
start time for surgery. At least two large-bore (20 gauge)
intravenous lines are required for access due to the potential for signicant blood loss or uid shifts during surgery.
An arterial line is placed and a central line is also helpful to
monitor the total body intravascular volume for both operative and postoperative uid management. A Foley catheter is
useful to record urinary output, and a thermistor is used to
record core body temperature. A Doppler monitor is placed
over the heart to monitor blood ow and is used to detect
the possibility of unanticipated intraoperative air embolism.
Steroids and anticonvulsants are not routinely used. Prophylactic antibiotics are given just prior to the incision.
In young children, the hair is clipped to allow the surgeon
full visualization of the degree of the skull deformity, so that
the surgical correction can be tailored accordingly. This also
helps facilitate the scalp closure and postoperative wound
care by the nursing staff and the parents.
Once the intraoral endotracheal intubation has been performed, we have found it helpful to secure the tube with
either a circummandibular or a circumdental wire, thus obviating the need for taping and allowing full access to the
face during surgery. Temporary tarsorrhaphy sutures are
also placed for intraoperative corneal protection.

Operative Procedure

Figure 31 Skull deformity in unilateral coronal synostosis. The ipsilateral forehead is attened and the superior and lateral orbital rims are
recessed. (A) Compensatory growth (depicted by arrows) occurs at adjacent sutures. Compensatory growth at the metopic and contralateral
open coronal sutures causes unilateral frontal bossing. Growth at the
sagittal and open coronal sutures leads to a contralateral parietal bulge.
(B) Skull-base deformity along the anterior cranial fossa also occurs.

14535_C03.indd 14

The patient with unilateral coronal synostosis is placed in a

supine position on the operating table with the head resting in
slight extension in a Mayeld headrest. The patient with bilateral coronal synostosis is positioned differently, the details of
which will be discussed separately from the unilateral deformity. The headring is reinforced with additional soft padding
to prevent excessive pressure over bony prominences during
the lengthy procedure. The scalp and face are prepped with
Betadine (providone-iodine) solution, with emphasis placed
on scrubbing the external auditory canals, which tend to colonize with bacteria. The head, face, and neck are then draped to
the clavicles, and staples are used to secure the drapes. A 180
degree access to the head and facial region is required, and
the surgical table is rotated so that the anesthesiologist is positioned at the patients side at the foot level. The nursing staff
and all instrumentation, which has been placed on a single
large table, is positioned opposite the anesthesiologist, who is
also at the foot of the bed. A smaller Mayo stand is positioned
over the patients abdomen, and only those instruments most
currently in use are kept for ready access.

Skin Incision and Flap Elevation

A standard wavy bicoronal incision is performed extending
from just behind one ear across to the opposite side. Care is

4/11/08 11:31:22 AM

Figure 32 Skull deformity in bilateral coronal synostosis. Bilateral
coronal synostosis leads to signicant bilateral forehead attening (with
a decrease in the overall anteroposterior dimension of the skull) and re-

taken to make the incision posterior to the anterior hairline,

yet forward enough to allow for access to the orbital region
once the scalp ap is dissected. The scalp ap is elevated
anteriorly down to the level of the supraorbital rim. The
supraorbital neurovascular bundle is preserved and may
occasionally need to be freed from its foramen with a thin
osteotome. Dissection is then extended laterally down each
lateral orbital rim detaching the lateral canthi to the junction with the inferior orbital rim, and medially up to, but
not detaching, the insertion of the medial canthal tendons.
The nasolacrimal apparati are also carefully preserved. The
nasion is exposed during this part of the dissection as well.
Inferolaterally, the anterior aspect of the maxilla, the malar
eminence, and the anterior aspect of the zygomatic arch are
also exposed. The temporalis muscles are elevated off their
insertions and left attached to the undersurface of the scalp
ap, thus allowing access to the infratemporal hollow. The
temporal and sphenoid bones are exposed from the lateral
orbital rim close to the junction where the zygomatic arch
meets the posterior temporal bone. This area will allow the
formation of a tenon extension, once the orbital osteotomies are performed, of the supraorbital bone unit that will
be advanced and reshaped to compensate for the temporal
narrowing seen in this condition.

14535_C03.indd 15

Unilateral and Bicoronal Craniosynostosis


cession of the orbital rims. Compensatory growth (depicted by arrows)

at the squamosal suture causes vertical elongation of the skull. Growth
at the sagittal suture causes the skull to widen.

Craniotomy and Craniofacial Reconstruction

Unilateral Coronal Craniosynostosis
Emphasis has been placed on the concept that despite unilateral sutural involvement, the deformity is, in almost all
cases and occasionally signicantly, bilateral. Whereas the
ipsilateral side reects growth restriction, the contralateral
side exhibits the effects of compensatory changes driven by
the growth of the brain.
A bifrontal craniotomy is performed with the posterior
extent of the cuts being posterior to both the fused and nonfused coronal sutures and the anterior cut ~1 cm above the
level of the supraorbital rims. Retraction of the frontal and
temporal lobes of the brain is then performed, taking care to
remain anterior to each olfactory bulb. Three-quarter orbital
osteotomies are completed across the orbital roof, superior
aspect of the medial orbital wall, lateral orbital wall, and
the lateral aspect of the orbital oor into the inferior orbital
ssure. Tenon extensions are made extending laterally into
the sphenoid and temporal bones. The nal remaining cut
is made across the nasion just above the nasofrontal suture
(Fig. 33). This forms a single orbital unit that is removed
in its entirety to be reshaped. The remaining portion of the

4/11/08 11:31:26 AM

16 Pediatric Neurosurgery

Figure 33 Craniotomy and orbital rim osteotomies for unilateral

coronal synostosis. The surgery begins with a bifrontal craniotomy
that includes both coronal sutures. Radial osteotomies are performed,
and the frontal bone is recontoured. Bilateral three-quarter orbital osteotomies are then performed, elevating the visor as a single unit.
Dotted lines depict areas of osteotomies.

Figure 34 Orbital rim reconstruction in unilateral coronal synostosis.

During the reconstruction, the ipsilateral superior and lateral orbital
rims are advanced and reshaped to match the contralateral side. The
contralateral orbital rim often needs to be recessed by removing a portion of the contralateral tenon. A template of bicortical graft is then
placed over the contralteral orbital rim and is used as an onlay graft for
the ipsilateral orbital rim. Further reshaping of the ipsilateral rim often
requires a combination of burring down the inner cortex of the orbital
rims, thus softening them enough to use the Tessier bone benders.


abnormally shaped and positioned greater wing of the sphenoid bone is then carefully rongeured medially up to the
fused frontosphenoid suture, and into the superior orbital
ssure. This will allow for subsequent brain expansion behind the newly congured orbital unit.
The goals for reshaping the orbital unit include: (1) advancement of the ipsilateral lateral orbital rim; (2) advancement of the retruded supraorbital rim in relationship to the
inferior orbital rim in the anteroposterior (AP) plane; (3)
recreation of the overall shape of the orbit to match the opposite orbit; and (4) recessment of the contralateral lateral
orbital rim to take out any compensatory changes. These
changes are effected by a combination of burring down the
inner cortex of the orbital rims, thus softening them enough
to use the Tessier bone benders to reshape the bone in the
proper conguration. The recessed portion is given additional projection via advancement of the tenon extension
along the lateral temporal bone. A portion of the distal end
of the tenon extension on the contralateral side is removed
to allow for recessing, again at the temporal bone region
(Fig. 34). Finally, the retruded supraorbital rim and the
reshaping of the orbital box are addressed simultaneously
by placing an onlay bone graft, harvested from the bifrontal
bone piece and xed with an absorbable lag screw, over the
decient area and burred to the matching conguration of
the opposite side (Fig. 35).

14535_C03.indd 16

Figure 35 Orbital rim reconstruction and advancement in unilateral

coronal synostosis. The recessed portion is given additional projection
via advancement of the tenon extension along the lateral temporal
bone. A portion of the distal end of the tenon extension on the contralateral side is removed to allow for recessing, again at the temporal
bone region. The onlay graft is xed to the ipsilateral orbital rim with
lag screws. In addition, a portion of the greater wing of the sphenoid
is also removed up to the frontosphenoid suture and into the superior
orbital ssure to allow for subsequent brain expansion into the previously constricted space.

4/11/08 11:31:29 AM

The newly congured orbital unit is then returned to its
original position, albeit advanced on the affected side and
recessed on the opposite side, and secured with 2 mm thick
resorbing plates and screws bridging the tenon extensions
to the adjacent temporal skull. The segment of frontal bone
is also reshaped through a combination of Tessier bone
benders, inner and outer cortex burring, and barrel-stavenlike osteotomies to match the new curve of the supraorbital unit and to recreate a smooth and symmetric forehead.
The segment can be rotated 180 degrees to use the more
properly shaped curve of the posterior edge to match the
curve of the supraorbital unit if needed. This segment can
be secured with either resorbable plates and screws or even
absorbing sutures to avoid any possibility of future growth
restriction or transcranial migration of any xation hardware (Fig. 36).
To prevent early relapse of the deformity, we believe that
rmer rigid xation via plate-and-screw use should be employed but judiciously and only in those areas where signicant postoperative pressure can be expected. Prior to
closure, lateral canthopexies are performed by attaching
the lateral canthi to the orbital rim with permanent sutures
anchored through drill holes in the bone.

Figure 36 Final reconstruction in unilateral coronal synostosis. The

orbital rims have been reconstructed using an onlay graft and by advancing the recessed rim and recessing the advanced rim. The orbital
rims are attached to the parietal bone using absorbable plates. The
segment of frontal bone is also reshaped through a combination of
Tessier bone benders, inner and outer cortex burring, and barrelstaven-like osteotomies to match the new curve of the supraorbital
unit and to recreate a smooth and symmetric forehead. The segment
can be rotated 180 degrees to use the more properly shaped curve
of the posterior edge to match the curve of the supraorbital unit if
needed. This segment can be secured with either resorbable plates
and screws or even absorbing sutures to avoid any possibility of future
growth restriction or transcranial migration of any xation hardware.

14535_C03.indd 17

Unilateral and Bicoronal Craniosynostosis


Bilateral Coronal Craniosynostosis

The problem with the bilateral deformity is twofold: rst,
the height of the skull and the recession of both supraorbital
rims and lateral orbital rims and second, the brachycephaly
that presents a signicant problem because it is difcult to
correct and failure to correct it will compromise the overall result (Fig. 32). The issue is whether its correction is
warranted in every instance. This can be addressed in two
ways: (1) osteotomy and advancement of the single orbital
unit consisting of both orbits and the supraorbital bar, as
previously described; and (2) expansion of the entire cranial
base region that allows for a downward settling of the top
portion of the skull along the vertex, thus reducing overall
skull height. It is this second goal that necessitates a change
in the operative positioning from that used in the unilateral deformity. The patient is placed in a modied prone
position, the so-called sphinx position, to correct both the
frontal and height abnormalities. Before placing the patient
in this position, however, it is important to assess the stability of the cervical spine and the craniovertebral junction by
preoperative lateral cervical spine roentgenograms in exion and extension. Positioning the patient on the operating
table is greatly aided by a vacuum-stiffened bean bag to
mold the upper body and neck. The face and arms are padded with thick, cushioning foam to prevent pressure sores
and compression nerve palsies (Fig. 37).
Bur holes are placed in the pterion regions bilaterally, and
parasagittally in the anterior parietal bone, just posterior
to the coronal suture. Similarly, a biparieto-occipital bone
graft is outlined with multiple bur holes adjacent to the
sagittal and transverse sinuses. Once the bone is elevated
both frontally and parieto-occipitally, further dissection epi-

Figure 37 Operative positioning for bilateral coronal synostosis. The

patient is placed in a modied prone position, the so-called sphinx
position, to correct both the frontal and height abnormalities. Before
placing the patient in this position, however, it is important to assess
the stability of the cervical spine and the craniovertebral junction by
preoperative lateral cervical spine roentgenograms in exion and extension. Positioning the patient on the operating table is greatly aided
by a vacuum-stiffened bean bag to mold the upper body and neck.

4/11/08 11:31:31 AM

18 Pediatric Neurosurgery
durally may be performed below the level of the transverse
sinus to allow the surgeon to fracture outwardly (in the
posterior direction) the occipital bone. The outfractures or
barrel-staven-like osteotomies increase the bony capacity
by enlarging the perimeter of the skull locally. This allows
later brain and dural displacement or settling into this region as the height of the skull becomes reduced secondarily
by gravitational forces (Fig. 38). Barrel-staven-like osteotomies in the occipital bone in the midline and paramedian
regions are longer than those placed further laterally, to
achieve elongation along the AP axis of the skull, without
further widening of the parieto-occiput. The thickened and
abnormally elevated superior portion of the greater wing
of the sphenoid bone is removed by rongeur in a manner
similar to that described for the unilateral deformity. The
abnormally convex squamous portion of the temporal bone

Figure 38 Craniotomies and osteotomies for bilateral coronal synostosis. The craniotomized segments include midline biparietal, bifrontal,
and bioparieto-occipital aps. Once the bone is elevated both frontally
and parieto-occipitally, further dissection epidurally may be performed
below the level of the transverse sinus to allow the surgeon to fracture
outwardly (in the posterior direction) the occipital bone. The outfractures or barrel-stavenlike osteotomies increase the bony capacity by
enlarging the perimeter of the skull locally. The barrel-staven-like osteotomies in the occipital bone are longer than those placed further laterally, to achieve elongation along the anteroposterior axis of the skull.
The abnormally convex squamous portion of the temporal bone is left
in place but is addressed in a similar manner with barrel-staven-like osteotomies. The bilateral three-quarter orbital osteotomies are elevated,
reshaped, and advanced.

14535_C03.indd 18

is left in place but is addressed in a similar manner with

barrel-staven-like osteotomies. The bone is reshaped and
straightened by radial osteotomies into the center of the
convexity and controlled outfracturing of the bone segments
with Tessier bone benders. By leaving the temporal bone
in place, the overall stability in the lateral portions of the
skull is increased. The craniotomized frontal and parietooccipital bones are recongured using similar bone-shaping
techniques before being returned to their orthotopic positions. These are then held loosely in their desired positions
by xation with absorbable sutures to the rigidly xed advanced orbital unit anteriorly, to the radially cut temporal
bone laterally, and to the outfractured occipital base bone
posteriorly. The vertex of the infant, composed primarily of
the brous anterior fontanelle, is loosely reattached at the
top of the skull with absorbable sutures to the adjoining
bones, purposely avoiding any downwardly directly external
cinching (Fig. 39).
We believe that as a consequence of the expansion of the
skull-base perimeter, gravitational forces will act to settle
the brain and dural into a more inferior overall position,
thus having the effect of decreasing the overall height of
the skull. This obviates the need for any deliberate downward cinching of the skull vertex as previously described,
which did occasionally lead to adverse effects on cerebral

Figure 39 Final reconstruction in bilateral coronal synostosis. The

craniotomized bifrontal and biparieto-occipital bones are recongured using similar bone-shaping techniques before being returned to
their orthotopic positions. These are then rigidly xed to each other
and held in their desired positions with absorbable sutures to the rigidly xed advanced orbital unit anteriorly, to the radially cut temporal
bone laterally, and to the outfractured occipital base bone posteriorly.
The vertex of the infant, composed primarily of the brous anterior
fontanelle, is loosely reattached at the top of the skull with absorbable
sutures to the adjoining bones, purposely avoiding any downwardly
directly external cinching.

4/11/08 11:31:35 AM

The operative eld is copiously irrigated to remove all
nonviable debris and bone dust that could act as a nidus
for future infection. Because the temporalis muscles were
never detached from the overlying scalp ap, there is no
need to reattach the muscles to their insertions, as they will
naturally return to their proper position as the scalp ap is
brought posteriorly.
The scalp ap is reapproximated and closed in a two-layer
fashion with buried absorbable sutures both in the subgaleal plane and on the skin. A snug but nonconstricting
dressing using a sterile surgical towel wrapped as a turban
is placed over the entire head, and the head is kept elevated
at all times. The tarsorrhaphy sutures and the wire around
the endotracheal tube are then removed.

Specialized Instrumentation
A signicant advance in instrumentation for pediatric craniofacial surgery occurred with the introduction of resorbable
plate-and-screw xation hardware. These new biomaterials
consisting of polymers of polylactic acids are designed to be
totally resorbed within 9 to 15 months following implantation. Studies have conrmed that they have tensile strength
properties comparable to previously used metallic hardware
at the time of their initial use, which allows for the same
adaptability when used in three-dimensional calvarial reconstruction. Due to the relatively long retention lifespan
of the xation hardware, the recongured and repositioned
bone segments can heal with additional support from the
plates and screws. Any concern about future translocation of
the hardware or restriction of further craniofacial growth is
obviated because of predictable resorption of the hardware.
High-speed drill and saw systems have made the surgery
technically easier.

Postoperative Management Including

Possible Complications
Immediate postoperative care is given in the pediatric intensive care unit. Vital signs and central venous pressures are
monitored, and laboratory values are obtained for review of
evidence of hypovolemia and blood loss. Depending on the
degree of blood decit, blood transfusions are frequently
necessary and the concomitant replacement of coagulation
factors may be required. Appropriate dosages of analgesia
are given intravenously to keep the patient comfortable. The
child is positioned on his/her back in bed to keep pressure
off the forehead and orbits, and the head is elevated at 30
degrees to prevent excessive postoperative swelling. As a result of the periorbital manipulations, the eyelids commonly
swell shut within 24 hours following surgery. This can cause
anxiety for both the parents and the child, and reassurance

14535_C03.indd 19

Unilateral and Bicoronal Craniosynostosis


must be given that the majority of the swelling will resolve

in 3 to 4 days.
In most instances, the child is ready for transfer to the
regular ward within 24 to 48 hours. Usually, the child will
run a slightly elevated temperature between 38C and 39C
at day 3 or 4 following surgery. A routine fever workup is
performed, but only rarely is the fever caused by infection.
The child is considered stable for discharge from the hospital
once the child is tolerating a regular diet and eyelid swelling has resolved enough to permit eye opening. A follow-up
visit is arranged for 1 week after hospital discharge.

Early postoperative complications arise from intraoperative
or perioperative blood loss, which can be compounded by
inadequate blood replacement. Tears in the sagittal sinus
during the craniotomy can have immediately devastating
consequences and must undergo repair quickly. Meticulous
attention must be directed when looking for possible tears
in the dura mater, which can lead to a persistent cerebrospinal uid leak. Once recognized, the tear can be easily
repaired with a single absorbable suture of 40 Nurolon. An
air embolism is also a possibility, particularly if the patient
is slightly volume-depleted and a sinus is inadvertently entered. This can be detected by the precordial Doppler ultrasound, as well as by the end-tidal volume gas spectrometer.
Treatment includes placing the patient in a Trendelenburg
position and ooding the eld with saline to prevent further intake of air into the circulation. A small amount of air
is usually tolerated with minimal deleterious effect. Injury
to the brain itself, as well as the globes, can occur if proper
precautionary measures are not taken during the osteotomies, although this risk remains low with an experienced
craniofacial surgical team. Pressure sores must be avoided
by attention to and protection of the areas where the skull
rests during the procedure.
The most frequent late complications include infection
and recurrence or relapse of the original defect due to suture
restenosis. To prevent infection postoperatively, strict sterile
technique must be adhered to and prophylactic antibiotics
used up to the time of removal of the drain. If osteomyelitis
occurs, there is a high risk of loss of the bone grafts.
This chapter is a revision of the chapter, Treatment of Unilateral or Bilateral Coronal Synostosis by John A. Persing,
M.D., and John A. Jane, Sr., M.D. The chapter appeared in the
Neurosurgical Operative Atlas, Volume # 1, edited by Setti S.
Rengachary and Robert H. Wilkins. The Neurosurgical Operative Atlas was published by the American Association of
Neurological Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank John A. Persing,
M.D., for his help and efforts on the original chapter published in the rst edition of this work.

4/11/08 11:31:39 AM

Transoral Surgery for Craniovertebral
Junction Abnormalities
Arnold H. Menezes
Abnormalities of the craniovertebral junction (CVJ) have
been documented for many centuries with postmortem examination. The antemortem recognition of these lesions, in
the past ve decades, has now stimulated surgical therapy.
Several approaches to the anterior craniovertebral border
have been developed. Of these, the transoral-transpalatopharyngeal route is the most frequently used for decompression of the ventral cervicomedullary junction. This author has performed more than 700 of these procedures and
shares his experience with the readership. The advances in
microsurgical instrumentation and neurodiagnostic imaging have expanded the use of this approach.

Patient Selection
Bony abnormalities of the craniovertebral junction can be
divided into those that are reducible and the irreducible
categories. The primary treatment for reducible CVJ lesions is stabilization. Surgical decompression of the ventral
cervicomedullary junction is performed when irreducible
pathology is encountered. This decompression must be performed in a manner in which encroachment has occurred.
The transoral-transpalatopharyngeal approach is utilized
when a lesion is ventrally situated. The other possible approaches to the ventral CVJ are the lateral extrapharyngeal
and the maxillary dropdown procedures. In lateral compressions, a lateral or posterolateral approach is utilized.
When dorsal compression is evident, a posterior approach
is made. If instability exists following any of the situations,
a posterior xation becomes mandated. Thus, the factors
that guide the surgical approaches to the lesions of the CVJ
are: (1) reducibility of the lesion; (2) the direction of the
encroachment; and (3) the type of the lesion.

Preoperative Preparation
The relevant diagnostic imaging for treating abnormalities
at the CVJ consists of plain radiographs of the region that
include the skull and cervical spine. Dynamic studies in the
extended and exed positions assess stability and possible
reduction. Computed tomography (CT) of the CVJ is an integral part of the assessment. This should be augmented
with three-dimensional CT reconstructions to provide an
understanding of the location of the skull base as well as


the atlas and the axis and the odontoid process in particular. The latter, the odontoid process, is usually smaller than
in the normal adult, when combined with segmentation
Magnetic resonance imaging (MRI) is now the mainstay
in neurodiagnostic imaging of the craniocervical junction
(Fig. 41). This should be performed with thin-section midsagittal images in both the T1 and T2 modes and should
include the dynamic exed and extended positions. It provides information about the relationship to neural structures, the osseous abnormality, and the vascularity. Magnetic resonance angiography (MRA) is an essential part of
the investigation, especially in patients with neurological
dysfunction that cannot be explained on the basis of the
previous studies. This is done with the patient in the exed
and in the extended positions as well as in a rotated position to look for vascular occlusions that can change with
position. It is crucial to attempt reduction in children. This
is because in 80% of children with atlantoaxial dislocation
or basilar invagination below the age of 12 to 14 years, a
reduction is possible so as to relieve compression on the
neural structures and thus avoid the ventral procedure. In
that circumstance where a reducible lesion is documented,
stabilization is the operation required.

Assessment of Nutritional Status and Dental Hygiene

This is particularly important in children who present with
brain stem dysfunction or a failure to thrive. Preoperative
nutritional support is mandatory. Dental hygiene is assessed
to eradicate causes of operative bacterial contamination.
Dental guards are also fashioned to protect the upper and
lower dentition during the oncoming operative procedure.

Assessment of Co-Morbidities
It is necessary to assess pulmonary function as well as to
assess sleep apnea in those children who have obvious brain
stem dysfunction or abnormalities that pertain to the glossopharyngeal, vagus, and hypoglossal nerves. This may require a tracheostomy to be performed before the start of the
operative procedure. However, in the authors experience in
the past 15 years, a tracheostomy has not been performed
preoperatively because postoperative improvement is the
rule. On the other hand, two patients have had to have a tra-

Transoral Surgery for Craniovertebral Junction Abnormalities


procedure and similarly Peridex (chlohexidine gluconate)

It is necessary to recognize that the entrance to the oral
cavity must provide a working distance of 2.5 to 3 cm between the upper and lower incisor teeth. This span may not
be available in children with juvenile rheumatoid arthritis
and with involvement of the temporomandibular joint, who
have ventral cervicomedullary compression. There were
four instances of children with segmentation abnormalities
of the cervical spine and acute kyphosis of the craniocervical junction that necessitated mandibular split with midline
glossotomy due to small oral access.

Preoperative Cervical Traction


Figure 41 (A) Composite of mid-sagittal T2- and T1-weighted magnetic resonance imaging (MRI) of posterior fossa and cervical spine. This
8-year-old girl presented with difculty swallowing, sleep apnea, and
distal arm weakness. Note the ventral bony abnormality of the clivusodontoid indenting into the ventral mid-medulla oblongata. There is
tonsillar ectopia through the foramen magnum, and the cervicomedullary buckle is behind the body of C-2. A syringohydromyelia is evident
at the cervicothoracic junction. (B) Composite of T2- and T1-weighted
MRI of the craniocervical junction and cervical spine of patient in Fig.
1A. These postoperative images were made a year following transpalatopharyngeal resection of the lower clivus and the odontoid process
with medullary decompression. Note the collapse of the cervicothoracic
syrinx in addition to the medullary decompression.

Skeletal traction is usually applied through an MRIcompatible crown halo device so as to assess the reducibility of the bony lesion. In the pediatric patient between
the ages of 8 and 16 years, this is instituted 4 days prior to
the planned anterior and posterior surgical procedures at
the craniocervical junction. Should the lesion be reducible, a
dorsal xation is made. On the other hand, if it is irreducible,
both the ventral and dorsal procedures are performed in the
same anesthetic. In situations where an irreducible lesion
is dened at the initial assessment, the crown halo traction
is placed only after the induction of general anesthesia to
stabilize the CVJ.
In children between the ages of 2 and 8 years, it is not
feasible to do preoperative traction. This author has utilized intraoperative traction and uoroscopy to document
irreducibility in the very young child. The same principles
regarding reducibility apply.
The transoral-transpalatopharyngeal route to the craniocervical border has been safe and has provided the author
with a direct approach to this region. It has been utilized between 1977 and 2005 in more than 700 patients (Fig. 42).
The infection rate for the operation is less than 1%. A dorsal
occipitocervical fusion has been necessary in all children.
The indication for ventral transoral-transpalatopharyngeal
approach to the craniocervical border is irreducible ventral
bony abnormalities (including pannus) with compression of
the cervicomedullary junction. This indication is reserved
for extradural bony and soft tissue masses and a few intracranial-intradural tumors. Preoperative antibiotics consist
of 1 g of penicillin G started 2 hours before the commencement of the transoral procedure.


Operative Procedure
cheostomy after failing weaning from the endotracheal intubation 10 days after the ventral decompressive procedure.

Oropharyngeal Cultures
Three days prior to the operation oropharyngeal and nasal
cultures are obtained. No antibiotics are instituted if normal
oral ora is present. Mupirocin nasal ointment is utilized
in the nasal passages twice a day prior to the operation.
Nystatin rinses are instituted 3 days prior to the operative

The pediatric patient is brought into the operating suite

with a cervical collar in place as a precaution for the intubation maneuvers as well as for positioning. Topical
oropharyngeal and nasopharyngeal analgesia is utilized to
facilitate the ber-optic endotracheal intubation. An awake
ber-optic intubation is made in the older pediatric patient.
In the young child, general endotracheal and intravenous
anesthesia is then followed by a ber-optic endotracheal
intubation, which is accomplished through the mask. The

22 Pediatric Neurosurgery

Figure 42 Illustration of rostral-caudal exposure (between arrows)

obtained via the transoral operation.

position of the patient during intubation is judged by the

preoperative dynamic studies. In the awake patient following intubation, the patient is positioned for the operative
procedure and examined awake to check the neurological
status. General anesthesia is then administered. The author
avoids nasopharyngeal intubation because it disrupts the
integrity of the high nasopharyngeal mucosa and obstructs
the operative eld.
The patient is now positioned supine on the operating
table with the head resting on a padded headrest, with mild
extension with cervical traction in place at 7 lb weight, in
the older child. In children between the ages of 5 and 8
years, this weight should be reduced to 4 to 5 lb. The oral
endotracheal tube is now secured to the skin over the middle half of the mandible on the left with sutures and adhesive tape. The nasal passages are topically anesthetized
with 4% cocaine and are then packed with nasal paddies.
The laryngopharynx is occluded with a throat pack, and the
cleansing of the oral cavity and pharynx now commences.
This is done with successive rinses of 10% povidone-iodine
followed by normal saline and then hydrogen peroxide and
subsequently the saline rinse. The dental guards are now
used over the upper and lower dentition to protect them

during the operative procedure. The author also prepares

the right anterior abdominal wall for possible harvesting of
donor external oblique aponeurosis and fat should this be
necessary if the dura is opened during the transoral procedure. The circumoral area as well as the anterior abdominal
wall are prepared and draped in a sterile fashion.
A Dingman mouth retractor with an incorporated tongue
blade is utilized for automatic exposure. In circumstances
where the operative procedure is at the foramen magnum
and above, it is necessary to split the soft palate and at times
the hard palate. On the other hand, if the procedure is limited to the level of the atlas and axis vertebrae, the soft palate may be elevated by catheters attached to the soft palate
via the nasal passages and secured to either side of the soft
palate and then withdrawn into the high nasopharynx to
allow for exposure. A gauze pack occludes the laryngopharynx. Lidocaine solution (0.5%) with 1:200,000 epinephrine
is injected into the median raphe of the soft palate. The microscope is now used for magnication and a concentrated
light source, and to aid in the operative dissection.
The soft palate incision (when performed) starts at the
base of the uvula to one side of the midline and ascends
into the median raphe within 1 cm to then proceed up in the
midline to the hard palate (Fig. 43A). Stay sutures provide
for lateral retraction and are held in place via the springs of
the Dingman mouth retractor. Should exposure be necessary into the lower clivus, an extension of the hard palate
incision is made in the midline and a subperiosteal exposure
of the midline hard palate is then accomplished. Topical
cocaine is now used to anesthetize the posterior pharyngeal
wall. The median raphe of the pharynx is then inltrated
with 0.5% lidocaine solution with 1:200,000 epinephrine. A
midline incision is made in the posterior pharyngeal median
raphe extending from the rostral clivus to the C-3 level (Fig.
43B). The leaves of the posterior pharyngeal wall are now
reected laterally and folded upon themselves with stay
sutures to prevent damage to the orices of the eustachian
tubes and to allow for exposure. The prevertebral fascia
and longus colli muscles are reected free of their osseous
ligamentous attachments to expose the caudal clivus and
the ventral surface of the atlas and the axis vertebrae (Fig.
43C). The anterior longitudinal ligament and the occipital
ligaments are dissected free of their bony attachments to expose further the caudal clivus, the anterior arch of the atlas,
and the anterior surface of the axis body (Fig. 43D). This
provides an exposure of 3 to 3.5 cm in width. One should
avoid an attempt at further lateral exposure because this
brings a risk of damage to the eustachian tube orices as
well as the vertebral arteries and their entrance into the vertebral canal and possible injury to the hypoglossal nerve.
The anterior arch of the atlas is now removed with a highspeed drill for a width of ~20 to 25 mm (Fig. 43E). The
author utilizes a 4 mm cutting bur that is then changed to
a diamond bur using an electric drill. In situations where
the odontoid process has invaginated into the posterior
fossa, the inferior edge of the hard palate may need to be
resected. This latter procedure is accomplished with Kerrison rongeurs.
If the inferior portion of the clivus needs to be removed,
it is done by thinning the ventral aspect with a diamond bur

Transoral Surgery for Craniovertebral Junction Abnormalities




Figure 43 (Continued on 24)

24 Pediatric Neurosurgery

Figure 43 (A) Illustration of exposure of the oral cavity and the pharynx with the Dingman mouth retractor in place. The soft palate has
been incised exposing a portion of the hard palate at the apex of the
incision. The view through the operating microscope is within the circle. (B) Drawing and view through the microscope. The soft palate is
retracted laterally with stay sutures and the posterior pharyngeal wall
has been incised. (C) The longus colli muscles as well as the posterior
pharyngeal wall are retracted laterally to expose the fascia overlying
the anterior arch of the atlas as well as the body of C-2. (D) The dens
is revealed when the anterior arch of the atlas is removed. The caudal
clivus is exposed with the invaginating odontoid process underneath
the caudal clivus. (E) The shell of the odontoid process is removed
along with a portion of the axis body. (F) The cored out odontoid
process and shell are being removed with curettes and rongeurs. (G)
The cruciate ligament is visible after removal of the odontoid process.
Dura now lls the area of decompression.

and dening the inferior border. It should be separated from

the circular venous sinus and the dura as well as the tectorial membrane, which is attached to the posterior aspect
of the clivus. The author prefers the use of ne Kerrison
rongeurs for clivus resection after the thinning out has been
done to prevent inadvertent entrance into the dura or the
circular sinus.
The odontoid process removal is accomplished by starting
at the tip and then proceeding downward in a rostral-caudal
manner. The odontoid process is rst cored out and subsequently the lateral margins are dened using ne curettes
and rongeurs. It is essential to sharply divide the apical ligament as well as the alar ligaments with curettes (Fig. 43F).
Resection of the odontoid process should proceed into the
body of the axis dictated by preoperative diagnostic studies,
which must be referred to during the operation. Fine up-biting Lee-Smith Kerrison rongeurs with 1 mm footplates are
used as well as microcurettes and microbiopsy forceps to facilitate removal of the odontoid process. Removal of pannus
from behind the odontoid process (especially in unstable
situations when it is present) and from the posterior fossa
is done in a piecemeal fashion with bipolar cauterization
with angled and ringed curettes. The lateral extent of the
exposure should encompass the amount of decompression
necessary, dictated by the preoperative CT and MRI.

At this point, the cruciate ligament should be visualized

(Fig. 43G). In such situations when it is frayed, such as with
tumor, it should be removed. Otherwise, this author has preferred to leave it in place. Likewise, the tectorial membrane
must be preserved. The cervical traction is maintained during the operation for inherent or potential instability. In
circumstances where pannus or granulation tissue acts as
a mass, especially in children, this is removed with bipolar
cauterization and piecemeal reduction. It is not necessary
to remove the entire granulation tissue because this will
ultimately brose and aid in the healing process.
Aerobic and anaerobic bacterial cultures are obtained
from the depths of the wound at the end of the decompression and resection. The author has utilized a combination
of bacitracin powder and microbrillar collagen to layer
the resection bed. The longus capitis and the longus colli
muscles are now approximated in the midline with interrupted sutures of 30 polyglycolic using a gure-of-eight
technique. A similar strength suture is utilized to bring together the pharyngeal constrictor muscles and separately
the posterior pharyngeal mucosa.
In young children there is exuberant adenoid tissue
present. The mucosa over the clivus is quite delicate and
friable, and attempts to bring this together will only cause
further bleeding. In this circumstance, large cerclage sutures
are made and compressed Avitine is placed over the area. A
blanket of Gelfoam soaked in thrombin is then placed over
the reconstituted pharyngeal wall. The throat pack is now
A nasopharyngeal-gastric feeding tube is now passed via
the nostril into the pharynx and under direct vision into the
esophagus and subsequently into the stomach (Fig. 44).
The position of this tube is conrmed by insufation of air
into the tubing and the anesthesiologist auscultating over
the abdomen. This tube is then secured to the ala of the
The closure of the soft palate must be in a layered fashion.
It is done by bringing together the nasal mucosa together
with interrupted 30 polyglycolic suture. The muscular layer
as well as the oral mucosa of the soft palate are approximated
with interrupted vertical mattress sutures of similar strength.
Tension in these sutures must be carefully assessed so as not
to be too snug, which leads to dehiscence.
A dorsal occipitocervical fusion combined with posterior
fossa decompression is usually mandated and accomplished
with the same anesthetic.

Special Circumstances
Transpalatal Route
In congenital abnormalities where the underlying pathology consists of a foreshortened clivus, the clivus-odontoid
articulation is high and the clivus itself tends to be more
horizontal than vertical. In this circumstance, removal of
the posterior 1 cm of hard palate in the sagittal dimension
may be necessary, and the width of this resection is ~8 mm
to either side of the midline to allow for exposure of the
high nasopharynx and the clivus. The posterior edge of the

Transoral Surgery for Craniovertebral Junction Abnormalities


Postoperative Management Including

Possible Complications

Figure 44 The photomicrograph through the operating microscope

after closure of the longus colli muscles, the posterior pharyngeal musculature, and the posterior pharyngeal mucosa in a layered fashion. A
nasogastric feeding tube is now visualized in the posterior pharynx.

vomer becomes visualized and may be removed. Closure of

this soft palate is made in a layered fashion as previously
described. This technique allows high nasopharyngeal exposure without splitting the mandible or doing a median

Intradural Extension of Tumor and Repair of CSF Leak

An intradural lesion at the upper cervical spine, craniocervical border, or behind the clivus necessitates preoperative recognition of the dural extension and placement of a
lumbar subarachnoid drain prior to commencement of the
operation. The dural turgidity may be reduced by draining
spinal uid during the operation. The dural incision is made
in a cruciate fashion. The vertical component of this starts
inferiorly, proceeding in a rostral dimension with careful
cauterization of the circular sinus at the foramen magnum.
Once the intradural operation is complete, it is essential to
bring the dural leaves together with 40 polyglycolic suture.
This is then covered with external oblique aponeurosis graft,
which is sutured to the dura as a transxing suture through
the fascia. This latter maneuver prevents migration of the
graft during further closure. Adherence to the dura is conrmed with the application of brin glue. The fascia and
dura are then covered by the fat obtained from the anterior
abdominal wall, after which the posterior pharyngeal closure is made in the described layered fashion.
During and after the operative procedure, CSF tension is
reduced by drainage from the lumbar subarachnoid space.
Triple antibiotic therapy consisting of cefotaxime, metronidazole, and methicillin are continued for the rst 5 days
after operation. If no bacterial ora is identied on CSF examination, the third-generation cephalosporin is discontinued and the antibiotics terminated at the end of 10 days. At
this time the CSF drainage is discontinued.

The endotracheal tube is maintained in place for postoperative convalescence until the swelling in the tongue and
oral tissues has receded. This usually lasts for 3 to 4 days.
It is our policy to maintain a daily caloric intake of 2500 to
3000 calories by the third to fourth postoperative day. The
transpyloric placement of the nasogastric tube facilitates
The postoperative management consists of ventilator and
respiratory care and nutritional support, antibiotic regimen,
pain management, and immobilization.
The vast majority of children who have undergone a transoral-transpalatopharyngeal decompression of the craniocervical border will leave the operating theater with an
endotracheal tube in place secured to the circumoral area,
intravenous and intra-arterial lines that have been required
for intraoperative care, and nasogastric feeding tube. In
addition, the child should have had a dorsal stabilization
procedure at the craniocervical junction. The child is then
transported to the intensive care location with a soft cervical collar in place. The halo ring that was utilized intraoperatively for traction and stabilization is left in place if it is
later to be incorporated into a halo vest.
Intravenous penicillin G is administered for 48 hours
and then discontinued. However, if wound cultures grow
pathological ora, appropriate antibiotics are instituted. The
Peridex and nystatin rinses that were instituted preoperatively are now done at 8-hour intervals to minimize bacterial ora, especially in light of the operative procedure and
the prolonged intubation necessary. The dental guards serve
to prevent inadvertent biting on the endotracheal tube. The
cuffed endotracheal tube is intermittently deated only after proper oral care is accomplished.
The soft tissues of the retropharyngeal space are monitored with lateral cervical radiographs. By the end of the
third or fourth postoperative day, when lingual swelling
has receded and the pharyngeal edema has resolved, an
attempt at endotracheal extubation is made. This should
be performed by an intensivist and with a neurosurgeon in
attendance. Prior to this, nasogastric feedings are discontinued and the endotracheal tube cuff is deated to detect
leakage of air around the tubing. Our policy is to attempt
extubation only with the utilization of a tube exchanger
replacing the endotracheal tube. This smaller diameter tube
exchanger allows for satisfactory ventilation and should this
be possible, the tube exchanger is removed. If not, the tube
exchanger acts as a stylet to allow for rapid reintubation.
The position of the nasogastric feeding tube placed during the operation is conrmed with appropriate chest and
abdominal lms. It is preferable to have the feeding tube in
a transpyloric location. Enteral feedings are instituted in a
graduated manner at the end of 24 hours. A clear oral intake
is initiated by the fth postoperative day, at which time the
oral endotracheal tube had been removed. A full liquid diet
is started by the 10th day. By this time, the patients caloric
intake is between 2500 and 3000 calories. A soft diet is instituted by the end of the 15th day.

26 Pediatric Neurosurgery
Postoperative immobilization is accomplished with a soft
collar until the nal decision is made to use either a halo
vest or an occipitocervical brace. An immediate placement
of a halo vest after the operation hampers postoperative
oral care as well as jeopardizes extubation and respiratory

The complications are divided into perioperative complications of transoropharyngeal surgery and delayed complications.

Perioperative Complications
It is necessary to make sure that the distance between the
incisor teeth is more than 25 mm to accomplish the operation. Otherwise, a mandibular split or a tongue split may be
necessary or another approach contemplated. However, the
situation may improve with the administration of intravenous paralyzing agents during the operation.
The damage to the eustachian tubes and hypoglossal
nerves can be avoided by limiting the lateral exposure from
the midline to <2 cm. In situations where one is unable to
reach the clivus for resection due to platybasia, it is necessary to divide the soft palate and the hard palate, and the
surgeons location is conrmed with intraoperative uoroscopy. Resection of the base of the odontoid process in an attempt to pull down the anchored tip is always fraught with
difculty. Hence, the resection should start from the dens
and then proceed downward. A novice may use frameless
stereotaxy or intraoperative uoroscopy. Persistent bleeding from the circular sinus necessitates the use of hemostatic
agents and even possibly an attempt to clip both the leaves
of the bleeding dural site. Pannus and arterial bleeding must
be controlled with cauterization. CSF leakage should be addressed as previously described.

Delayed Complications of Transoropharyngeal Surgery

Postoperative lingual swelling may be controlled partially
by the use of intravenous Decadron (dexamethasone) dur-

ing the operation and postoperatively. At times, intermittent

release of the tongue depressor helps with relieving venous
stasis as well as lymphatic obstruction in prolonged operations. If meningitis occurs, it is mandatory to obtain CSF and
also place a lumbar drain. During this time, no oral intake
should be allowed and intravenous antibiotics must be instituted. An attempt at closure of the leakage site should
be made. Early palatal dehiscence requires an immediate
attempt at reclosure. A delayed small palatal opening should
be treated with hyperalimentation, and if this does not close
over several weeks, it will require freshening of the opening
and reclosure.
Pharyngeal dehiscence is a dreaded complication. When it
occurs within the rst week, reclosure must be attempted.
If, however, this occurs on a delayed basis after a week, the
patient is treated with hyperalimentation and intravenous
Neurological worsening requires an assessment of the
alignment to make sure that there is no meningitis, abscess
formation, or retained lesion. Thus, an MRI must be repeated
and possibly an MRA to look for vascular compromise.
A retropharyngeal abscess requires drainage in an extra-pharyngeal manner and not through the mouth. This is to prevent
stula formation. In circumstances of retropharyngeal abscess,
one must check for meningitis and osteomyelitis.
Delayed pharyngeal bleeding must signal the treating
physician to look for infection, osteomyelitis, and possible
vertebral artery erosion and a false aneurysm. Angiography
is necessary. If the vertebral artery abnormality is recognized, this should be occluded with either coils or a balloon.
Velopalatine incompetence is mostly seen in children and
occurs about 4 to 6 months after operation. It signies brosis of the palate or pharyngeal wall. The treatment for this
has been pharyngeal retraining, the use of a palatal prosthesis, and at times a pharyngeal ap. Persistent hoarseness of
voice 4 to 6 weeks after the operation necessitates visualization of the vocal cords to look for granulomas and polyps.

Malposition of the Orbits
John A. Persing and Bianca I. Knoll

Malposition of the orbits is a congenital deformity that is

frequently seen in craniofacial anomaly centers. There are
two commonly encountered forms of this abnormality involving the mediolateral axis of the orbits.
The rst is telorbitism, which refers to a widened distance
between the two orbits. This is to be distinguished from the
second type, hypercanthorum, in which there is a widened
distance between the medial canthi alone (the distance between the lateral canthi is normal). The widened distance
between the canthi may be due to bone, soft tissue, or both
being displaced laterally. Other forms of orbit malposition
also exist.
Superior and inferior displacements (vertical dystopia)
as well as anterior and posterior displacements (AP dystopia) with or without accompanying rotational components
(rotary dystopia) occur. Dysfunction may include diplopia,
amblyopia, and absence of normal depth perception. The
goal of surgical treatment in these patients, therefore, has
been to create an aesthetic cranioorbitalfacial image while
restoring or maintaining visual function. Only the anomalies
of the mediolateral axis will be discussed in this review.

Patient Selection
The appropriate operative procedure for correction for
the patient with hypercanthorum entails translocation
of the medial portion of the orbits only, whereas in the
patient with telorbitism, both the medial and lateral orbital walls are translocated medially. Patients are chosen for the two different operative approaches based
on the clinical measurement of a widened intercanthal
distance (medial and lateral), plain radiograph and computed tomography scan demonstration of widened bony
intercanthal distance. Surgery is elected usually at approximately 5 years of age, unless a concurrent abnormality such as an encephalocele is present, providing
the opportunity for one-stage treatment at an earlier
age. Age 5 is chosen because this allows for correction of
the abnormality when the bone is sufficiently strong to
avoid inadvertent fracture during surgical mobilization,
the orbit has reached relative maturity, and the child has
not reached school age, so that major deformity can be
corrected, or at least ameliorated, prior to critical peer
interaction in school.

Preoperative Preparation
The risks of the operative procedure include potential injury
to the brain and visual system, including enophthalmos,
extraocular muscle entrapment, optic nerve damage, cerebrospinal uid leakage, and recurrence of the intercanthal deformity postoperatively. Translocation of the orbits
medially in early childhood may negatively affect midfacial
growth, particularly if the osteotomy is signicantly below
the inferior orbital rim. Damage to developing tooth roots,
particularly the canine tooth root, as it is located most cephalad, is possible. Injury there may further reduce growth
potential of the maxilla.
Preoperative preparation includes prophylactic use of intravenous antibiotics at the time of surgery to cover aerobic
and anaerobic ora in the nose, and prophylactic, shortterm anticonvulsants. No steroids are used.
A comprehensive anesthetic technique is advocated,
which includes orotracheal intubation, monitoring for
blood loss with central venous catheters, arterial line, Foley
catheter, and monitoring for air embolus by Doppler, endtidal CO2, and nitrogen monitors. Hypotension is induced
by increasing concentrations of the inhalation of anesthetic
agent at the time of craniotomy to minimize blood loss. Autologous or designated donor blood transfusion is preferred
if blood transfusion is necessary. Spinal drains are placed in
the lumbar cistern following the induction of anesthesia.
The drains are not opened, however, until bur holes have
been made and the craniotomy is about to be performed. A
Bair Hugger blanket (Arizant Healthcare, Eden Prairie, Minnesota) is used to keep body temperature above 36.5C. All
irrigation uids are warmed before use to reduce the likelihood of hypothermia.

Operative Procedure
The initial preparation for the treatment of patients with
hypercanthorum or telorbitism is the same. The patient is
placed supine with the head on a well-padded headrest and
the neck is slightly exed. Draping of the patient includes
full exposure of the scalp to the region of the midportion
of the vertex of the skull and to the level of the mouth caudally. Hair removal in the scalp, if performed at all, is minimal, and corresponds to the course of the coronal incision.


14535_C05.indd 27

4/23/08 3:14:28 PM

28 Pediatric Neurosurgery
Hair anterior to the skin incision line is braided with rubber
bands to remain out of the way during the operative procedure and is covered with masking tape or heavy aluminum
foil. A coronal incision is performed, extending down to the
level of the superior tragus anteriorly, to allow for periorbital dissection. The dissection of the anterior scalp ap
is performed in a supraperiosteal plane to the level of the
orbital rims, followed by elevation of the periosteal ap,
based anteriorly, on the supraorbital vessels from the same
coronal incision site to be used later as a covering ap for
the anterior cranial fossa oor defect created by the orbit
translocation procedure.
Bilateral pterion bur holes and one parasagittal bur hole
posterior to the coronal suture are placed. A bifrontal craniotomy line is drawn on the skull leaving ~1 cm height of
frontal bone superior to the apex of the superior orbital rim.
A bifrontal craniotomy is performed. The outer table of the
midline frontal bone of the glabelar region is removed with
a side cutting bur, and the frontal bone is fractured forward.
Alternatively, as frequently is the case in the region of excess
bone in the glabelar region, a bur hole may be placed in the
area of intended bone removal (Fig. 51A).

In patients with hypercanthorum, the remaining bifrontal
bone segment is bisected, leaving a supraorbital bar, ~5 to
6 mm in width, cephalad to the medial portion of the orbital
rim. If the nasal prole is acceptable, a segment of midline nasal bone may be left ~3 mm wide to simulate a new
nasal bridge. Two additional approaches exist, however, if
the nasal prole is unacceptable. The midline bone may be
removed entirely, leaving a 5 mm wide bone segment laterally on each medial orbital rim (Fig. 51B). When the orbital
rims are translocated medially, the medial border of the
orbital rim denes a new, more acceptable nasal prole.
The second and most often used alternative is to leave a 3 mm
segment of nasal bone in the midline to serve as a base for
scaffolding for on-lay bone graft augmentation of the dorsum of the nose. The bone used for augmentation is usually
from the rib because of the potential to include a pressureresistant cartilaginous cap. With all these techniques, the
medial orbital osteotomy is usually performed with a sagittal or oscillating saw to avoid unwanted fracture of the nasal
and lacrimal bones.
The frontal lobes are allowed to reposition posteriorly by
the cerebrospinal uid drainage for an osteotomy in the orbital roof extending posterior to the midpoint of the globes
anteroposterior axis (Fig. 51C). The medial limit of the
osteotomy is the lateral cribriform plate, avoiding injury,
at this time, to the olfactory nerve bers. Characteristically,
the cribriform area is excessively widened and frequently
obstructs medial translocation of the orbital rim. Therefore,
the anterior-most olfactory bers are divided, and the proximal segments of these nerve bers and surrounding dura
are oversewn to prevent cerebrospinal uid leakage postoperatively. The antereolateral portion of the ethmoid air cells
are removed by rongeur to allow subsequent unobstructed
movement of the orbits medially.

14535_C05.indd 28

If the medial canthi position relative to one another and to

the anteroposterior axis of the nasal bone is unacceptable,
effort is made to preserve their attachment to the lacrimal
bone. To avoid displacement of the canthi during dissection, a subciliary, transconjunctival Caldwell-Luc incision is
made infraorbitally to complete the caudal dissection and
osteotomy paralleling the superior orbit osteotomy (Figs.
A transversely oriented osteotomy is placed the paramedical frontal bone, leaving a 5 to 6 mm width of frontal bone
as the medial extension of the supraorbital bar.
If the midline nasal bone is to remain in situ, the osteotomy line does not extend into the plane of the medial nasal
bones. The segment complex of frontal ethmoid, and more
laterally situated nasal bones, is removed by osteotome bilaterally (Fig. 52A).
The medial rim with superior and inferior rim extensions,
in the form of the letter C, are mobilized. Ordinarily, the
greatest resistance to movement is encountered in the deep
nasomaxillary region. Prying of the rim with an osteotome
usually frees the bony and soft tissue attachments. Care
must be taken, however, to avoid injury to the nasolacrimal
duct during this maneuver. As the orbits are translocated
medially, infolding of nasal cartilage (septal and upper lateral) and mucosa usually results, requiring trimming of the
excess tissue. The openings in the mucosa are oversewn
to prevent gross air and bacterial contamination into the
epidural space postoperatively. The medial orbital walls and
the nasal bones are trimmed with the air drill to ~33 to 4
mm wide so that the overall distance between dacryon and
dacryon is ~10 mm or less (Fig. 52B). Trimming is done at
this time rather than earlier to avoid fracture of the medial
orbit during the prying maneuvers. The nasal process of the
maxilla is aggressively trimmed to avoid occlusion of the
airway as the orbital rim is moved medially. The orbital rims
are placed in position but are not yet secured.
If the medial canthi need to be repositioned, a transnasal
medial canthapexy may be performed. It is easier to perform
the initial stages of the canthapexy prior to stabilization for
the orbital bones because of the greater visibility afforded
by the mobile and widely separated bone segments. A drill
is used to open the entry point into the posterior superior
lacrimal bone (Figs. 52C, 52D, and 53A) to avoid fracture
of this fragile bone. If the medial orbital wall is exceedingly
thin and does fracture, a split calvarial bone graft (Fig. 53B)
from the parietal region may serve as a substitute for the
medial orbital wall. It will provide stabilization of the medial canthus and avoid anterior as well as lateral migration
of the canthus postoperatively. The medial orbital rims are
then translocated medially and secured to each other and
the frontal supraorbital bar.
If the midline nasal prole remains unacceptable after
translocation of the medial orbits, a cantilevered costochondral bone graft may be secured to the existing nasal prole
(Figs. 53C and 53D). Costochondral rib grafts are the primary choice graft material because the cartilaginous tip,
unlike even calvarial membranous bone, resists resorption
extremely well. To lessen the likelihood of epidural infection
postoperatively, the pericranial ap previously elevated at
the time of anterior scalp ap dissection is tacked over the

4/23/08 3:14:28 PM

Figure 51 (A) A bur hole is placed in the glabelar region at the site of
intended bone removal (green) to allow for safe dissection of the midline dura and sagittal sinus. Orbital osteotomies are located, as shown,
to include provision for removal of the medial inferior portion of the
nasal process of the maxilla so as not to impinge on the nasal airway
following medial orbital translocation. A supporting frontal bone bar,
~5 mm tall, is left above the medial superior orbital rim. (B1) The midline nasal bone has satisfactory projection, but the breadth is too great.
The midline nasal bone is in situ, and resection of excess bone occurs in
the paramedical local left (green). (B2) The midline projection is unacceptable. The midline bone is removed, and the medial orbital walls,
when translocated medially, form the new nasal prole. (B3) The existing nasal prole is decient, but, rather than excising the midline nasal
bone, it is allowed to remain in situ to serve as base scaffolding for dorsal

14535_C05.indd 29

Malposition of the Orbits


augmentation by placement of a cantilevered bone graft. (C1) View of

the anterior skull base from above. The orbital roof osteotomy extends
from the midportion of the orbit laterally to the cribriform plate medially. (C2) The osteotomy extends posteriorly well behind the midpoint
of the axis of the globe. Bone is removed medially (green) adjacent to
the cribriform plate to allow for medial translocation of the orbital roof.
(D) To place osteotomies in the inferior orbital region, either a transconjunctival Caldwell-Luc or subciliary incision is made. (E) In transconjunctival and subciliary approaches, a pre septal dissection is preferred to expose the inferior orbital rim. (F) The inferior orbital osteotomy is placed
at the level of the infraorbital foramen in children to avoid damage to
developing tooth buds. (From Persing J, Edgerton M, Jane J, Eds, Scientic
Functions and Surgical Treatment of Craniosynostosis 1989. Baltimore: Williams and Wilkins. 263269. Reprinted by permission.)

4/23/08 3:14:28 PM

30 Pediatric Neurosurgery

Figure 52 (A) Dissection is performed posterior to the lacrimal crest

medially. An oscillating saw is used to complete the superior and inferior orbital osteotomies. The posterior medial orbital wall osteotomy is
complete by an osteotome. Particular attention is necessary to remove
enough bone in the ethmoid air cell region to allow for unimpeded
translocation of the medial portion of the orbital rim. (B) Following
mobilization of the orbital rims, the midline bone and remaining medial orbital rims are trimmed (green) with a shaping bur to achieve the

14535_C05.indd 30

thinnest possible nal nasal midline (ordinarily 8 to 10 mm encompassing the medial orbital walls and the midline bony strut). (C) The medial
canthi are elevated with an accompanying periosteal pennant from the
dorsum of the nose. (D) A drill hole (green) is placed in the posterior
superior lacrimal bone for the periosteal pennant and the canthus to be
passed transnasally. (From Persing J, Edgerton M, Jane J, Eds, Scientic
Functions and Surgical Treatment of Craniosynostosis 1989. Baltimore: Williams & Williams. 263269. Reprinted by permission.)

4/23/08 3:14:32 PM

Figure 53 (A) An air-driven bur is used to penetrate the lacrimal bone

perpendicular plate of the ethmoid to allow threading of the canthal
tendon. Any signicant opening in the nasal mucosa is oversewn. (B) If
the lacrimal bone is too fragile or otherwise unusable, a split calvarial
bone graft can be used to serve as a buttress substitute for the lacrimal
bone. The canthal periosteal pennants are tightened following medial
translocation and xation of the orbital rims. (C) Augmentation of the

Malposition of the Orbits


nasal bridge is accomplished following translocation of the orbits by

placing a wedge of bone beneath the antilevered costochondal cartilage
graft. (D) The graft is secured to the underlying medial orbital rims by
transosseous wiring. (From Persing J, Edgerton M, Jane J, Eds, Scientic
Functions and Surgical Treatment of Craniosynostosis 1989. Baltimore: Williams & Wilkins. 263269. Reprinted by permission.


14535_C05.indd 31

4/23/08 3:14:37 PM

32 Pediatric Neurosurgery
ethmoid air cells on the anterior cranial base. The scalp ap
is then closed in two layers, galea and skin.
The translocation of the medial orbit rim medial in hypercanthorum effectively enlarges the orbit. This may result in
enophthalmos if the movement is more than 3 to 4 mm. In
this case, adding intraorbital bone grafts or bone substitutes
is highly desirable.
In rare circumstances, medial translocation of the medial
orbital walls at the level of dacryon is desirable. This approach avoids an open craniotomy and may be very useful
in patients with indwelling ventriculoperitoneal shunts, or
other co-morbidities. Here, the nasal osteotomy is performed
much as described for the medial translocation of the whole
medial wall, except that it is done at an infracranial level.
Bone is removed from caudal to cephalad with the superior
point at the level of dacryon or below the cranial base.

Hypertelorism (Telorbitism)
For patients with hypertelorism (telorbitism), the operative
approach is much the same as just described for hypercanthorum, but, in addition, the lateral orbit is moved with the
medial as a single unit. The movement of the lateral skull
with the medial maintains orbital volume, and it obviates
the need for intraorbital volume augmentation.
The patient is positioned supine, following placement of
a lumbar cerebrospinal uid drain, and a bifrontal craniotomy, including temporalis muscle elevation, is performed.
The supraorbital bar is left 1 cm wide at the level of the
orbital rim apex (Figs. 54A54C). The supraorbital bar
will be transversely bisected when translocation of the orbit is performed, leaving a 5 mm wide supraorbital bar for
xation, and a 5 mm wide orbital rim a the rims apex. The
orbital roof and lateral orbital wall are cut posterior to the
midpoint of the globe following protection of the anterior
tip of the temporal lobe. If correction of an accompanying
orbital malrotation is not necessary, horizontally oriented
osteotomies are placed on the anterior surface of the maxilla through or below the level of the infraorbital foramen
(Fig. 54D) in adults. In young children, the more cephalad
osteotomy is desirable to avoid injury to developing tooth
roots. Olfactory ber section and frontonasoethmoid resections are performed as described previously for the treatment of hypercanthorum.
In patients with hypertelorism, the medial canthi may
require repositioning. This is performed by transnasal canthopexy as described earlier (hypercanthorum). Likewise,
the need for repositioning of the lateral canthi ~2 mm above
the medial canthi on the horizontal axis also may be evident. After the orbits have been translocated and secured,
the attachment point for the lateral canthi, in most cases, is
placed just inside the orbital rim (Figs. 54E54H). When

14535_C05.indd 32

severe exorbitism (globe protrusion beyond the eyelids secondary to a constricted orbit volume) coexists, the canthi
are attached to orbital rim bone on the external surface of
the zygomatic process of the frontal bone. This reduces the
projection of the globe beyond the eyelid. The temporal
fossa is lled with calvarial bone chip grafts, and the temporalis muscle is advanced forward to be attached to the
orbital rim, to prevent an hourglass deformity or hollowing postoperatively in the temporal region.
The incisions are closed in two layers, galeal and skin.
The nose is packed with petroleum-based gauze. No drain is
inserted in the galeal region to avoid aspiration of nasopharyngeal bacteria into the subgaleal space.

Postoperative Management Including

Possible Complications
The patients are monitored postoperatively by clinical measures, and intracranial pressure (ICP) monitoring is not used
routinely. This is because placement of a lumbar cistern
drain allows for false low recordings of ICP despite clinical
evidence of cerebral edema. This may lead to a tardy diagnosis of elevated ICP or intracranial hemorrhage. Because
of this, pain control postoperatively should still allow for
clinical assessment of neurological status.

Complications from the operative procedure are relatively
few. The major immediate concerns relate to cerebral
edema and/or intracranial hemorrhage and injury to the
visual system, either to the globe or optic nerve by trauma
or hematoma, or to the extraocular muscle system. Also, it
is important to note that if sufcient bone is not removed
from the medial portion of the cribriform plate as the orbit is translocated medially, there is the possibility of impingement of the medial rectus muscle on the corner of the
remaining bone. This may require reoperation to correct.
Later concerns include cerebrospinal uid leakage, subdural
or epidural infection, and osteomyelitis. The possibility of
cerebrospinal uid leakage and meningitis should be signicantly reduced by watertight dural closure supported by the
use of brin glue at the suture line, with further support by
the pericranial ap overlying the dural closure. Unresolved
problems are soft tissue relaxation at the medial canthal
region resulting in an apparent redevelopment of hypercanthorum, and the possibility of growth disturbance on the
nasomaxillary and midface regions with surgery performed
in early childhood.

4/23/08 3:14:39 PM

Figure 54 (A) A bifrontal craniotomy is performed with a 1 cm tall supraorbital bar left above the apex of the orbital rim. It is bisected leaving a
5 mm thick supraorbital rim that may be translocated medially, following
removal of paramedian frontal and nasal bone (green), and a 5 mm supraorbital bar to which the orbital rim bone is afxed. (B) Note resection of
the nasal process of the maxilla to avoid impingement on the nasal airway.
(C) The orbits are then translocated medially. (D) Bone grafts are inserted
posterior to the lateral rim of the orbit and in the region of the sigma to
prevent a postoperative hourglass deformity. A drill is used to trim the
lateral portion of the supraorbital bar. (E) The normal position of the lateral
canthus is approximately at or 2 mm above the level of the medial canthus.

14535_C05.indd 33

Malposition of the Orbits


(F) Drill holes are placed in the frontal process of the zygoma. The lateral
canthus is attached transosseously to the internal surface of the zygomatic
process of the frontal bone, if the patient does not demonstrate globe proptosis. (G) If the patient has accompanying signicant proptosis, the canthus
is placed on the external surface of the sigma in an effort to reduce the
malrelationship between the globe and the eyelids and to restore normal
lid-globe anatomical relationships. (H) A slight overcorrection of the lateral
ssure position superiorly will allow normalization of its position with subsequent soft-tissue relaxation. (From Persing J, Edgerton M, Jane J, Eds, Scientic Functions and Surgical Treatment of Craniosynostosis 1989. Baltimore:
Williams & Wilkins Reprinted by permission.)

4/23/08 3:14:40 PM

Ventriculoatrial Shunting
John Drygas and Stephen J. Haines

The ventriculoatrial (VA) shunt was introduced as a method

of treatment for hydrocephalus in the 1950s and was the
rst predictably successful valve-regulated cerebrospinal
uid (CSF) shunt. Technical developments have since made
ventriculoperitoneal (VP) shunts reliable and because they
are easier and quicker to insert and revise, and have less
propensity to systemic bacteremia if infected, they are the
initial treatment of choice for hydrocephalus for most patients. VA shunts are now the second most common shunt
procedure, representing 2 to 10% of all implanted shunts in
recent series. Where both VP and VA shunting are contraindicated, shunts to the pleural space, the gall bladder, the
ureter, the bone marrow, the subarachnoid space, and other
areas, including direct heart catheterization, have been reported to be successful.

Patient Selection
VA shunts are indicated for the treatment of hydrocephalus,
either obstructive or communicating, which is not transient
in nature or amenable to endoscopic third ventriculostomy.
Other indications for shunting, which may occasionally require VA shunting, include the treatment of pseudotumor
cerebri and drainage of arachnoid cysts and subdural hygromas that are unresponsive to other therapeutic measures. There are some patients in whom a VA shunt is the
procedure of rst choice, such as those patients with brosis
or inammation in the peritoneum from remote or recent
infections or multiple previous abdominal operations. The
peritoneal cavity or pleura of some patients, particularly
small infants, may occasionally not have sufcient absorptive capacity to handle the necessary amount of CSF.

Bacteremia or infection of the CSF or proposed shunt tract are
absolute contraindications to the placement of a VA shunt.
In the presence of infection elsewhere in the body, a shunt
should be inserted only in very unusual circumstances. Congestive heart failure and pulmonary hypertension may both
interfere with shunt function and be aggravated by the additional uid load delivered to the heart; these are relative
contraindications to the procedure, especially in infants.

Abnormal venous anatomy, previous jugular or subclavian

vein thrombosis, and unusual cardiac irritability are relative

Preoperative Preparation
Preoperatively the patient and his or her parents are informed that the major risks of the procedure are those of
infection approximating 8% and shunt malfunction of up to
40% in the rst year, either of which would necessitate revision or replacement of the shunt. When a VA shunt is placed
in an infant, malfunction due to growth-related migration
of the atrial catheter into the superior vena cava (SVC) is so
predictable that elective revision at about 2 years of age has
been recommended by some. There is a slight risk of intracranial hemorrhage (which may be increased in patients
with marked hydrocephalus). Rare surgical risks of air embolism, cardiac rupture and tamponade, and thromboembolism are mentioned, as are the attendant risks of general
anesthesia. Remote risks include immune complex glomerulonephritis and pulmonary hypertension from chronic
thromboembolism or volume overload.
Where possible, an antiseptic scrub is administered preoperatively for both inpatients and outpatients. The hair
may be shaved immediately preoperatively, although recent
studies suggest that this is not necessary. An appropriate
dose of an anti-staphylococcal antibiotic is administered
at least 30 minutes prior to incision. This has been shown
to reduce infection in standard surgical procedures and in
shunt procedures with an infection rate higher than 5%.

Anesthetic Considerations
Most patients with hydrocephalus can be presumed to have
some degree of increased intracranial pressure (ICP). Because of this a gentle anesthetic induction is preferred, being
careful to avoid any manipulations that would increase ICP,
that is, Valsalva maneuvers, coughing, prolonged hypoventilation. Appropriate inhalational or intravenous anesthetics
that decrease ICP and preserve cerebral autoregulation are
used, combined with hyperventilation if deemed necessary.
Succinylcholine is avoided because of its propensity to increase ICP.


14535_C06.indd 34

4/23/08 3:14:50 PM

6 Ventriculoatrial Shunting


Special Equipment
An ultrasound imaging system is very useful for cannulating the vein for placement of the distal catheter and for
visualizing the ventricle in infants with an open fontanelle.
A C-arm uoroscopic unit is extremely helpful in verifying correct catheter placement. Operating room personnel
must remember to don lead aprons prior to scrubbing. It is
also useful to have intravenous contrast material available.
Heparinized saline is necessary to ush the atrial catheter,
as well as an antibiotic saline solution with which to irrigate
the shunt system and wounds. The appropriate shunt system is chosen preoperatively. Programmable valves may be
used. The distal atrial catheter should have a radio-opaque
marker of some type to aid in identication on images.
There is some suggestion that antibiotic-impregnated shunt
catheters may reduce the risk of shunt infection. A central
venous pressure or electrocardiogram (EKG) monitoring device must be available to monitor pressure waves or EKG
tracings as the atrial catheter is advanced if these methods
are to be used for locating the tip of the atrial catheter.

Figure 61 (A) Lateral and (B) vertex views of incision landmarks and
(C) side view of padding placed under the head and shoulders for optimal positioning. Note the intersecting lines drawn to represent two
planes. The ventricular catheter should be passed along the line that
is the intersection of these two planes.

Operative Procedure
The basic principle of positioning the patient is to provide
clear access to the head for ventricular puncture and to the
neck for cannulation of the venous system. Therefore, following induction of satisfactory general endotracheal anesthesia, with the patient in the supine position the head is
turned to the appropriate side. In the preferred setting, the
head is turned to the left to provide access to the right neck.
Because of the vascular anatomy, right-sided cannulations
are often easier than left-sided ones. The ear maybe taped
forward and thereby easily draped out of the eld. Soft padding is placed beneath the shoulders to expose the anterior
triangle of the neck (Fig. 61C). The skin and/or hair in the
operative eld and surrounding area is then prepared with
an appropriate antiseptic. Prior to draping, the landmarks
for ventricular access and access to the venous system are
drawn on the skin (Figs. 61A and 61B). A mark is placed
~2.5 to 3 cm from the midline and 11 to 12 cm posterior to
the nasion in the adult (or approximately one-seventh of the
distance from the coronal suture to the nasion in the child).
A line is then drawn from this mark toward the inner canthus of the ipsilateral eye. Another line is drawn that passes
through this point and a point midway between the tragus
and the lateral canthus of the ipsilateral eye. If these lines
are taken to represent imaginary planes in partial sagittal
and coronal directions, their intersection forms a line that
should pass through the foramen of Monro (Fig. 62). Sterile
drapes are then applied in such a manner as to allow for access to head and neck sites. In draping the head, it is helpful
to place a sterile towel with one border on the midline to
use as a landmark in placing the bur hole. If intraoperative
ultrasound is to be used in the case of an infant, the anterior
fontanelle should be draped into the operative eld.

Figure 62 A ventricular catheter traveling along the intersection of

the two drawn planes will be directed at the foramen of Monro.

14535_C06.indd 35

4/23/08 3:14:50 PM

36 Pediatric Neurosurgery
The shunt hardware, including ventricular and atrial
catheters, valve system, reservoir, and connectors, should
be chosen before the procedure, opened, and placed in an
antibiotic-saline solution on the instrument table before the
incision is made. We generally use gentamicin and avoid the
use of bacitracin because its foaming action may interfere
with the function of some valves. Unless a programmable
valve is to be used, it is prudent to have several valves with
different pressures readily available. The valve system to
be used should be tested according to the manufacturers
recommendations to ensure its appropriate function. (Most
manufacturers now advise that although patency testing
is sensible, specic pressure testing in the operating room
cannot reproduce laboratory testing requirements and
therefore cannot accurately evaluate the functioning pressure of the valve. Therefore we only evaluate patency at the
time of surgery.) It should be lled with saline solution, have
all air bubbles removed, be connected to distal and proximal tubing with permanent suture, and be clamped on the
proximal end with a mosquito clamp shod with suture boots
to keep it full of uid. To the greatest extent possible, one
should avoid touching the skin and shunt system with the
gloved hand to minimize the risk of postoperative infection.
It is also recommended that the shunt system not touch
cloth drapes, as electrostatic forces may cause cloth bers
to adhere to the shunt and be implanted with it.

The operation is performed in a logical, stepwise fashion.
We prefer to place the ventricular and vascular catheters as
the last part of the operation because of the risk of dislodging either catheter before it is secured in nal position.

Step OnePercutaneous Access of the Internal Jugular Vein

or Isolation of the Common Facial Vein
We prefer to use a percutaneous technique similar to that
used for placing central venous catheters. Usually the internal jugular vein is used for access, though the subclavian
may be used, and the procedure is identical to cannulation for standard central venous catheters. For the internal
jugular vein, a thin-walled, 18 or 20 gauge needle that is
contained in most central venous access kits is used. The
skin is punctured ~4 to 6 cm (23 nger breadths) above
the clavicle between the heads of the sternocleidomastoid
muscle or a proportionate distance in a child. The tip of
the needle should be aimed at the sternal notch (Fig. 63).
Alternatively, the carotid artery can be palpated and the
needle aimed lateral to it. The use of ultrasonic guidance
by the surgeon can facilitate the percutaneous access to
the internal jugular vein from the most routine adult to the
most difcult pediatric patient. For the subclavian vein, the
needle is inserted just below the junction of the middle and
inner thirds of the clavicle and aimed at the sternal notch.
Once the vein is punctured, the syringe is removed and a
exible J-wire (0.021 in. in diameter for older children and
adults) is inserted through the needle; its position can be
checked with uoroscopy. The needle is removed when the
wire is veried to be in the SVC. A small incision is made

14535_C06.indd 36

Figure 63 Guidelines for percutaneous placement of the atrial catheter.


at the point of entry to facilitate passage of the introducer

and to connect the shunt system with the atrial catheter
and bury it. A standard, tear-away introducer sheath over
a vessel dilator is then passed over the wire through the
subcutaneous tissue and into the vein. The wire and dilator are removed, and a nger is placed over the sheath to
prevent the introduction of air. The atrial catheter can now
be threaded through the sheath to its proper position described in Step Four below.
If percutaneous access is unsafe or not possible, the open
approach through the common facial vein may be used. To
isolate the common facial vein, a 2 cm transverse neck incision is made one nger breadth below and parallel to the
ramus of the mandible in an adult, and a proportionally
smaller distance below the jaw in children. This should be
centered just medial to the medial border of the sternocleidomastoid muscle (Fig. 64). The platysma is split in
the direction of its bers, and using a combination of blunt
and sharp dissection along the avascular plane medial to
the sternocleidomastoid muscle, the common facial vein is
identied. The vein is isolated for ~1 cm of its length, and

4/23/08 3:14:52 PM

6 Ventriculoatrial Shunting


Step ThreePlacement of the Valve System

At this point the valve system is passed subcutaneously
from the cranial incision to the neck incision, passing in
front of the parietal boss and behind the ear. Care must be
taken to avoid the very thin skin just behind the ear and
not to have the bulky portion of the valve or reservoir over
a bony prominence where it may cause pressure erosion of
the skin. Depending on the size of the patient, a small transverse incision may be necessary behind the ear so that the
distal tubing may be passed the remainder of the distance to
the neck incision. The proximal shunt system tubing is then
cut to the appropriate length. Both ends of the system are
clamped with rubber shod clamps to maintain the system
uid-lled and free of air bubbles, and to avoid inadvertently pulling the tubing out of the incision, and they are
covered with sponges moistened in an antibiotic solution.
The purpose of placing the valve-reservoir system rst is to
minimize the number of manipulations once the vein and
the ventricle have been catheterized. When the venous or
ventricular catheter is placed early in the procedure, it may
be dislodged inadvertently. Because of this possibility, we
then place the atrial catheter, and we place the ventricular
catheter last.

Step FourPlacement of the Atrial Catheter

Figure 64 Complete ventriculoatrial shunt in place.


two ligatures of a size appropriate to the size of the vessel

are placed around it. The vein is then ligated using the most
cephalad tie.

Step TwoPlacement of a Cranial Bur Hole

The ventricular catheter may be placed into the frontal
or occipital horn of the lateral ventricle depending on the
preference of the surgeon and the individual patients ventricular anatomy. We prefer the frontal approach because
our experience indicates that such shunts function better
than parieto-occipitally placed ones, although the evidence
in this regard is contradictory. The patients nondominant
hemisphere is selected because of the remote risk of parenchymal damage caused by hemorrhage at the site of insertion. A curved incision is made with its base directed
inferiorly so that no portion of the incision will be overlying
the shunt apparatus. A subgaleal pocket is created using
blunt dissection to allow placement of the valve system.
The bur hole is then made using a standard hand or power
drill at the predetermined site, bone wax is applied, and
the exposed dura is covered with a saline-soaked sponge or
cottonoid patty until later in the procedure.

14535_C06.indd 37

The goal in placement of the atrial catheter is to position the

catheter midway between the SVCright atrium junction
and the tricuspid valve. A variety of techniques are available
to conrm this placement, the least reliable of which is the
chest x-ray, upon which we do not rely. A more accurate, yet
simple, way of correctly placing the catheter utilizes continuous pressure wave recordings as the catheter is advanced.
Another simple technique optimizes the EKG recording of
the right atrium, using the shunt tubing as an EKG lead.
The atrial catheter is introduced into the jugular vein either through the percutaneous sheath or directly through
the common facial vein or jugular vein. In the latter cases
a purse string suture should be placed prior to making the
venotomy. The ligature is cinched around the catheter just
enough to prevent back bleeding yet allow the catheter to
advance easily. The catheter has been filled with an isotonic heparinized saline solution with the proximal end
connected to a pressure transducer or EKG monitor for
continuous monitoring in a closed system, and to prevent
air aspiration. When attached to a pressure transducer a
characteristic change is noted as the catheter is advanced
first from the SVC into the right atrium and then into the
right ventricle (Fig. 65). The catheter is then withdrawn
until the atrial pressure tracing is again identified. Fluoroscopy is often used during advancement of the catheter.
If necessary, the catheter may be made more visible by the
injection of an appropriate quantity of contrast material.
When the proper position has been identified, the position is maintained by clamping the catheter with a shod
mosquito clamp. If there is a ligature around the catheter,
it is tied around the catheter so that it cannot move but
still will permit free flow of CSF. The catheter should be
aspirated to remove air bubbles and then flushed with

4/23/08 3:14:53 PM

38 Pediatric Neurosurgery

Figure 65 Typical pressure wave form readings seen with advancing

catheter positions.

heparinized saline. A shod clamp is placed proximally, and

a straight connector is tied to the proximal end of the catheter, which is then covered with an antibiotic-moistened

tip and an internal stylet in place is then advanced through

the dural opening utilizing the guidance lines and planes
described above in Fig. 62 until CSF return is evident. (We
do not use anged ventricular catheters because of the tendency of the choroid plexus to grow into the interstices of
the catheter. This can lead to intraventricular hemorrhage
when the catheter is removed during a revision.) Usually
the resistance encountered when the catheter pops through
the ventricular ependymal surface is easily felt. The stylet is
then removed and the catheter advanced ~1 cm. A depth of
5 to 6 cm is usual in adults and is proportionally less in children. A manometer may be connected to the catheter, and
the ventricular pressure may be measured and recorded.
The catheter is then clamped with a shod clamp, cut, and
connected to the valve system with a permanent suture,
keeping the knot on the side to prevent skin erosion. The
connector should then be secured to the periosteum with a
single stitch to prevent migration out of the ventricle during
growth or movement.
The system is now inspected to ensure spontaneous ow
of CSF from the valves distal tubing, and the atrial catheter is again aspirated and ushed and connected to the
valve system with the previously placed straight connector
and a permanent suture. The sutures can be tied across the
connector, to further protect against disconnection. There
should be no excess slack at this connection.
The position of the atrial catheter is again checked uoroscopically. The tip of the catheter should be seen overlying
the seventh thoracic vertebra. The ventricular catheter position can also be checked uoroscopically. By use of intraoperative uoroscopy one can avoid leaving the operating
room with an inadvertently poor catheter placement. All
wounds are inspected and irrigated with antibiotic-saline
solution and then closed in the standard fashion. The galea is
rst closed with interrupted 3-0 or 4-0 absorbable sutures,
and the skin is closed according to the surgeons preference.
In children we use 4-0 or 5-0 absorbable monolament suture, which avoids the need for suture removal. In adults
ne nylon suture is often used. The posterior auricular and
cervical incisions are closed in two layers, and sterile dressings are applied.

Intraoperative Problems

Step FivePlacement of the Ventricular Catheter

Difculty in Locating the Ventricles

The goal in placement of the ventricular catheter is to place

the tip in the frontal horn just anterior to the foramen of
Monro. In an infant, ultrasound through the anterior fontanelle can often be used to direct the catheter placement in
this fashion.
The center of the exposed dura is coagulated with a needle-point monopolar cautery to create an opening with a
diameter equal to that of the catheter. This reduces the risk
of CSF egress around the catheter, reducing the incidence of
CSF leak, especially in pediatric patients with a thin cortical mantle. The catheter should be marked in some fashion
so that it is not advanced excessively (beyond 7 cm in the
adult). A ventricular catheter with multiple side holes at its

We nd the above method of passing the catheter most

successful in cannulating the anterior part of the frontal
horn of the lateral ventricle. If unsuccessful, three passes
are made at slightly different mediolateral angles, at no
time passing the catheter to a depth greater than 7 cm in
an adult. Passing the catheter perpendicular to a plane
tangent to the skull at the point of entry is often helpful. As mentioned, ultrasound can be used in a child with
an open anterior fontanelle, or in an adult by creating an
adjacent bur hole. If difculty in ventricular catheter placement is anticipated (usually because of slit ventricles) image-guided catheter placement (frameless or framed stereotaxy) should be used.

14535_C06.indd 38

4/23/08 3:14:55 PM

6 Ventriculoatrial Shunting


Difculty Passing the Atrial Catheter in a Patient with

Normal Vascular Anatomy

the ICP is particularly high to protect against collapse of the

cerebral mantle.

Once the catheter enters the vein, it usually passes straight

into the SVC and right atrium. If any difculty is encountered, a standard J-wire commonly used for central vascular
access can be passed into the vein and almost always easily
directed into the atrium under uoroscopy. The catheter can
then be threaded over the wire.


Difculty Passing the Atrial Catheter in a Patient with

Abnormal Vascular Anatomy
If a problem with venous access is anticipated, such as a
patient who has had multiple VA shunts, previous SVC or
jugular thrombosis, or multiple previous central venous access lines, it is advisable to obtain preoperative venography
to dene the anatomy. Where the internal jugular has been
previously sacriced or thrombosed, the external jugular
may be of use. It is also possible with the aid of an interventional radiologist to catheterize a femoral or hepatic vein
in a retrograde fashion using an angiographic catheter. One
may then cut down directly on the wire or snare, and this
may be used to direct the catheter into the atrium.

Difculty Positioning the Atrial Catheter

We nd the method of using continuous pressure recordings the easiest and most accurate way of placing the tip of
the catheter correctly in the atrium. The use of contrast material under intraoperative uoroscopy is another method.
Or, using the saline-lled atrial catheter as an EKG lead, the
characteristic biphasic P-wave changes seen in the atrium
may be identied. Two-dimensional intraoperative echocardiography has also been used successfully for this purpose,
but is technically cumbersome.

Postoperative Management Including

Possible Complications
Antero-posterior and lateral x-ray lms of the skull and thorax should be obtained within the rst 2 postoperative days
to verify catheter positions and continuity of the system.
Since the C-arm was used for the thoracic localization, it can
be directed cranially for ventricular catheter localization
before leaving the operating room. A computed tomography
scan should be done a few weeks after surgery in the case of
a rst-time shunt placement to document ventricular size at
a time when the shunt is known to be functional. This can be
invaluable in assessing shunt function in the future.
We monitor the EKG for the rst 24 hours postoperatively
because of the possibility of arrhythmia. One may wish to
nurse the patient in the supine position for the rst day if

14535_C06.indd 39

Malfunction of the shunt system and infection are the two

most common complications of the procedure and may occur at any time after operation, including in the recovery
room. The former may be a consequence of an indolent
CSF infection, obstruction of any part of the valve or tubing by proteinaceous debris or brosis, or disconnection.
We believe the technique of securing all connections with
a nonabsorbable suture tied across the connector can help
prevent disconnection.
Infection can be minimized by the use of perioperative antibiotics, by keeping the handling of the shunt system to a
minimum, and possibly by the use of antibiotic-impregnated
shunt materials. When at all possible, the system should be
manipulated with instruments rather than with the gloved
hand. When tapping the shunt reservoir the skin should always be adequately prepared with an appropriate antiseptic.
Patients with a VA shunt should be advised to follow standard
bacterial endocarditis antibiotic prophylaxis before any surgical or dental procedure. Signicant but rare complications
include pulmonary venous thromboembolism, pulmonary
hypertension, sepsis, immune complex glomerulonephritis,
and hemorrhage from erosion of the SVC. Elective revision
can be anticipated after 2 years in an infant or child due to
growth-related migration, which can be followed by routine
thoracic radiographs.

Ventricular shunting procedures are commonly thought of
as minor neurosurgical procedures. Given short shrift in
training and low priority in practice, the operation may give
suboptimal results. However, with meticulous technique and
skillful execution, the VA shunt can be a safe and effective
alternative in the neurosurgeons armamentarium for the
treatment of hydrocephalus in patients who would not benet from endoscopic third ventriculostomy or VP shunt.
This chapter is a revision of the chapter, Ventriculoatrial
Shunting by Paul J. Camarata, M.D. and Stephen J. Haines,
M.D. The chapter appeared in the Neurosurgical Operative
Atlas, Volume # 1, edited by Setti S. Rengachary and Robert H.
Wilkins. The Neurosurgical Operative Atlas was published by
the American Association of Neurological Surgeons (AANS)
from 1991 to 2000.
We would like to acknowledge and thank Paul J. Camarata, M.D., for his help and efforts on the original chapter
published in the rst edition of this work.

4/23/08 3:14:56 PM

Repair of Growing Skull Fracture
Tadanori Tomita

Growing fracture is a rare complication of skull fracture

occurring in infancy and early childhood. This late complication of skull fracture is also known as a leptomeningeal
cyst. Growing fracture is somewhat of a misnomer, but it
is characterized by progressive diastatic enlargement of the
fracture line. Although skull fracture is a common occurrence
in the pediatric age groups, the incidence of growing fracture
is only 0.05 to 1% among skull fractures in childhood.

Patient Selection
The usual presentation of the growing fracture is a progressive, often pulsatile, lump on the head. Neurological symptoms such as seizure, hemiparesis, and mental retardation
are less frequent. Often these patients are perfectly asymptomatic, and a palpable mass or widening of the fracture
line is the sole sign of neurological sequalae noted incidentally by the parents. Usually a growing fracture develops within a few months following the initial skull fracture, but it may not be recognized for many years. Growing
skull fractures usually occur during the rst 3 years of life
(most often during infancy), and almost never occur after
8 years of age. Although fractures may form in any part
of the skull, the most common site for growing fracture is
over the skull vault in the parietal region. Dural laceration is
always present along the fracture line, and it is an essential
factor for the development of a growing fracture. The dural
laceration enlarges with the growing fracture. Computed
tomography (CT) or magnetic resonance imaging (MRI) often demonstrates a focal dilatation of the lateral ventricle
near the growing fracture. Lack of resistance of both dura
and skull leads to focal amplication of the pulse wave of
the intracranial pressure, causing herniation of the brain or
subarachnoid space through the fracture line and the dural
defect. The growth of the fracture line is caused by bone
resorption due to continuous pulsatile pressure at the edge
of the fracture line. A rapidly developing infantile brain and
associated pathological conditions such as brain edema or
hydrocephalus also contribute an outward driving force to
cause brain herniation through the dural and skull defect.
This pulsatile force of the brain during the period of its rapid
growth produces the brain herniation through the dural laceration and fracture line, causing the enlargement of the
fracture line of the thin skull.

One of the risk factors for the development of a growing

fracture is the severity of head trauma. A linear skull fracture
with underlying hemorrhagic contusion of the brain suggests
a severe injury, signicant enough to cause a dural laceration.
Initial CT scans for the evaluation of head trauma in patients
who ultimately develop a growing fracture usually reveal
signicant hemorrhage or contusion subjacent to the skull
fracture. When a growing fracture is inspected at the time of
surgical repair, the herniated brain is seen to be developing
a cerebromeningeal cicatrix. In some cases, loculated subarachnoid cerebrospinal uid (CSF) cyst(s) may be noted with
underlying gliotic, atrophic brain. Although the loculated subarachnoid apace may become cystic (leptomeningeal cyst),
true leptomeningeal cysts are rare. The cystic changes in the
growing fracture usually represent cystic encephalomalacia.
Depressed fractures usually do not cause growing fractures, but a linear fracture extending from the depressed
fracture can lead to a growing fracture. The child who on
initial x-ray lms of the skull has diastasis of the fracture
>4 mm is considered to be at risk for future development of
a growing fracture. Diastasis of a cranial suture, however, is
an unusual site for a growing fracture.
A growing fracture at the skull base can occur in an older
age group, especially where the bone is thin such as in the
orbital roof, if a linear fracture is accompanied by a dural
laceration. Growing fracture and a meningoencephalocele
can develop with a similar mechanism as those occurring
in the skull vault of the young patient.

Radiological Studies
X-ray lms of the skull show wide diastases of the fracture
line. If initial skull lms are available, one can compare the
lms to conrm growth of the fracture line during the
interval. When multiple fractures are noted in the same patient, healing of the fracture in one area may be noted as
opposed to a growing fracture in another area. The fracture
line can cross the coronal or lambdoid sutures but is usually
limited to one parietal bone.
Neuroimagings such as CT and MRI provide information
regarding the sequelae within the growing fracture and any
intracranial pathological changes. Furthermore, if they are
available from the time of initial trauma, it should be possible to demonstrate progressive changes. It is not unusual
that the initial neuroimagings show hemorrhagic contu-


14535_C07.indd 40

4/11/08 11:24:54 AM

7 Repair of Growing Skull Fracture


sion, or subarachnoid or extraparenchymal hemorrhage. At

the time of discovery of the growing fracture, neuroimagings demonstrate the diastasis of the fracture line and often cystic lesions near the fracture site. These cystic lesions
represent encephalomalacia, a loculated arachnoidal cyst, or
cortical atrophy. The ipsilateral ventricle tends to show focal
porencephalic dilatation with ipsilateral shift of the midline
structure. This phenomenon may be due not only to lack of
dural resistance but also to cerebral atrophy.

Surgical intervention is indicated with a growing fracture
line, seizure disorder, or progressive neurologic decits. A
progressive cystic degeneration in the brain that has herniated through the dural and cranial defects can occur; therefore, surgical correction is recommended in young children
even when seizures or other neurological symptoms or signs
are absent. However, incidental, asymptomatic, and stable
fractures in late childhood or adulthood probably do not
require surgery. The goal of surgery for growing skull fractures is to repair the dural laceration and cranial defect, and
to resect seizure foci. Growth of the growing fracture may
arrest after CSF diversion shunting by a decrease of the CSF
pulse pressure, but this does not correct a seizure disorder.
Placing a shunt for primary treatment of these patients is
not advised unless hydrocephalus is present. Shunting for
nonhydrocephalic patients creates undesirable shunt dependency.

Operative Procedure
The scalp incision should be large enough to expose the entire
length of the skull defect. An S-shaped or semicircular skin incision is made, and the scalp ap is turned subgaleally, leaving
the underlying periosteal tissue intact (Fig. 71A). By palpation, the entire length of the cranial defect covered by pericranium is exposed in surgical view. The site of the cranial defect
is often bulging and may be accompanied by blush appearance
due to an underlying subarachnoid cyst. As the cranial defect
is dissected by incising the pericranium along the edge of the
bony defect (Fig. 71B), soft tissues adherent to the edge of
the cranium defect are scraped off by a sharp dissector.
The surgeon should remember that the dural edge is invariably larger than the cranial defect, and that the pericranium
is directly adherent to the underlying cerebral tissue at the
cranial defect. An effort to expose the dural edge by removing
the cranial edge should not be undertaken, as this procedure
is often complicated by removing the dura simultaneously
with the skull bone due to the adhesive nature of the dural edge. To identify the dura, several bur holes are made
away from the skull defect with a distance of at least 50% of
the width of the cranial defect. At this time, a large enough
amount of pericranium is removed from the neighboring
skull to use it for repair of the dural defect. Once the dura is
identied at each bur hole site, the dura is separated from the
inner table of the skull toward the defect (Fig. 71C). A craniotomy is made around the skull defect by connecting the bur

14535_C07.indd 41

Figure 71 (A) The scalp ap is turned subperiosteally. The cranial

defect is usually covered by the pericranium. (B) The pericranium is incised along the edge of the cranial defect. Then, the edge of the cranial
defect is exposed by scraping off the soft tissues adherent to it. (C) The
pericranium is removed from the surrounding skull surface and preserved for dural repair. Four bur holes are made in the surrounding skull
for a craniotomy. After the conrmation of intact dura matter under
the bur hole, the dura is separated from the bur hole toward the cranial
defect. The surgeon should not attempt to identify the dura by removing the bone from the edge of the cranial defect. The craniotomy is
performed on both sides of the growing fracture. The two bone aps
are removed and preserved for autologous bone cranioplasty.

holes with a craniotomy. Two pieces of the craniotomy ap

are obtained, one from each side of the growing fracture.
After the craniotomy is completed (Fig. 72A), reactive periosteal tissue and the cerebromeningeal cicatrix are identied in the dural defect. Under magnied vision by means of
surgical loupes, the cicatrix including the periosteal tissue is
lifted, and all abnormal tissue is separated and transected using a bipolar cautery until normal white matter is exposed
(Fig. 72B). The edge of the dura is separated from the cerebral tissue, carefully avoiding trauma to the cerebral blood
vessels. In this region, abnormal tissue such as cystic changes
or xanthochromic discoloration due to previous hemorrhage
is often noted.
After adequate debridement of the cicatrix at the growing
fracture and freeing of the intact dural edge from the cortical surface, the dural defect is closed using the periosteal
graft (Fig. 72C). Autologous pericranium is preferable to
cadaver dura. A watertight closure of the dura is important
to avoid a recurrence of the growing fracture or postoperative CSF leakage.
Each of the obtained craniotomy aps is split at the diploic space with an osteotome, separating it into inner and
outer tables (Fig. 72D). The cranial defect is then repaired

4/11/08 11:24:55 AM

42 Pediatric Neurosurgery
by laying in the split autologous skull grafts. Usually four
pieces are laid next to each other side by side to ll the cranial defect. These aps are secured to each other with either
nylon sutures or stainless steel wires through drill holes
(Fig. 72E). These aps are further secured to the craniotomy edge. If the defect of the skull is too large or the skull
is too thin to separate into inner and outer tables, one may
consider autologous rib grafts. These autologous bone grafts
are well incorporated, and healing is excellent. Foreign materials such as methyl methacrylate should be avoided for
cranioplasty in the growing skull.

Specic Considerations
The growing fracture may extend toward a dural venous
sinus such as the superior sagittal or lateral sinus. Although
these venous sinuses were spared from direct injury at the
initial trauma, direct exposure of them is not advised or

necessary. When the fracture line extends perpendicularly

to these sinuses, the closest end to the sinus does not need
dural repair. However, if the growing fracture is parallel and
near to the sinus, dural repair may be difcult due to the
lack of enough dural edge next to the sinus. In these cases,
one may repair the dural defect with a periosteal graft sutured to the periosteum of the skull above the sinus.

Postoperative Management Including

Possible Complications
CSF diversion shunting has been recommended for persistent postoperative CSF leakage from the craniotomy wound.
It is justied if coexisting hydrocephalus is evident, or if
CSF leakage occurs despite adequate repair of the growing
fracture. A lumboperitoneal shunt or temporary lumbar CSF
drainage is to be considered under these circumstances.

Figure 72 (A) After the craniotomy, the intact
dura mater is exposed around the dural defect,
which is covered by the periosteum. Underneath
the overgrowing periosteum is a cerebromeningeal cicatrix that is removed using bipolar cautery
and sharp dissection until healthy white matter
is exposed. (B) After all pathological tissues have
been removed, the edge of the surrounding dura is
separated from the intact cortical surface. (C) The
previously removed periosteum is used to repair
the dural defect. A watertight closure is achieved
with 40 sutures. (D) The bone grafts are split at
the diploic space between the inner and outer tables by means of an osteotome. (E) The obtained
split bone aps are used to repair the cranial defect. The bone aps are secured to each other and
to the edge of the cranial defect with nylon sutures
or stainless steel wires.

14535_C07.indd 42

4/11/08 11:24:55 AM

Occipital Encephaloceles
William O. Bell

Encephaloceles are uncommon congenital malformations

of the central nervous system (CNS) occurring in ~1 to 3 of
every 10,000 live births in Western civilization. Seventy to
eighty percent of all encephaloceles occur in the occipital
area, with the remainder located anteriorly or at the base of
the skull. In the Far East, anterior and skull-base encephaloceles are the most common, and cranial CNS malformations are more common than spinal malformations.
Occipital encephaloceles range in size from quite small
(12 cm) to larger than the neonates head. Almost all are
covered with partial-thickness skin, although the skin over
smaller defects is usually full-thickness. Identication of
these anomalies is usually straightforward as they are almost always quite obvious. In my experience the majority
of encephaloceles are small, sessile, and nearly planar with
the surrounding skin, thereby making their identication
somewhat more difcult. The presence of an encephalocele
can have prognostic implications, especially if hydrocephalus is also present.
The occurrence of hydrocephalus in the presence of encephalocele is due to aberrant development of the brain
stem and cerebrospinal uid (CSF) pathways. Hydrocephalus may be present at birth or develop following repair of
the encephalocele. Up to 50% of infants with a large occipital encephalocele will require a CSF shunt for control
of hydrocephalus. However, hydrocephalus is rare in small,
skin-covered encephaloceles.

Patient Selection
All except the very smallest encephaloceles must be repaired, but because the majority of these are covered with
skin, there is no emergent need to take the child to surgery
within 24 to 48 hours of birth. This delay allows adequate
time for preoperative planning and a full discussion of the
implications of an encephalocele with the infants parents.
Although the risk of seizures after repair is high, I do not
begin prophylactic anticonvulsants preoperatively but wait
for seizures to occur before starting these drugs.
Smaller encephaloceles are usually slightly raised with
darker, thicker hair at the central portion. They are covered
by full-thickness skin and invariably are located at the vertex. Operative closure of these very small encephaloceles is
not a necessity, but I recommend it for cosmetic purposes.

In my experience many of these lesions are hyperpathic and

should be excised for the comfort of the infant. For these
smaller lesions I recommend waiting until the infant is 3 to
6 months of age for excision, as that allows the anesthesia
and surgery to be less risky.

Preoperative Preparation
Operative closure/excision of an occipital encephalocele is
usually straightforward, as long as certain important points
are kept in mind. A computed tomography (CT) or magnetic
resonance imaging (MRI) scan should be obtained preoperatively to assess the intracranial contents for gross brain
structure and ventricular size. I prefer a CT scan using 3 mm
cuts because the information sought can be obtained easily
by this procedure and monitoring the infant during an MRI
scan may be problematic. Current-generation CT scanners
can obtain the necessary information in a matter of minutes.
Very often, there are brain abnormalities that will affect
prognosis, and these should be discussed with the parents
before the operative repair is begun.
The majority of occipital encephaloceles are located infratentorially. The exact locations of the major venous sinuses and their relationship with the encephalocele can be
determined accurately with MRI if needed.

Operative Procedure
A general anesthetic is required for this procedure. In the
vast majority of instances, a balanced anesthetic technique
is sufcient. The infant must be positioned prone (Fig. 81).
If the encephalocele is large, this positioning will result in
undue pressure being placed on the globes unless appropriate care is taken to keep the area of the orbits free from
any encumbrance. I use umbilical tape around the padded
horseshoe headrest at the level of the orbits for this purpose
(Fig. 81). For lesions at the vertex, the neck may be placed
in a neutral position, but for lesions in the suboccipital area,
the neck must be exed as much as possible for the surgeon
to work effectively. For very large defects, I recommend that
the encephalocele be suspended from an overhead device
to allow proper skin preparation and to keep the weight off
the head until the defect can be excised. The anesthesiolo-


14535_C08.indd 43

4/23/08 3:15:19 PM

44 Pediatric Neurosurgery

Figure 81 The patients position for repair of an occipital encephalocele. Note the umbilical tape around the horseshoe headrest at the level of
the eyes and the rolls placed beneath the child.

gist should be aware that the neck will be exed and that
this maneuver may change the position of the endotracheal
tube once the child is turned prone. Extraordinary care must
be taken to ensure that the endotracheal tube is securely
taped and secured to avoid its dislodgement during the procedure.
The childs torso is placed on soft rolls oriented either vertically or horizontally, and care is taken to avoid any pressure
points. I usually do not place an arterial line or a Foley catheter because these operations are neither bloody nor lengthy.
I use 3M Steri-Drape1010 surgical drapes (3M, St. Paul,
Minnesota) because they nicely establish the perimeter of
the area to be draped and reduce the amount of exposed
skin, thereby allowing the infant to retain body heat during
the procedure. In addition, these drapes prevent the skin
preparation solution, bodily uids, and irrigation uid from
dislodging the tape holding the endotracheal tube in place.
For skin-covered defects, I prepare the skin using diluted
providone-iodine scrub, tincture of iodine, and alcohol in
that order. For defects with exposed tissue, I use Betadine
scrub and Betadine solution, followed by a normal saline
rinse. This avoids applying alcohol to the exposed tissue.
For small, skin-covered lesions, I recommend a horizontal
ellipse-shaped opening encompassing the lesion. Dissection is done along the skull from all directions toward the
skull defect. The skull defect most commonly measures in
millimeters and has good-sized arteries and veins coursing
through it. The periosteum surrounding the skull defect is
usually thickened, and I excise this also. The dural defect is
either very small or nonexistent and can be closed with a re-

14535_C08.indd 44

absorbable suture. If there is a substantial skull defect, then

a piece(s) of bone can be harvested from the adjoining skull
and placed into the defect. The skin is closed in a standard fashion with galeal and skin sutures. I recommend reabsorbable sutures for both layers. Undermining the scalp in
the subgaleal space around the defect will ensure that there
is no tension on the skin closure.
For larger lesions, I open the encephalocele sac either vertically or horizontally with a scalpel and then use Metzenbaum scissors (Figs. 82A and 82B) to obtain a direct
view of the interior of the sac. The walls of the sac may
be resected or they may be everted with stay sutures (Fig.
82C), whichever provides the best exposure. Frequently,
there are multiple concentric layers of arachnoid that need
to be opened with forceps to expose the neural tissue that is
located at the base of the encephalocele sac (Fig. 82C).
What to do with neural tissue located outside the cranial cavity is somewhat controversial. Some have advocated
pushing it back inside the cranium; others have suggested
rst performing electroencephalographic or evoked potential recordings to determine whether it is functional neural
tissue. Whether the externalized tissue is functional or not,
forcing it inside a cranium that has not been housing it may
exacerbate hydrocephalus or disrupt intracranial dynamics.
Histological sections of this external neural tissue invariably
show disorganized neural tissue without layered cerebral or
cerebellar cortex. The neurons are usually interspersed in a
glial background in what appears to be a random manner.
My usual practice is to excise the exposed neural tissue
(Figs. 83A and 83B). Frequently, there are reasonably

4/23/08 3:15:19 PM

Occipital Encephaloceles


Figure 82 (A) The initial incision may be made vertically or horizontally and (B) then opened further with the scissors. (C) Generally, there are
arachnoidal layers that must be opened with forceps so that the abnormal neural tissue at the base can be identied.

large vascular channels (both arterial and venous) coursing

through the tissue, and these must be electrocoagulated
carefully with the bipolar forceps (Fig. 83A) before the
scissors are used to excise the tissue (Fig. 83B). If the location of the major venous sinuses is known beforehand and
reasonable care is taken during the excision, the torcular
herophili and transverse sinuses are rarely encountered.

14535_C08.indd 45

There is always a dural defect, and it is always somewhat

smaller than the associated skull defect. To obtain a watertight dural closure, I use the surrounding periosteum (Figs.
84A84C). A no. 15 blade scalpel is used to incise the
periosteum, which is then reected with an elevator such
as the Dingman periosteal elevator (Figs. 84A and 84B).
The periosteum/dura is then closed with an interrupted or

4/23/08 3:15:20 PM

46 Pediatric Neurosurgery

Figure 83 (A) After careful electrocoagulation with the bipolar forceps, (B) the abnormal tissue is then excised with the scissors.

running absorbable suture such as 40 Vicryl in a vestover-pants fashion (Figs. 84C and 85A). Because dura
and periosteum have been used, this type of closure may
allow some ossication of the skull defect, but complete
ossication has been rare in my experience. If the skull defect is large, a piece of adjoining skull can be used to cover
the defect, sutured in place with absorbable 30 or 40 sutures. The area from which the graft is taken will reossify

14535_C08.indd 46

The skin may be closed in a vertical, horizontal, or oblique direction. The rst step toward closure of the skin is
to trim away excess partial-thickness skin (Fig. 85B) and
then to begin blunt dissection in the subgaleal space (Fig.
86A). The most distance for skin closure is obtained in the
cephalocaudal direction, and it is for this reason that I usually choose a horizontal skin closure (Fig. 86B). The galea
is closed with interrupted, buried 40 Vicryl sutures and the
skin with a running 40 monolament or re-absorbable su-

4/23/08 3:15:22 PM

Occipital Encephaloceles


Figure 84 (A) The periosteum is incised with a scalpel and (B) then reected using a periosteal elevator. (C) The dura is closed in a vest-overpants fashion using absorbable suture.

14535_C08.indd 47

4/23/08 3:15:24 PM

48 Pediatric Neurosurgery

Figure 85 (A) Once the dural closure has been completed, the bone edges of the defect are identied. (B) The excess partial-thickness skin may
then be trimmed.

ture. I prefer a re-absorbable stitch for skin closure because

this eliminates the need for suture removal in a squirming
infant postoperatively.

sion until adequate healing has occurred, and that one be

vigilant for developing hydrocephalus. Before the infant is
discharged from the hospital, if overt hydrocephalus has not
developed, a follow-up CT scan or ultrasound study should
be obtained as a baseline.

Postoperative Management Including

Possible Complications
Postoperative care is routine, with the only admonition being that the infant be kept from lying directly on the inci-

14535_C08.indd 48

4/23/08 3:15:25 PM

Occipital Encephaloceles


Figure 86 (A) Undermining is done in the subgaleal space to allow for skin mobilization sufcient for closure. (B) After the galea is closed, the
skin is closed using a running monolament suture.

14535_C08.indd 49

4/23/08 3:15:27 PM

Surgical Management of Pansynostosis
James Tait Goodrich and David L. Staffenberg

Patient Selection
Children born with severe pansynostosis are usually diagnosed at birth or very shortly thereafter. The craniofacial
team is ideally notied of the childs condition very early
on, and is involved in the surgical planning as soon as possible. The most important factor in timing the surgery is
the presence of increased intracranial pressure (ICP). In our
experience, almost all children born with pansynostosis
have increased ICP (>90%). The most difcult child to sort
out is one with pansynostosis secondary to microcephaly.
Fortunately a good pediatric workup and genetic screen can
isolate this child. The typical pansynostotic child has all the
signs of increased ICP including severe beaten metal appearance or thumbprinting of the skull. There are usually
associated clinical signs of lethargy, increased irritability,
early signs of developmental delay, and so on. In some cases
these signs or symptoms present so early that the surgery
has to be scheduled within the rst month of life. This leads
us into the surgical timing problems that involve multiple

Preoperative Preparation
Timing of Surgery and Other Preoperative
Children with pansynostosis are typically diagnosed within
the rst 2 or 3 weeks of life, if not at birth. In craniofacial
surgery, the ideal time in elective cases is 4 to 6 months of
age. This growth period allows for an ideal surgical weight
plus a fully developed hematological system. In children
with severe pansynostosis and associated constricted
calvarial growth, the luxury of waiting 6 months is very
rarely, if ever, possible. Typically the surgery will have to
be scheduled within the rst 6 weeks to prevent damage
to the developing brain. The longer the surgical team can
wait the safer the surgery will be. The decision to operate is
a joint one among the pediatricians (including neonatalogists and intensive care personnel) plus the anesthesiologists, family, and craniofacial team. The potential risks to
the very young child are the signicant blood loss (almost
all of these children will require blood transfusions), and


the surgical team is operating on a child with fetal hemoglobin, and in many cases other associated congenital
anomalies. Ideally, in children with pansynostosis, the preoperative evaluation includes a thorough evaluation for
any cardiac, renal, or other associated systemic anomalies,
which are not at all uncommon. Absolutely essential is a
full radiographic workup including standard computed
tomography (CT) and magnetic resonance imaging (MRI)
scans. Congenital anomalies such as holoprosencephaly,
underlying maldevelopmental problems of the brain,
should all be worked out prior to surgery. In many cases
we ask our bio-ethics team to be involved to make sure the
child is a reasonable candidate for surgery and to make the
family aware of future developmental problems. Included
in the radiological studies should be an evaluation of the
ventricular system, as a high percentage of these children
do develop hydrocephalus because of associated severe
cranial-based anomalies. Because of unusual positioning
problems in these children (i.e., it is not uncommon to have
the child in hyperextension) the skull base and cervical
medullary junction need full evaluation. Chiari malformations, narrow or distorted foramen magnum, etc., are all
seen in this population. To prevent a child from awakening
from surgery as a quadriplegic, it is important to sort out
these details rst. If, for example, a Chiari malformation is
detected, then the surgeon will appropriately modify the
positioning to avoid hyperextension or undue stress on the
cervical region.
The use of prophylactic anticonvulsant drugs (ACDs) preoperatively remains controversial; on our service we do
not routinely place a child on anticonvulsants unless there
has been a seizure history. If ACDs are used, then the appropriate loading dose should be given at least 24 hours
in advance of the surgery and continued for 30 days after
surgery. Because these children typically have severe skullbase anomalies, we complete a full endocrine workup to
rule out any hypothalamic-pituitary dysfunction. A routine
coagulation prole is done prior to surgery. As mentioned
earlier, a full and complete medical workup is done to rule
out any cardiac, liver, or renal dysfunction. If any are detected, then the surgical management plan includes the appropriate treatment of these medical problems. It cannot be
overemphasized how thorough this preoperative workup
should beit is not at all unusual to nd one or more other
system abnormalities in the preoperative workup.

Surgical Management of Pansynostosis (Craniosynostosis)

Anesthetic and Monitoring

Surgical correction of a pansynostosis is a long and technically complex procedure. Our anesthesia team manages
these children as if they all have increased ICP so that inhalation agents that increase ICP are avoided. Use of paralytic
agents is encouraged to prevent any movement during the
case and also to assist in reducing ICP. We no longer use
dehydration agents (mannitol 0.5. g per kg) prior to induction; the exception is if the brain looks tight at the opening.
A Foley catheter is placed to monitor urine output. A temperature probe is placed either in the esophagus or rectally
to monitor body temperature. Heating blankets (e.g., Bair
Hugger blanket; Arizant Healthcare, Eden Prairie, MN) are
placed both below and on top of the patient to maintain an
euthermic body temperature.
There is almost always a signicant blood loss associated
with these cases; therefore, at least two large-bore intravenous lines need to be placed. If vascular access is difcult,
then a subclavian or jugular line can be placed. Arterial monitoring is essential to allow for evaluation of blood pressure,
serial electrolytes, and blood gases. The child is typically
kept in hypotensive range with a mean arterial pressure of
40 to 50 mmHg to help reduce blood loss. We now routinely
use oxygen saturation monitors, which can be placed on the
childs nger or toe. We position the child in mild hyperextension. We also try to place the head level with the heart to
help reduce air embolism; fortunately air embolism is rare
in these types of cases. With meticulous skin incisions and
careful coagulation of all bleeding sites, plus careful waxing
of the small emissary veins that travel through the skull
bone, the risk of air embolism can be markedly reduced.
It has been our experience that these children, because
of increased ICP, often have dilated venous patterns over
the scalp and, in addition, large diploic veins. These can be
treacherous sources of blood loss plus sources of venous air
embolism and must be carefully waxed or coagulated.
Perioperative antibiotics are used; typically only the skin
organisms are covered, for example, Staphylococcus and
Streptococcus. The drug of choice is oxacillin (or vancomycin if the child is allergic). If the child has another source of
infection such as a otitis media or upper respiratory infection, these are treated preoperatively and cleared prior to
the start of any surgery.

Operative Procedure
The typical child with pansynostosis requires almost the entire calvaria remodeled and repositioned (Fig. 9196). As
a result the entire calvaria must be exposed in the operative
eld. This surgical eld includes from the level of the orbits
to the level of the inion and sometimes the foramen magnum. To do this we place the child in the prone position with
the head extended (after radiological evaluation has shown
no abnormalities of the cervical medullary junction). The
chin rests on a horseshoe that is well padded (Figs. 91A,


91B, and 94A). The draping is done with a split sheet,

keeping the orbits and entire calvaria in the operative eld.
Bilateral tarsorrhaphies are done at the beginning the case
to protect the eyes. We no longer do a full head shave, rather
we either part the hair or trim the hair for ~2 cm where the
skin incision is to be made.
Fig. 97 shows the operating room layout. The operating
table is usually reversed head to foot so that the surgeons
have more room to place their legs comfortably under the
table. The operator works at the end of the table with the
assistant to the side. The nurse with the instrumentation is
placed to the right of the assistant. A Mayo stand is placed
over the table and on this the nurse places the instruments
directly in use. Behind the nurse is a larger table that holds
the main instrument groups. A Mayo stand is placed to the
left of the operating surgeon and directly to side of the
childs head. When the drapes are in place, this then allows
the anesthesiologist an access tunnel to the childs airway
and to the childs peripheral lines. All the suction tubes,
cautery lines, etc., are run off the opposite end of the table.
This allows the surgeon and assistant to sit and not to have
their chairs run over the various tubes and cables that can
be present. The only item at the foot of the surgeon should
be the cautery pedals and the pedals to run the air drills
when necessary.
An additional point to consider in the draping is the use
of waterproof drapes. As these cases require considerable
bone removal, the saline irrigation used can be copious. It is
important that the drapes in contact with the body not get
wet during the case. In addition we place heating blankets
both below and on top of the child. Because these are small
children the heat loss can be considerable and rapid, so hypothermia has to be carefully monitored and avoided.

Skin Incision
These cases require a large and wide exposure of the calvaria.
The simplest and most cosmetically efcient incision is the
bicoronal incision done in a zig-zag fashion. This incision is
started behind one ear and carried over to the opposite ear
and normally follows a midpoint between the nasion and
inion. This incision will easily allow the operator to expose
from the level of the orbit and nasion over the convexity and
to below the nuchal line. By this approach the entire calvaria
can be visualized. During the skin incision the surgeon must
be careful not to cut into the temporalis muscle fascia. The
pericranium is elevated as a separate layer, and where it
merges with the temporalis fascia it is incised. The temporalis muscle is then elevated as a separate layer and is hinged
to its base. By doing this the temporalis is available to be
laid down at the end of the case intact and sewn back to the
pericranial layer. If the exposure is carried down to the foramen magnum, then the nuchal muscles can be elevated as
a ap in continuity with the periosteal layer. This technique
allows the nuchal muscles to be laid down at the end of
the case when the periosteum is laid back. When elevating
the frontal periosteal membrane the supraorbital nerve and
artery complex can be easily elevated with the periosteum.
This will allow preservation of this important nerve and

52 Pediatric Neurosurgery

Figure 91 (A) Operative positioning, lateral view. The child is placed
prone in a mild hyperextended position. The child is face down resting
on a well-padded horseshoe headrest. The electrocautery grounding
pad is placed on the childs back. (B) A frontal intraoperative view of
a child with severe pansynostosis and a severe cloverleaf deformity
of the skull. This child was diagnosed at birth and by 3 weeks of age
had stopped all head growth. It is important to appreciate the number
of deformities in the skull. Because of fusion of the coronal and squamosal sutures there is a severe dimpling in the temporal region. Over
the calvaria several spots could be palpated secondary to brain escaping through the skull. The lambdoid sutures were so tightly fused the
child did not develop the usual inion point. The head assumed the typi10.1055/978-1-60406-039-3c009_f001

cal early cloverleaf appearance that occurs when all the sutures fuse. (C)
An intraoperative view of the forehead corresponds with the view in Fig.
3A after the skin aps and pericranium have been elevated. Methylene
blue has been used to mark out the osteotomy sites. The orbital bandeau will be harvested from just behind the coronal suture. The new
forehead will come from the childs left temporoparietal region (see Fig.
93A). (D) An axial computed tomography scan of this child showing
the severe thumbprinting and bone invagination into the brainall
concepts that must be kept in mind when elevating the bone aps. Getting these invaginated bones out of the brain can be extremely tedious
and difcult.

Surgical Management of Pansynostosis (Craniosynostosis)

artery that run in the supraorbital notch. Occasionally the

notch will need to be opened, especially in older children.
The notch can be opened with either a small osteotome or
a 2 mm Kerrison punch.

Operative Technique
To illustrate the operative technique used in this type of
craniofacial reconstruction we have selected two cases, each
done with a different surgical technique. We have selected
these two cases to illustrate various surgical procedures
available to the craniofacial team.

Case One
History: A 6-week-old child who presented with severe pansynostosis involving all the sutures resulting in increased
ICP and severe thumbprinting appearance of the skull.
Craniofacial reconstruction is going to require complete
disassembly of the calvaria from orbit to posterior fossa (illustrated in Figs. 9193).
The patient is positioned in the supine position with head
extended as shown in Fig. 91A and Fig. 91B. After the bicoronal incision is made and the periosteum is elevated, the
entire calvaria is exposed in the eld. As this childs forehead
and orbital rims are going to be advanced, a new bandeau and
forehead unit needs to be marked out. Fig. 91C shows the
intraoperative view after methylene blue is used to mark the
plates to be harvested. The bandeau will come from an area
just anterior to the coronal suture. The only area where a normal forehead unit could be located was over the left temporal
parietal region. The rest of the osteotomies are designed to
provide large pieces of bone for the reconstruction after the
bandeau and forehead have been advanced.
As this child has severe skull molding and thumbprinting of the inner table, the bone elevation has to be done
with great care to avoid tearing the dura or entering a venous sinus (see Fig. 92). Using a high-speed footed drill
and walking the footplate along the grooves and digital
markings, these units can be elevated safely. Areas that are
particularly treacherous are over the sinuses and between
the orbits. On the CT the surgical team noted this child has
a tongue of brain coming in deep between the orbits. Appreciating this anatomy early prevents an unacceptable situation of tissue damage to the frontal lobes. In many cases
the sphenoid wings can be severely sclerosed (giving the
harlequin eyes), and in these cases the excessive sphenoid
ridges are removed bilaterally.
Once the bone plates have been elevated, they are taken
off the eld to another sterile table where the surgeons can
reconstruct the calvaria in the fashion best suited to give an
advancement, keeping in mind the amount of bone available. In Figs. 92B and 92C we have reconstructed the new
calvaria using the forehead unit and new orbital bandeau. To
give an adequate advancement one can appreciate the lack
of bone that occurs when the reconstruction is done. Important points to remember are a good forehead and bandeau
unit rst. Bone has to be placed over the calvaria along the
sagittal sinus to prevent a later turricephaly that can occur.
We have found that if the convexity/vertex is left open, the


brain will go the path of least resistance and as a result turricephaly can occur. With proper attention placed to strut
placement the brain (i.e., frontal lobes) will expand forward
and not upward. These struts are placed along the convexity
to help orient the direction of growth. The other pieces of
bone are then placed in a mosaic fashion to cover as much of
the brain as possible. In this particular child there was such
a severe constriction of the head that after the calvaria was
remodeled with the available bone, we realized only 60 to
70% of the brain was covered with bone. In young children,
particularly under 1 year of age, reossication will occur
easily and quickly. In some cases the bone harvested can be
additionally split with ne sharp osteotomes.
In children with severe pouching of the dura caused by
thumbprinting there can occur dural tears. These tears must
be attended to and sutured so that no cerebrospinal uid
(CSF) leaks occur. CSF leaks can be very troublesome days
later with subgaleal collections so it is important to repair
any CSF leaks or dural tears done in the opening.
Fig. 92C shows the calvarial unit after placement. Absorbable sutures (e.g., 30 Vicryl) are used for stabilization
and holding in position the forehead units. In cases where
extra stabilization is needed, absorbable miniplates can be
used. We no longer use any wire or metal miniplates because of unacceptable migration of these materials as the
child grows. In several cases, at a later reoperation, we found
these metallic materials had literally migrated through
dura, and in a couple of cases they had actually come to rest
within the brain.

Case Two
History: An 8-month-old child was noted to have severe
pansynostosis with lack of head growth from 4 months of
age. After careful workup for microcephaly, which was ruled
out, the craniofacial team determined the child had severe
synostosis of all the calvarial sutures. However, in this child
the orbital units and forehead were not severely affected,
so the reconstruction was based on the units behind the
coronal sutures.
The positioning techniques and preliminary steps were
the same as in case one. The signicance of this childs synostosis can be appreciated in Fig. 94, which shows the
anteroposterior and superior views of the child. In this child
all the calvarial sutures were either closed or severely sclerosed on x-ray. However, the aesthetic examination showed
the orbits and forehead to be symmetrical. On the basis of
this, rather than do a complete calvariectomy, we kept the
forehead and orbital unit intact and reconstructed the calvaria and occipital region. In Figs. 94C and 94D the childs
head is shown from a lateral and superior view so as to see
the methylene blue markings and where the osteotomies
will be done (Fig. 96 shows a schematic reconstruction). To
correct the deformity over the occipital region and to allow
growth and advancement, a new bandeau is harvested from
over the occipital calvaria. This will act as the base unit for
reconstruction. To anchor this bandeau down, a step-off
has been marked out over the asterion region (Fig. 94D).
The entire occipital region and posterior fossa will be removed in two units, and these pieces will be used in the

54 Pediatric Neurosurgery

Figure 92 (A) The unit of bone that came from the area labeled B in
Figure 93B. The view is of the underside of the bone, that is, the inner
table. One can appreciate the severe thumbprinting that can occur in
these cases. The dura is pushing through these areas, and in addition
there are points where the bone has been completely eroded through
the dura and brain has been exposed. The elevation of this bone has
to be done very gently and carefully to avoid injury to the sinuses and
to the brain. See Fig. 91D for the CT view. (B) Once the various bone
units are harvested, they are taken to a separate table and positioned

to see how they will t on the child. It is easy to appreciate from the
photograph that severe pocketing and erosion of the bone has occurred.
With the advancement necessary to decompress the brain, there is insufcient bone available to cover the calvaria. Two important pieces of
bone are those struts that will be placed over the convexity to prevent a
turricephaly from occurring. (C) An intraoperative forehead view of the
child prior to closure. This view is an equivalent view of Fig. 93B. (D)
The child at 10 months of age.

Surgical Management of Pansynostosis (Craniosynostosis)


Figure 93 (A) An artistic reconstruction of the Figure 91C showing the various sites for osteotomies and harvesting. The bandeau is
elevated from just behind the coronal suture. The new forehead unit
(labeled A) is marked out with a Marchac template and elevated from
the temporal-parietal region. (B) An artistic reconstruction of Figure
92 only here showing the various osteotomies on the left and their

repositioning on the right. The areas marked out as A and B will form
the new forehead and posterior occipital unit. The bandeau is harvested
and advanced forward and positioned in the right gure. A sagittal strip
of bone is placed in the hopes of preventing a turricephaly from developing. The blank areas are exposed dura mater, which is common due to
lack of bone that develops in the various advancements.

56 Pediatric Neurosurgery

Figure 94 (A) An intraoperative view of an 8-month-old child with
severe synostosis of all the calvarial sutures. The dimpling about the
temporal region and the attening and distortion over the parietooccipital region can be appreciated. In this child the frontal unit and orbital bandeau are symmetrical and not distorted. The surgical team decided on a reconstruction behind the coronal suture rather than a total
calvarial remodeling. (B) An intraoperative view similar to (A) with more
of an oblique angle of the child's head showing the severe distortion of
the calvaria that has occurred over the posterior parietal and occipital
regions. (C) A vertex view similar to (A) after reection of the skin and

periosteal aps, with the various osteotomies marked out in methylene

blue. The orbital bandeau will be taken from just behind the coronal
suture, which can be seen in this view. The area over the child's left occipitoparietal area was determined to have the best conguration to reconstruct a new occipital plate. (D) A lateral view after reection of the
skin and periosteal aps. Using methylene blue the osteostomies have
been marked out. The bandeau will come from just behind the coronal
suture. This unit will be moved back and placed tongue-in-groove into
the asterion step-off unit, which has also been marked out.

Surgical Management of Pansynostosis (Craniosynostosis)

reconstruction. The piece over the right occipital region (labeled A in Fig. 96A) was determined to have the closest
symmetry to a new occipital plate, and this was marked out
using a Marchac template (Walter Lorenz, Surgical, Jacksonville, Florida; see Figs. 95A and 96B). As the sagittal sinus
lies under the sagittal suture, it is sometime useful to make
an additional osteotomy lateral to each side of the suture.
Then the plate of bone over the sagittal sinus is elevated last,
allowing better control in case of any sagittal sinus bleeding.
In some children the bone can be tightly attached at the suture lines so additional care must be taken in elevating the
bone at these points. Once the bone has been harvested the
bandeau is placed into position, tongue-in-grooved to the
asterion unit. On this bandeau the new occipital unit, which
had been previously fashioned, is placed. This acts as the
framework upon which to build the rest of the calvaria in a
mosaic fashion. In Fig. 95A we show the various bone units
gull-winged out to show the location and position prior to
suturing into position. In the older child, with rmer bone, it
is occasionally useful to place radial cuts in the bone to allow
the expanding brain to spread out (this can be seen in Fig.
95D). Fig. 95C is a superior view showing the bandeau in
position; the new occipital unit has been placed and around
this framework the calvaria is reconstructed in a mosaic
fashion using the remaining bone (Figs. 95 and 96). As in
case one, it is important to remember to cover the sagittal
sinus rst with bone, to act as a strut, and subsequently to
prevent a turricephalic growth pattern from developing. Absorbable miniplates are again used in those parts that need
the additional support. Absorbable suture material is used
primarily to stabilize the various bone units. Where extra
stabilization is needed we use absorbable miniplates.

Closure Technique
After the calvarial bone units have been placed, the closure is done in a reverse fashion of the opening. The surface is then copiously irrigated with warm physiological
saline solution. As these are long cases, with long exposure
times, copious irrigation is essential to remove any debris
and bacterial contaminants. The pericranium is elevated and
carefully stretched out over the calvaria. The nuchal muscles
should come up easily with the posterior pericranial layer.
The frontal layer is then brought up and retention sutures
are then placed between the front and back layers. These
layers will not meet due to shrinkage and the advancement;
however, they can be kept in position by retention sutures.
Next the temporalis muscles have to be reattached to the
pericranium, or sometimes they can be sewn to the sutures
in the bone units. In any case the reattachment of the muscles is key for both aesthetics and muscle function.
In the past we would place a drain in the subgaleal space,
which we now no longer do. The reason we stopped this
practice is that the drains seemed to increase the blood loss
and increase the rate of infection. Rather than drains we
pay very close attention to hemostasis prior to closure of
the skin.
Remembering that the reconstruction is a very fragile
unit, a soft uffy dressing is placed over the wound. We
do not use any tight wraps or constricting type of dressing


as this can lead to collapse of the reconstruction. A snug

Kling type of dressing is normally adequate for holding the
dressings in place.

Specialized Instrumentation
These operations are done with a standard craniotomy set. It
is recommended that a high-speed drill system with craniotome (e.g., Midas Rex; Medtronic Inc., Fort Worth, Texas)
be used for harvesting the calvarial bone. These children
typically have thinned skulls with severe inner table erosion, and so to prevent injury to the brain and dura the
operator will need a well-controlled high-speed drill system
with a ne foot plate to separate bone and dura.
To mark out the forehead unit we have been using the
Marchac forehead templates. These templates are quite
useful in locating a unit of bone on the calvaria that most
closely approximates the normal forehead.

Postoperative Management Including

Possible Complications
There are several known complications associated with
these complex craniofacial procedures. It has been our experience that if the surgical team members think of them as
they go through the preoperative, intraoperative, and postoperative preparations, in most cases they can be avoided.

Blood loss occurs as a normal part of any operation. Because
of the extensive nature of the craniectomies, amount of ap
exposure, and osteotomies the potential loss of blood can be
signicant. To help reduce blood loss several procedures can
be done. We routinely ask the family to arrange for pedigree
blood to be given. Donor-directed reduces the risk of various infectious agents that can be carried in blood. The anesthesiology team keeps the child in the hypotensive (mean
arterial pressure of 50 mmHg) range during the procedure.
Blood transfusion is done only when the hematocrit drops
below 25 or the child becomes clinically symptomatic. We
routinely plan on transfusing at least 1 unit of blood and
in ~20% of the cases a second unit will be needed. It goes
without saying that attention to meticulous technique can
help reduce the loss of blood. Use of epinephrine inltration
during the skin incision, skin clips, careful use of cautery,
etc., all can help reduce blood loss.

For any craniofacial procedure the child is preoperatively
started on antibiotics for skin organisms (usually oxacillin)
and treatment is continued for 24 hours postoperative. The
rate of infection is fortunately quite low in these cases. Due
to the length of these procedures and the debris that accumulates, it cannot be overemphasized how important it is to
copiously irrigate at the completion of each case to remove
the collected debris. These patients typically develop fevers

58 Pediatric Neurosurgery


Figure 95 (A) An intraoperative view (equivalent to Figs. 94A and

94B) after the various units have been repositioned. The new bandeau has been attached to the asterion, and the various bone units
have been mounted on the bandeau unit. In this view the surgeon has
laid the bone out to show where each of the units is being placed. (B)
An intraoperative view equivalent to (A) with the bone laid into position and just prior to stablization. (C) An intraoperative view equivalent to (A) and (B) with the bone units sutured into position. (D) A
posterior view (i.e., looking at the occipital region) showing the radial
cuts made in the occipital unit to help mold and better position the
bone. In the older child the bone is not as malleable, and this is a
useful technique to help the brain remodel the bone as it grows. (E)
An intraoperative view at the completion of the operation, equivalent
to Figs. 94A94C. The immediate correction that occurs with this
type of technique can be appreciated here.

Surgical Management of Pansynostosis (Craniosynostosis)


Figure 96 (Continued on 60)

60 Pediatric Neurosurgery


Figure 96 (A) An artist reconstruction of Fig. 94A showing the

various osteotomies. (B) An artist's reconstruction of Fig. 94D, only
from the childs right side showing the various units in their new
position. The asterion step-off can be seen here, and this is where
the bandeau will be anchored. (C) An artists reconstruction of Fig.
94D (opposite side view) showing the harvesting and direction of
the new bandeau unit. This unit acts as the framework upon which
the rest of the calvarial reconstruction will take place. (D) An artists
reconstruction of Fig. 95C, from the childs right side, showing the
various units in their new positions. (E) An artists reconstruction of
Figure 95C showing the bone units wired into position. The units
have been labeled and are equivalent to the schematic in (A). The
donor sites and their eventual location can be appreciated better here
in this schematic.

Figure 97 Operating room schematic. The surgeon typically sits at the head of the child with the assistant to the
right. The nurse is positioned to the right of the assistant. A Mayo stand is brought over the child at the buttock level. Another Mayo stand is placed to the left of the
child and parallel with the child. This provides a tunnel for
the anesthesiologist to view the child and the respiratory
apparatus. The anesthesia machinery is placed behind the
anesthesiologist. A large table is placed behind the nurse
where most of the instruments are placed. The Mayo stand
is used only for the immediate instrumentation.

Surgical Management of Pansynostosis (Craniosynostosis)

in the rst week postoperative. Routine fever workups are

done, but rarely is a source found. We attribute these fevers
to the resorption of blood that occurs from under the skin
ap. In the case of an extensive advancement it is not uncommon to leave a large dead space. These spaces have
the highest incidence of infection and if a child develops
evidence of sepsis, these spaces many need to be tapped
for a bacteriology culture. The incidence of this happening
is fortunately low but always must be considered in the

Patient Positioning
Because of the need to place the child in hyperextension,
it is important to get preoperative CTs or MRIs to rule out
any abnormalities of the cervical medullary junction, a not
uncommon nding in this patient population. Before a child
is hyperextended the surgeon must rule out any evidence of
a Chiari malformation, bony anomalies, or narrowing of the
foramen magnum. If any of these are noted, the positioning
has to be corrected to allow for it.

Cerebrospinal Leaks
All CSF leaks should be identied and repaired prior to closure. A helpful technique in locating potential CSF leaks is
to have the anesthesiologist provide a sustained positive
pressure Valsalva maneuver prior to closure. CSF leaks can
lead to large subgaleal collections and an increased risk of

Use of Wire and Metal Fixation Miniplates

Early on we used wire to stabilize the bone units. Due to unacceptable wire erosion through the skin, we have given up
the use of this material and have replaced it with absorbable
sutures and absorbable miniplates. Over the years we have


found the metal miniplates to also migrate in unacceptable

patterns in the growing child. As a result we no longer use
metal xation plates in a child under the age of 3. However,
the use of absorbable miniplates has proven to be helpful in
those areas where extra stabilization is needed.

There is a signicant incidence of hydrocephalus reported
in children with multiple suture closures. There are also differing views as to when to treat the hydrocephalus, that is,
before or after the craniofacial reconstruction. It has been
our policy to do the craniofacial procedure rst and shunt
second. We have found that in some cases the treatment of
the pansynostosis alleviates the need to treat hydrocephalus
(as was the situation in the rst case). Plus the early correction of hydrocephalus can cause a dramatic collapse of
the brain, which will work against the reconstruction. On
the other hand, if the child is very sick from increased ICP,
the shunt is placed to relieve ICP. Another factor to consider
in a secondary placement of the shunt is the increased risk
of infection due to the shunt being present during a lengthy
This chapter is a revision of the chapter, Pansynostosis:
Surgical Management of Multiple Premature Suture Closure by James T. Goodrich, M.D., and Craig D. Hall, M.D.
The chapter appeared in the Neurosurgical Operative Atlas,
Volume # 2, edited by Setti S. Rengachary and Robert H.
Wilkins. The Neurosurgical Operative Atlas was published by
the American Association of Neurological Surgeons (AANS)
from 1991 to 2000.
We would like to acknowledge and thank Craig Hall, M.D.,
for his help and efforts on the original chapter published in
the rst edition of this work.

Tethered Spinal Cord, Intramedullary Spinal
Lipomas, and Lipomyelomeningoceles
Elizabeth C. Tyler-Kabara and W. Jerry Oakes
Progressive dysfunction of the spinal cord may be caused by
xation or by compression from a neoplasm. This chapter
focuses on three congenital entities that cause symptoms
by different mechanisms. The patient with a tethered spinal
cord develops symptoms from tension by a thickened and
taut lum terminale on the distal spinal cord. Intramedullary spinal lipomas distort and compress the surrounding
cord but do not have a component of xation that causes the
neurological deterioration. The lipomyelomeningocele has
components of both cord xation and compression, which
is thought to be the underlying explanation of the progressive loss of neurological function associated with this lesion.
Because the surgical approaches to these lesions are quite
different they will be discussed separately.

Tethered Spinal Cord

The concept that xation of the distal spinal cord by a thickened lum terminale can cause progressive spinal cord
dysfunction has become accepted during the past 50 years.
Surgical procedures to section the lum have only become
common in the past 30 years. The dysfunction of the spinal
cord occurs from a combination of repeated small injuries or
contusions as the cord is put under tension with movement
(especially spinal exion) and by apparent vascular compromise. The change in the supercial vasculature of the cord as
well as its disturbed ability to metabolize oxygen has been
demonstrated both in laboratory animals and in humans.
The vascular changes that occur with tension of the cord
seem to improve following release of the tension.

Patient Selection
Patients may present for clinical attention with a variety of
complaints. One group of patients will present for clinical
attention only as a result of cutaneous evidence of occult
spinal dysraphism. This may take many forms including a
at capillary hemangioma or a small dermal appendage.
When one of these cutaneous ndings is present, it suggests further investigation. The majority of patients with
a at capillary hemangioma will have normal intradural
anatomy; however, a signicant minority (as high as 10%)
will have a tethered spinal cord. Waiting for the patient to
demonstrate the clinical ndings of a neurogenic bladder is

likely to result in a xed neurogenic decit that may not be

reversible with surgical intervention. Prevention of further
bladder dysfunction by early investigation and prophylactic
surgery is warranted.
Commonly, the initial clinical symptom in patients with
tethered cord syndrome is the gradual and progressive loss
of coordinated bladder activity. This may become manifest
as repeated bouts of urinary infection or primary or secondary urinary incontinence. The urinary symptoms may
be combined with evidence of spasticity of the lower extremities. The lower extremity involvement is frequently
a combination of hyperactive deep tendon reexes with
upgoing toes and muscle wasting, fasciculations, and shortened foot or leg length. The combination of upper and lower
motor neuron disturbances in the lower extremities is the
signature of this problem. Even though the tension could be
thought to be relatively symmetric on the distal cord, the leg
ndings are typically asymmetrical. Rectal incontinence is
usually delayed until late in the course. Nonradicular pain in
the back and legs may be the primary cause of presentation
in the adult population. Occasional precipitous deterioration has been recognized. Sudden lower spinal exion as in
patients assuming the lithotomy position under anesthesia
has resulted in dramatic worsening and even abrupt onset
of paraplegia in occasional patients.
In at least 10 to 20% of patients with congenital anorectal
atresia, a tethered spinal cord will also be found. This group
of patients clearly warrants early investigation in an effort
to prevent the development of a neurogenic bladder and to
improve the neurological input into the dysmorphic rectal
sphincter. Many patients with a thickened lum terminale
are seen to have some other form of neural tube abnormality. This may range from an obvious myelomeningocele to
the combination of intact normal skin and an occult split
cord anomaly or terminal syringohydromyelia.
The suspicion of a tethered spinal cord is conrmed with
magnetic resonance imaging (MRI). The use of computed tomography (CT) myelography in this patient population has
been almost entirely replaced by MRI. In general, three criteria are necessary to radiographically conrm the clinical
impression of a tethered spinal cord: caudal descent of the
conus, fatty inltration and thickening of the lum terminale, and a drawn-out or funnel-like appearance of the distal conus. Normally the conus should not descend below the
L1-2 disk space. The conus progressively ascends within the


14535_C10.indd 62

4/23/08 3:15:58 PM


Tethered Spinal Cord, Intramedullary Spinal Lipomas, and Lipomyelomeningoceles

spine throughout embryological development and early infancy. It normally attains its adult position by 3 to 6 months
of age and maintains this position throughout adult life. In
patients with tethered cord syndrome, the thin delicate structure of the normal lum terminale is lost. It becomes thickened (usually more than 3 mm in diameter) and is typically
inltrated by fat, which is easily appreciated by MRI on the
axial T1 sequences. This thickening results in tension on the
conus and a loss of the normal bulbar lumbar enlargement of
the distal cord. The cord assumes a funnel-like appearance,
dorsally displaced within the subarachnoid space and under
tension. When all radiographic ndings occur in a patient
with an appropriate clinical setting, the diagnosis is secure.
Unfortunately, there are occasions when patients present
with progressive clinical symptoms that could easily be attributable to a tethered cord, but the radiographic ndings
are confusing. The conus may be in a normal position but the
lum may be thick, inltrated by fat, and dorsally displaced
under tension. As with many clinical situations, judgment in
analyzing the clinical and radiographic ndings is essential
to arrive at an appropriate decision for the patient.
The natural history of the condition is not well understood. Many believe that the constant small trauma associated with tension on the distal cord by the thickened lum
is associated with the relentless loss of neurological function in most patients. This loss may occur within the rst
few months of life or more typically over a much longer
time. Occasional adult patients will demonstrate many
decades of symptom-free survival only to come to clinical
attention with irreversible bladder dysfunction. Pain-free
adolescents and adults with normal neurological function
but clear radiographic abnormalities attest to the incomplete ability of the clinician to predict the natural history
of this condition. Patients with progressive symptoms and
classic MRI changes are easily counseled to accept operation
in an attempt to prevent further loss of function. Asymptomatic patients can be approached with a risk-versus-benet
analysis depending on the strength of the radiographic abnormalities and other evidence of dysmorphism, that is,
anorectal atresia, hemivertebrae, etc.


Figure 101 (A) Prone position of an infant undergoing sectioning of
the lum terminale. (B) The typical skin incision for exposure of the
sacral dura. The skin incision extends from the spinous process of L5
to the mid sacrum.

Once a decision to operate has been offered and accepted,

the patient is positioned prone with bolsters placed under
the iliac crest and thorax to allow free excursion of the abdomen (Fig. 101A). Some surgeons feel more comfortable
utilizing intraoperative monitoring of rectal and/or urethral
electromyography. Ultra-short-acting muscle paralysis is
necessary to allow appropriate interpretation of these electrical parameters.

space to allow adequate visualization of the distal lum

terminale. The dura is opened in the midline and the arachnoid is opened somewhat off the midline. The thickened
lum can be recognized by the inltration of fat within it,
by its midline dorsal position, and by the size differential
between this structure and the surrounding thin delicate
roots that exit in a ventrolateral direction. Additionally, the
lum has a characteristic vessel on its ventral surface that
may be appreciated by rotating the lum. Once the lum is
identied it is carefully separated from all adherent nerve
roots, particularly those on the ventral surface (Fig. 102A).
One must rotate the lum to visualize the undersurface and
to ensure that no small roots are left adherent to the capsule
of the lum. When this is conrmed the lum is coagulated
and cut (Fig. 102B). This is done at two levels to allow the
excision of a section of 5 to 8 mm that can be sent for histological analysis and to ensure that the two ends of the lum
will not adhere to one another. Care is taken to ensure a watertight closure of the dura and the supercial soft tissues.

Operative Procedure

Postoperative Management

An incision is made from the spinous process of the tip of

the L5 vertebra to the mid sacrum (Fig. 101B). The dorsal
bony elements of S1 are removed by standard laminectomy
technique, and the sacral dura is exposed. There is rarely
need to expose more than 2 to 3 cm of the subarachnoid

Patients are routinely nursed at for several days to allow

adequate dural healing without the additional tension of
orthostatic force of the cerebrospinal uid (CSF). The likelihood of adherence of the cut ends of the lum is extremely
small. Long-term complications are very rarely seen.

Preoperative Preparation

14535_C10.indd 63

4/23/08 3:15:59 PM

64 Pediatric Neurosurgery
congenital lesions that may lay dormant for many years or
decades. They are not associated with spina bida occulta
and are relatively evenly distributed in the cord. There is a
predilection for involvement of the dorsal aspect of the cord,
but the fat is not connected with a defect of the arachnoid,
dura, laminae, or skin. When symptomatic, these lesions
are approached as any other intramedullary tumor with a
laminectomy over the involved area.

Operative Procedure
Using an ultrasonic aspirator or a CO2 laser at very low wattage, the fat can be removed and the brous septa between
the various compartments of fat can be disrupted. The cord
may be reconstructed into a tube following the resection
of the fat and the redundant dura is closed primarily and
tented dorsally in an effort to prevent adhesion from the
resection site to the undersurface of the dura.
The outlook for these patients is a function of the degree
of neurological disability prior to surgery and the success
of the resection. The goal of the operation is not to remove
every pocket of fat but rather to debulk signicant mass effect leaving the interface between the spinal cord and the
lipoma unmanipulated.


Figure 102 (A) Axial views through the sacrum before and after the
adherent roots ventral to the thickened lum terminale have been dissected. (B) The initial section through the lum terminale showing fat
in the center of the lum.

Results with regard to blunting of the previous rate of

loss of neurological function are quite favorable. The ability to restore lost function, particularly bladder synergy, is
poor. Relief of pain and sensory loss is gratifying. Progressive moderate scoliosis may be arrested. Restoration of motor function is less likely than sensory improvement or the
relief of pain. Urologic dysfunction, which is the most common cause of clinical presentation, remains the least likely
to improve once it is rmly established.

Intramedullary Spinal Lipoma

Patient Selection
Intramedullary fat that accumulates under the pia of the
spinal cord may cause the patient to present for clinical
attention with chronic myelopathy or pain. These are rare

14535_C10.indd 64

Subcutaneous fatty tumors over the midline lumbosacral

region usually emerge through a fascial, bony, or dural defect and ultimately involve the caudal spinal cord. These
lesions are best described by where they grow into the cord
(dorsal or caudal). Lipomyelomeningoceles are part of the
larger group of conditions termed occult spinal dysraphism
and may be found in conjunction with a split spinal cord,
tethered spinal cord, or other forms of this broader group.
They do seem to have an enhanced genetic predisposition
and may be associated with a Chiari I malformation or occasionally hydrocephalus. The special circumstance of a fatty
mass associated with a terminal syringohydromyelia will
be discussed later.

Patient Selection
Symptoms related to these lesions range from a cosmetic
presentation at birth with a signicant subcutaneous fatty
mass in the midline over the lumbosacral region to a subtle
subcutaneous lipoma associated with primary or secondary
urinary incontinence. It is unusual today to have children
present primarily with evidence of pain or motor disturbance involving the lower extremities. In large part this is
due to an increased awareness of the progressive natural
history of this lesion and the ease of conrmation of spinal
cord involvement by MRI. Decreased or absent rectal tone
combined with a neurogenic bladder is a clinical indication
of lower sacral root involvement from a neurogenic cause
and should be investigated further.
Today the procedures of choice to conrm the pathological anatomy are MRI and radiographs of the lumbar spine.
These two modalities have almost totally replaced CT my-

4/23/08 3:15:59 PM


Tethered Spinal Cord, Intramedullary Spinal Lipomas, and Lipomyelomeningoceles

elography. Additionally, useful information can be obtained

from urodynamic testing and occasionally from electromyography of the lower extremities.
The natural history of patients with a lipomyelomeningocele is a progressive loss of neurological function. Occasional exceptions to this natural history can be demonstrated,
but the vast majority of infants followed prospectively
without operative treatment will develop progressive and
relentless deterioration. Some forms of lipomyelomeningoceles can more clearly be positively inuenced by surgery
than others. These would include simple dorsal tubes of
fat having a discrete point of xation with the spinal cord.
Lesions associated with a blending of the neural elements
with the lateral aspect of the dural cul-de-sac benet less
certainly from operative intervention. Although most North
American pediatric neurosurgeons continue to favor early
intervention for all infants, other well-respected pediatric
neurosurgeons have questioned this approach and may wait
for the individual patient to demonstrate clear loss of neurological function before recommending intervention. Surgical exploration should include release of the traction on the
cord and resection of a signicant amount of the intramedullary lipoma. Reversal of bladder dysfunction is unlikely
once it is established. For this reason alone, early aggressive
surgery is logical and justied if felt to be feasible. It should
be mentioned that there is no place for a supercial excision
of the subcutaneous lipoma without disturbing the intradural contents. A supercial excision seriously complicates


the eventual intradural procedure by the development of

signicant intradural adhesions.

Preoperative Preparation
A patient chosen for exploration of the lesion is positioned
prone with soft supports under the iliac crest and chest to
allow free excursion of the abdomen. This signicantly lessens epidural bleeding. The hips are exed and, again, intraoperative rectal and urethral monitoring may be utilized.

Operative Procedure
The skin incision is made in the midline directly over the
subcutaneous mass and extends both above and below the
mass (Fig. 103A). A common error is not to allow sufcient
distal room for adequate exposure of the lesion. For routine
lesions it is necessary to have access from the lower sacrum to two segments above the level of the fascial defect.
The subcutaneous lipoma is almost always easily separable
from the lumbodorsal fascia and the skin. As the neck of
the lesion is circumferentially developed, large amounts
of supercial lipoma may be excised, reducing the bulk of
the lesion. The neck of fat coursing through the fascial defect is retained at this point (Fig. 103B). The soft tissues
and muscle adherent to the last intact spinous process and
laminae are reected laterally (Fig. 103C). The muscles and



14535_C10.indd 65

Figure 103 (A) A typical skin incision over the dome of a

lipomyelomeningocele. (B) The subcutaneous aspects of the
lipoma dissected, the neck of the lesion coursing through the
fascial defect is fully developed. Excessive lipoma can be removed safely to allow additional exposure (dashed lines). (C)
After excessive circumferential fat is removed, the paraspinal
muscles are then dissected off the last intact lamina. The constricting periosteal band immediately cephalad to the neck of
the lipoma can then be sectioned.

4/23/08 3:16:00 PM

66 Pediatric Neurosurgery
other soft tissues circumferentially adherent to the rudimentary laminae surrounding the neck of the lesion are also
dissected. Immediately caudal to the last intact laminae a
band of brous tissue corresponding to the periosteum of
an incompletely formed bony element will commonly be
encountered. With sectioning of this band the dura may expand signicantly into the area. This band may be associated
with acute angulation of the malformed cord as it is drawn
dorsally toward the subcutaneous lipoma. Sectioning this
band may signicantly relieve the tension on the cord and
reduce its posterior angulation (Fig. 103C).
With adequate exposure of the dural tube cephalad to the
lipoma, it is opened in the midline. Asymmetrical exiting
roots can then be seen. The junction of the dura to the neck
of the lipoma as it emerges through the dural defect is a key
landmark for further dissection. Special care must be given
at this point to appreciate the relationship of the dorsal
roots that have been displaced laterally by the lipoma and
the dura-lipoma complex (Fig. 104A). Obviously, no roots
should be sacriced, and yet the dura needs to be circumferentially dissected away from the dorsally displaced cord
(Fig. 104B). The asymmetrical arrangement of exiting roots
combined with rotation of the cord may further complicate
this maneuver. Once this critical maneuver is complete and
the cord has been moved into a relaxed ventral position
within the dura, attention is turned to the intramedullary
component of the lipoma. With the ultrasonic aspirator or
the CO2 laser, the lipoma is progressively thinned until a

smaller layer of fat remains against the neural tissue. No

attempt is made to completely excise all fat, but the bulk
of the lesion is removed, which then allows reconstitution
of a neural tube in most patients. This is done with ne
inverted nonreactive suture (Fig. 105). Attention is then
directed to the lum terminale, which may be thickened.
Again, it is separated from the surrounding exiting roots
and sectioned.
The dura is reapproximated, allowing a capacious CSF
space dorsal to the newly formed neural tube. The supercial soft tissues are reapproximated, but not at the expense
of the dorsal CSF space.
Numerous variations on this anatomical theme exist. One
that deserves special comment occurs when the subcutaneous lipoma is asymmetrically situated off the midline and
opposite to it is a cleft of grayish blue, thin epithelium. The
thinned epithelium is seen to be blue from an underlying
CSF collection being formed by a terminal syringohydromyelia (Fig. 106).
Operatively, this lesion is approached in a similar manner,
with circumferential dissection of the subcutaneous lipoma
and cyst. With the last intact lamina removed, the dura is
opened to expose the caudally displaced spinal cord. The
plane between the dura-lipoma complex and the fatty inltrated cord is again key. The cord may be quite expanded by
the combination of fatty inltration and enlargement of the
central canal. Draining the cyst at this point may facilitate
the circumferential dissection around the dural attachment


14535_C10.indd 66

Figure 104 (A) A three-dimensional view of a lipomyelomeningocele

demonstrating the dura-lipoma complex and the position of the exiting dorsal roots. This relationship varies somewhat from patient to
patient and with the degree of rotation of the cord. (B) A surgical view
showing a hooked knife being used to open the dura while the cord is
rotated slightly for better exposure of the dorsal roots. The upper part
of the lipoma is also being excised by the CO2 laser.

4/23/08 3:16:02 PM


Tethered Spinal Cord, Intramedullary Spinal Lipomas, and Lipomyelomeningoceles

Figure 105 Following resection of the bulk of the intramedullary

lipoma, the distal cord is reconstituted into a neural tube with inverted
ne nonreactive sutures (upper inset). Last, the thickened lum terminal

to the neck of the lipoma. The dura, too, may be inltrated

by fat. This makes adequate dural closure technically demanding. With careful dissection and constant changing of
perspective, eventually the circumferential opening of the
dura-fat interface can be accomplished. The dorsal roots
may be seen to be adherent to the undersurface of the dura,
and prior to each maneuver they should be sought. These
roots are laterally displaced and will be seen to exit in an
asymmetrical fashion. Excessive fat growing into the neural
tissue will then be excised. The very caudal portion of the
conus can be reconstructed into a neural tube. This may
lessen the likelihood of secondary readherence at the site of
separation. At times, a dural graft may be necessary. This can
usually be harvested from the two-ply lumbodorsal fascia.
Again, watertight dural closure with preservation of a capacious CSF space around the reformed neural tube is ideal.

14535_C10.indd 67


is sectioned. The dura is closed and a capacious CSF space dorsal to the
newly formed neural tube is created (lower inset). This lessens the likelihood of retethering at the operative site.

The soft tissue closure supercial to the dura is performed

in such a manner that the intradural contents are not constricted and, therefore, are less likely to become adherent.

Long-term results following aggressive resection and repair of lipomyelomeningoceles are still being accumulated.
What does seem clear is that the risk of a serious permanent
injury from the operative manipulation is low in experienced hands and should be much less than 10%. The risks of
spontaneous worsening without operation are high, probably greater than 90% within the rst two decades of life.
The likelihood of 5 to 10 years of clinical stability without
further loss of neurological function following surgery is

4/23/08 3:16:02 PM

68 Pediatric Neurosurgery

Figure 106 (A) Surface landmarks of a cystolipomyelomeningocele. Asymmetrical position of the lipoma to the right of
midline and a thin epithelial veil over the terminal syringohydromyelia on the left are apparent. (B) Anatomical relationships
demonstrating the explosive expansion of the caudal lipoma
and its inltrative nature with respect to the dura and terminal
cord. Expansion of the distal central canal (terminal ventricle)
into a syringohydromyelia is easily appreciated.

also high. If these are the representative risks and benets,

then early surgical intervention is reasonable and appropriate for most patients with this lesion.
There are many anatomical variations within this category of congenital lesions. The experience and judgment
necessary in the successful operative manipulation of this

14535_C10.indd 68

lesion are signicant. This particular lesion should not be

operated on by the surgeon who deals with it only occasionally. Serious loss of bladder and/or bowel function will
occur with the inadvertent sacrice of functioning nerve
roots by the surgeon who does not perform this procedure

4/23/08 3:16:03 PM

Encephaloceles of the Anterior Cranial Base
Jonathan P. Miller and Alan R. Cohen

The term encephalocele, derived from the Greek for hernia of

the brain, is used to describe a congenital protrusion of intracranial contents beyond the normal connes of the skull.
Encephaloceles are relatively rare malformations, occurring
with a worldwide incidence of ~1 per 5000 live births.
Because encephaloceles, which pass through the same
internal skull defect, may project to different locations externally, the terminology used to describe them is often
confusing and inconsistent. Conventionally, sac locations
are described in relation to internal defects in the skull, not
by proximity to external landmarks. Using this scheme, encephaloceles can be divided into two broad categories: those
related to the cranial vault and those related to the cranial
base. Cranial vault encephaloceles, which occur when the
defect is on the outer surface of the skull, present as visible
external masses and usually occur along the midline in the
frontal, parietal, or occipital region, or rarely off the midline
at the pterion. These are more common in females and account for the vast majority of encephaloceles encountered
in the Western Hemisphere; the occipital encephalocele is
the most frequent type seen in North America. Cranial base
encephaloceles result from a defect in the basal skull. They
are more common in males, and they are seen mostly in
Southeast Asia, especially Malaysia, Burma, Indonesia, and
Thailand, where they outnumber posterior cranial vault
encephaloceles 9.5 to 1. Cranial base encephaloceles are
subclassied into two types: frontoethmoidal (also called
sincipital), which project forward and produce a mass on
the face, and basal, which project downward and produce a
mass in the nasopharynx.
The frontoethmoidal encephalocele (Fig. 111) projects
anteriorly and is seen as a visible defect on the face at the
root of the nose, which accounts the term sincipital, derived
from the Latin for the anterior and upper part of the head.
The hernia sac usually passes through an enlarged foramen
cecum, a normally blind opening between the frontal crest
anteriorly and the crista galli posteriorly. Frontoethmoidal
encephaloceles are subdivided anatomically into three types
based upon the site of the external skull defect opening
into the face: nasofrontal, nasoethmoidal, and naso-orbital.
Nasofrontal encephaloceles project onto the face at the nasion through an opening between the frontal bone above
and the nasal and ethmoid bones below. This is the site of
the primitive fonticulus nasofrontalis, a brous membrane

bridging the nasal and frontal bones during embryological

development that separates the dura from the developing
skin. Nasoethmoidal encephaloceles are long-necked and
project somewhat more inferiorly, between the frontal and
nasal bones above and the displaced ethmoid bone and nasal cartilage below. Naso-orbital encephaloceles are rare and
project through the inferomedial wall of the orbit, passing
between the frontal process of the maxillary bone anteriorly
and the lacrimal bone and lamina papyracea of the ethmoid
bone posteriorly.
By contrast, basal encephaloceles (Fig. 112) protrude
downward into the nasopharynx, either directly through the
cribriform plate or body of the sphenoid bone, or through
an enlarged foramen cecum. Because herniation occurs internally, there may be no obvious visible external mass and
clinical presentation is often more insidious. They usually
present as nasal (or rarely pharyngeal) mass, with symptoms of nasal airway obstruction. Also, because the defect is
close to the suprasellar cistern, these are much more likely
to contain important neurovascular structures than frontoethmoidal encephaloceles. Basal encephaloceles are variously divided into subcategories such as sphenopharyngeal,
spheno-orbital, sphenomaxillary, sphenoethmoidal, and
transethmoidal, depending upon the exact location of the
defect. Other midline facial anomalies may be seen, such as
cleft lip, cleft palate, or hypertelorism; they are sometimes
discovered incidentally during craniofacial reconstruction
for hypertelorism.
The contents of the encephalocele sac are quite variable.
Most often, cerebrospinal uid (CSF) and gliotic brain are
found, but large amounts of brain tissue are not uncommon, and many critical structures such as the hypothalamus, optic apparatus, and anterior cerebral arteries have
been encountered in large herniations. Unlike those of the
cranial vault, anterior cranial base encephaloceles are not
usually associated with neurological abnormality, although
hydrocephalus and epilepsy have been reported. They are
sometimes associated with other abnormalities such as microphthalmia, corneal opacity, coloboma, craniosynostosis,
and corpus callosum agenesis or lipoma.
The purpose of this chapter is to describe the clinical
management of encephaloceles at the anterior skull base.
Cranial vault encephaloceles are discussed elsewhere in this


14535_C11.indd 69

4/11/08 11:30:02 AM

70 Pediatric Neurosurgery

Figure 112 Basal encephalocele. The stalk passes through a defect

in the oor of the anterior fossa at the site of the foramen cecum. The
sac terminates in the nasal cavity between the septum medially and
the middle turbinate laterally.

Preoperative Preparation

Figure 111 Frontoethmoidal (sincipital) encephalocele; three variants. Each of these projects anteriorly to create a visible deformity at
the root of the nose. (A) Nasofrontal encephalocele passing between
the frontal bone above and displaced nasal and ethmoid bones below.
(B) Nasoethmoidal encephalocele passing between the frontal and
nasal bones above and displaced ethmoid bone and nasal cartilage
below. (C) Naso-orbital encephalocele projecting externally through
the medial wall of the orbit.

14535_C11.indd 70

The most common clinical presentation of a frontoethmoidal encephalocele is a visible protrusion on the face at the
root of the nose. Basal encephaloceles, on the other hand,
are often more insidious, and they can present with nasal
obstruction, mouth breathing, snoring, CSF rhinorrhea, recurrent meningitis, or nasal discharge due to purulence in
the nasopharynx. The encephalocele sac is visible in the nasal cavity, lying between the septum medially and the middle nasal turbinate laterally. It may be visualized as a glistening pink or blue intranasal mass that pulsates with the
heartbeat and may swell with crying or Valsalva maneuver.
Unique features of nasal encephaloceles are their tendency
to enlarge with compression of the jugular veins (so-called
positive Furstenberg test) and their tendency to be crossuctuant with compression of the anterior fontanelle.
Encephaloceles must be differentiated from other nasal
masses. Nasal polyps are rare in infants and are located
lateral rather than medial to the middle turbinate. Other
congenital nasal lesions in the differential diagnosis include
dermoids, lipomas, bromas, hemangiomas, teratomas, and
nasopharyngeal cysts. In older individuals, basal encephaloceles must be distinguished from a variety of benign and
malignant skull base tumors, as well as inammatory lesions such as mucoceles and granulomas of the nose and
paranasal sinuses. Some authors distinguish the nasal en-

4/11/08 11:30:02 AM

11 Encephaloceles of the Anterior Cranial Base


cephalocele from a related clinical entity, the so-called nasal

glioma, although these lesions appear to share a common
embryology. The nasal encephalocele is distinguished from
a nasal glioma by the presence of a stalk connecting the
nasal mass with the brain. This distinction is sometimes
difcult to make clinically, but it may be signicant as nasal
gliomas (and not encephaloceles) may be biopsied safely via
a transnasal approach.
Radiographic demonstration of an anterior skull-base
encephalocele is best performed with magnetic resonance
imaging (MRI). The midsagittal image provides a beautiful
demonstration of the hernia sac passing through a defect in
the frontal oor to terminate in the nasal cavity. MRI is also
useful for detecting the rare instance in which important
neurovascular structures pass into the encephalocele sac.
Angiography is not routinely necessary but may be considered if MRI suggests involvement of the anterior cerebral
arteries in the herniation. Computed tomography (CT) is
often helpful to delineate the bony anatomy, and threedimensional CT can be helpful for planning more complex
craniofacial repairs.

Operative Procedure
Historically, anterior cranial base encephaloceles have been
approached both transnasally and transcranially. The transnasal route is now rarely performed because of signicantly
higher risks of bleeding, infection, and CSF stula formation.
The transcranial approach was rst described by Dandy in
1929, and variations are still commonly used. An encephalocele with a relatively narrow neck and a sac that terminates
in the nasal cavity can be approached either intradurally or
extradurally, and there is continued controversy over which is
the better way. If there is a broad stalk or inclusion of critical
neurovascular structures within the encephalocele, a combined extradural-intradural exploration may be appropriate.
Encephaloceles associated with hypertelorism or those projecting onto the face are more complex and are best repaired
in conjunction with a craniofacial team.
For both the intradural and extradural approaches, preoperative nasal cultures are obtained and the patient is given
anticonvulsants and antibiotics appropriate for staphylococci and streptococci. The operation is performed under
general anesthesia with oral endotracheal intubation. Arterial, venous, and Foley catheters are inserted. Although the
procedure is not usually associated with signicant blood
loss, even minor bleeding may have serious consequences
in young infants, so blood is made available for transfusion.
In older infants and adults, a lumbar subarachnoid catheter
is used for intraoperative drainage of CSF. The patient is
positioned supine on a warming blanket and covered with a
regular blanket to minimize heat loss. The head is supported
by a doughnut cushion, elevated 20 degrees, and extended
with the brow up. Adults are positioned in a similar fashion using a Mayeld three-pin xation device (Schaerer
Mayeld, Cincinnati, Ohio). The scalp is then shaved and
cleansed with alcohol followed by povidone-iodine soap
and solution. A coronal incision well behind the hairline
is outlined with a marking pen (Fig. 113). The operative

14535_C11.indd 71

Figure 113 Patient positioned for bifrontal craniotomy. A coronal

scalp incision is outlined (dashed line).

eld is walled off with towels and covered with an iodoform-soaked adhesive drape. The scalp is inltrated with
a solution of 0.1% lidocaine with a 1:1,000,000 dilution of
epinephrine. The incision is made with a no. 10 scalpel, and
scalp hemostasis is obtained using Raney clips and Dandy
clamps. Alternatively, in younger infants the scalp may be
opened using a Shaw hemostatic scalpel. The scalp is reected anteriorly in the avascular loose areolar plane, which
lies deep to the galea but supercial to the pericranium.
Great care is taken not to disturb the pericranium, which is
mobilized as a separate vascularlized layer to reconstruct
the skull base at the close of the procedure. The pericranium
is incised with a knife and reected anteriorly using squareended periosteal elevators. The pericranial ap is based on
a vascular pedicle supplied by the frontal and supraorbital
arteries. Any holes in the pericranial ap are repaired at this
point using 40 Nurolon sutures. The scalp and pericranial
graft are covered with moist sponges.
For the extradural or combined approach, a bifrontal free
bone ap is fashioned (Fig. 114). Bur holes are placed in the
keyhole region bilaterally, and a small bur hole is placed in the
low frontal midline using a high-speed drill (e.g., Midas Rex;
Medtronic Inc., Fort Worth, Texas). The frontal sinus does not
present a problem in infants because it is not yet pneumatized.
A posterior midline bur hole can be avoided by using a knife
and dental instrument to create suturotomies and separate

4/11/08 11:30:04 AM

72 Pediatric Neurosurgery

Figure 114 A bifrontal bone ap is (A) created and (B) removed. Note
the preservation of a vascularized pericranial ap, which is projected anteriorly prior to the craniotomy. After removal of the encephalocele, this

dura from bone on either side of the anterior fontanelle. The

bone ap is elevated using the high-speed drill and footplate
system to connect the holes, with cuts directed away from
the midline. The dura is tented up to the posterior margins
of the craniotomy defect using 40 Nurolon sutures. The
dura is then dissected gently off the oor of the anterior
cranial fossa bilaterally using a Freer elevator or a bayonet
forceps and cottonoid patties. The dissection is usually not
difcult, but it can be troublesome medially in the region
of the cribriform plate. When necessary, the dissection may
be facilitated by the use of hyperventilation, mannitol, and
drainage of CSF from the lumbar catheter to relax the brain.
The surgeon can then identify the encephalocele stalk as it

14535_C11.indd 72

ap is sutured to the basal frontal dura and serves as a barrier between

the remnant of the encephalocele stalk above and the bony defect
opening into the nasal cavity below.

enters the bony defect at the foramen cecum (Fig. 115). The
stalk is dissected circumferentially. A 40 Nurolon sutureligature is placed through and around the encephalocele
stalk, and the stalk is divided with scissors just distal to the
site of ligation (Fig. 116). The intracranial portion of the
encephalocele is thereby disconnected from the intranasal
portion of the sac, which may be grasped from below by an
otolaryngologist and delivered through the nose (Fig. 117).
If the nasal sac is tenacious, it is left alone and will either
shrivel up, requiring no further treatment, or alternatively
it may be removed easily after several weeks at a second
sitting. Hemostasis is secured, and the ligated basal frontal
dura is covered with a piece of thrombin-soaked Gelfoam,

4/11/08 11:30:06 AM

11 Encephaloceles of the Anterior Cranial Base


Figure 115 Extradural approach. The encephalocele stalk is identied and isolated.

which may be inltrated with brin glue. Bony reconstruction of the skull base is usually not necessary, but very large
defects may benet from a graft to support the dural repair
and lessen the risk of a postoperative CSF leak. The vascularized pericranial graft is brought down to cover the skull
base and is xed to the frontobasal dura with 40 Nurolon
sutures (Fig. 114B). After replacement of the bone ap, the
scalp is closed in two layers.
For the intradural approach, a unilateral craniotomy will
sufce. The craniotomy is therefore made to midline and
extended low toward the cranial base. The dura is opened,
and the brain is gently retracted upward to reveal the root of
the defect (Fig. 118). The stalk is amputated; at this point,
the encephalocele can be pulled out through the nose by an
otolaryngologist. The space at the base of the skull is then
covered with a small piece of fat or Gelfoam and brin glue.
A vascularized pericranial graft can be used to reinforce the
repair. The dura is then closed in a watertight fashion, and

14535_C11.indd 73

the bone ap and scalp are closed. In older children and

adults, the lumbar subarachnoid catheter is left in place for
up to 5 days to protect the operative repair.

Postoperative Management Including

Possible Complications
Complications are generally related to infection or the formation of a CSF stula, both of which may occur as a result
of communication between the subarachnoid space and the
nasal cavity. The risk of such a communication can be minimized by a meticulous dural closure protected by a vascularized graft of pericranium. For the same reason, a subgaleal
drain should not be placed. If a CSF stula develops, lumbar
drainage for a few days will often solve the problem, but if
not, the wound should be explored surgically.

4/11/08 11:30:07 AM

74 Pediatric Neurosurgery

Figure 116 (A,B) The stalk of the encephalocele is ligated with a suture ligature and (C,D) is transected.

Encephaloceles of the anterior cranial base are rare malformations that can present insidiously. Straightforward
lesions may be approached intracranially using either an
extradural or an intradural dissection, whereas more complex lesions may require a combined intradural-extradural

14535_C11.indd 74

approach. Goals of surgery should be protection of neurovascular structures and prevention of a CSF stula. Although
they are more difcult to repair than the posteriorly situated occipital encephaloceles, they are associated with a
much better neurological prognosis, and outcome for most
patients is excellent.

4/11/08 11:30:10 AM

11 Encephaloceles of the Anterior Cranial Base


Figure 117 (A) The sac of the encephalocele is removed transnasally (B) after its
stalk has been ligated and divided intracranially.

Figure 118 Intradural approach. After

unilateral craniotomy, the dura is opened
and the frontal lobe gently retracted to
reveal the encephalocele.

14535_C11.indd 75

4/11/08 11:30:11 AM

Constance M. Barone, David F. Jimenez, and James Tait Goodrich

Patient Selection
Exorbitism or proptotic eyes secondary to craniofacial dysostosis is a consequence of shallow orbital sockets and is
not due to enlargement of the orbital contents (Figs. 12
1A121C). Therefore, corrective surgery for this problem consists of increasing the depth of the orbital cavities.
This correction may be undertaken in two operations: as a
fronto-orbital advancement done at 3 to 6 months of age,
and later as a midface advancement done after the age of 4
years, preferably during adolescence. The midface operation
should not be performed on the very young because their
tooth buds are located high up in the maxilla and the osteotomies will disrupt them, leading to severe dental problems as the child grows. Some surgical teams have chosen
to perform both operations as a single-step procedure, that
is, the monoblock advancement. There is a much greater risk
of infection with the monoblock procedure and that, along
with the high incidence of tooth bud disruption, has led us
to perform a two-stage procedure.
Exorbitism and midface retrusion are commonly observed
in craniofacial dysostosis, for example, Crouzons and Ap-

erts syndromes. Exorbitism, however, may occur without

any midface deformity or dental malocclusion, as is demonstrated in the patient presented here. Correction in this particular case could be accomplished by orbitonasal osteotomies through bicoronal, lower lid, and buccal incisions. The
principle in such surgical procedures is to restore a satisfactory aesthetic balance by recontouring the face with both
advancements of bone units and onlay grafts if necessary.
Our craniofacial team has elected to do our facial onlays
and reconstructions with calvarial bone. Because it is membranous, the use of this type of graft lessens signicantly
the risk of graft resorption, which occurs occasionally with
rib (endochondral bone) grafts. However, the techniques of
both calvarial bone and rib harvesting will be discussed.

Preoperative Evaluation
At our medical centers all patients with craniofacial disorders are evaluated at the respective Center for Craniofacial
Disorders. The patient is evaluated by specialists from the
elds of pediatrics, pediatric neurology, pediatric neurosurgery, genetics, plastic surgery, ophthalmology, dentistry,

Figure 121 (AC) Three views of a patient with exorbitism.



dentofacial surgery, otolaryngology, and speech therapy. All
patients have plain x-ray lms of the face and skull as well
as computed tomography (CT) scanning with bone windows
in axial and coronal planes. We now routinely use threedimensional CT scan reconstruction preoperatively. With the
software programs available we can perform measurements
and surgical osteotomies on the reconstructed CT model preoperatively. With this technique we are able to calculate the
amount of advancement as well as the amount of bone to be
harvested for the reconstruction. We have not found magnetic resonance imaging to be helpful in these cases.

Preoperative Preparation
We routinely start an anti-staphylococcal antibiotic at the
start of the operation. Because the surgical manipulations
are all extradural, we do not routinely use anticonvulsant
medications or steroids.
Because there is considerable bone harvesting and multiple osteotomies in these procedures, the surgical team
should be prepared for blood loss that is signicant enough
to require transfusion. If the family is interested, we routinely plan for pedigree (donor-directed) blood donations.
If available, a cell saver unit can rescue up to 50% of the
patients lost blood volume. Because of the risk of extensive
blood loss, all patients require at least two large-bore intravenous lines of 16 gauge or larger. If there is any history
of cardiac or pulmonary problems, we routinely put in a
central venous pressure line. An arterial line is mandatory
for monitoring blood gases, hematocrit, electrolytes, etc.,
during the procedure.
For an advancement of the orbital rims alone, 2 units of
packed red blood cells are made available. However, if the
midface is also included in the advancement (i.e., a Le Fort III
disconnection), 4 units of packed cells are made available. If
a Le Fort III disconnection is planned and the patient is to be
placed in intermaxillary xation, then a tracheostomy should
be considered. If the orbital rims alone are all that will be
advanced, as in this case, then we routinely use orotracheal
intubation. Ideally, the endotracheal tube should be sutured
to the gingiva, around a tooth using a heavy silk suture, to
prevent inadvertent extubation during the operation.
To protect the eyes from corneal abrasions and exposure
during the procedure, all patients have bilateral tarsorrhaphies prior to formal draping. If corneal shields are available, these may be used in place of the tarsorrhaphies.



is essential during the operation to provide various views

of the facial region. Anesthesia equipment is on the left side
of the patient, placed below the patients shoulder level.
The operating surgeon sits at the head of the patient with
the assistant to the right of the surgeon. The instrument
tray comes in over the patients abdomen but is positioned
no higher than the midthoracic region. The surgical team
therefore has access to both sides of the patients head as
well as intraoral access. For this reason we avoid the use of
bulky overhead tables that reduce access and visualization
of the face.

Because a bicoronal incision as well as intraoral and infraorbital incisions will be used, the entire face and mouth must
be prepared in addition to the scalp. The mouth is irrigated
copiously with Betadine (providone-iodine) solution, and
the nares are also cleansed with cotton swabs soaked in
Betadine solution. The head is shaved along the proposed
bicoronal incision for a width of ~1.5 cm. We no longer do
full-head shaving. Patients (or their parents) are quite appreciative of being able to comb the hair over the wound 1
week later. We have had no increased incidence of wound
infections since starting this technique more than ten years
ago. The head and entire face are then cleaned with Betadine scrub, cleansed with alcohol, and painted with Betadine solution.
In an adult, the bicoronal incision site is then injected
with 0.5% lidocaine and a 1:200,000 epinephrine solution. In a child (<7 years old) we use 0.25% lidocaine and a
1:400,000 epinephrine solution. Sterile towels are placed
below the posterior portion of the bicoronal incision and
are also stapled to the posterior aspect of the scalp. A facial
drape is then placed over the lower lip, chin, and endotracheal tube. It is important that the drapes are loose so that
the head can be moved during the procedure.
We routinely run all our suction lines, cautery cords, etc.,
toward the foot of the patient. Because the surgical team
sits, this allows easy mobility of the chairs (i.e., the chairs
are not rolling over the cords and tubes).
Because the operative site is heavily irrigated during the
procedure, it is important to have waterproof outer drapes.
Some of the newer drape designs have large plastic bags for
uid collection; we have found these to be quite useful.

Skin Incisions and Flap Elevation

Operative Procedure
The patient is placed in the supine position with the head
resting on a cerebellar (horseshoe) headrest. The head is
placed in a slightly extended, brow-up position. Rigid xation devices such as a Mayeld clamp are specically avoided
because the surgeon will need to move the head (although
usually never more than 10 to 151 degrees: This exibility

To gain access to the frontal bone, the supraorbital rims,

and the nasal bridge, a bicoronal scalp incision is used. This
incision is carried from tragus to tragus, well behind the
hairline. A full-thickness skin ap is elevated, leaving the
pericranium intact. The pericranial ap is then elevated
separately as a second layer. Frequently the supraorbital
nerve and artery are encased in a bony canal along the supraorbital rim; they can be freed by removing the anterior
wall of the supraorbital canal with an osteotome or small
Kerrison punch. The temporalis muscles are left intact and
are elevated in a fan-shaped fashion from each temporal

78 Pediatric Neurosurgery
fossa using a monopolar electrocautery. The dissection is
carried forward and over the supraorbital rims following
the subperiosteal plane and extending into the orbits to a
depth of ~1.5 cm. The lateral portions of the orbit and the
frontozygomatic process are also exposed on each side. To
gain access to the inferior orbital rims, a lower eyelid incision is made on each side at the junction between the lower
lid skin and the cheek skin. This incision is carried through
the orbicularis oculi muscle onto the infraorbital rim. Subperiosteally, the orbital contents are elevated off the orbital
oor to a depth of ~3 cm. The medial orbital contents as
well as the lateral inferior orbital contents are elevated in
a similar manner. Care is taken to preserve the infraorbital
nerve. As an alternative, a subciliary incision can be used;
however, this makes the dissection somewhat more difcult
and introduces the possibility of an undesirable postoperative ectropion and/or ciliary inversion. At this point, the entire orbital contents are freed from their bony encasement
in a subperiosteal plane.
Next an intraoral, transbuccal incision is performed that
extends across the midline of the pyriform aperture. The
pyriform aperture and maxilla are exposed following the
subperiosteal plane. Once again, care must be taken to protect the infraorbital nerves. Gentle retraction of the globes
should be observed at all times.

Calvarial Bone Harvesting

A large amount of calvarial bone is readily exposed via the
bicoronal incision. Calvarial bone can be harvested utilizing
one of two methods. The rst method requires a craniotomy.
A full-thickness piece of calvarial bone is removed and then
split along the diploic space using the Midas Rex drill with
a C-1 attachment and a reciprocating saw. The thickest bone
is found over the parietal region. Once the bone has been
split, the inner table is placed back in the donor site. The
outer table is then used for reconstruction (the outer table
has smoother contours, making it better for the aesthetic reconstruction). A second technique for bone harvesting does
not require a craniotomy. A trough is developed around the
edges of the proposed piece of bone. The trough is easily
created with an olive-tip drill. Curved osteotomes are then
used to split the calvaria along the diploic space, thereby
elevating the outer table. Care must be taken to avoid getting too deep because the dura can be torn inadvertently. It
is best to harvest a bone graft from the side of the nondominant hemisphere.

Rib Graft Harvesting

Another technique for bone harvesting is the rib graft technique. This is easily done via an inframammary incision.
Through this incision, adjacent ribs can be removed; usually
ribs six and seven are selected. The ribs are dissected by
rst incising the anterior periosteum and then using a Key
periosteal elevator to raise the anterior periosteum from the
rib, an Alexander elevator is used to strip the superior and
inferior borders of the rib free. A Doyen periosteal elevator
is then used to separate the posterior (or deep) periosteum

from the undersurface of the rib. Care must be taken not to

injure the underlying pleura.
Prior to closure, irrigation with saline is performed while
the anesthesiologist manually inates the lungs to check for
bubbling, a sign of an air leak. If a leak is found, the pleura
has been entered inadvertently during the dissection. The
tear should be repaired and a small red rubber tube (no.
10 French) placed into the pleural space. The tube is then
placed to low water-seal suction (~20 ml water pressure).
The donor site is closed in layers. When the patient is in
the recovery room, an x-ray lm of the chest is obtained.
If there is no evidence of a pneumothorax, the chest tube
is removed and an occlusive petroleum gauze dressing is
The harvested ribs are split along the superior and inferior
borders using an osteotome and contoured using a Tessier
rib bender. This technique doubles the available amount of
rib bone for grafting.

Craniofacial Osteotomies
With a reciprocating saw, the osteotomy is initiated across
the nasal radix and extended to the medial orbital walls
above the level of the lacrimal fossa. This area is easily approached via the bicoronal incision. Laterally, the osteotomy
extends on the frontozygomatic process. This is done in
a stepwise fashion, with the assistant using a malleable
retractor to protect and retract the orbital contents (Fig.
122). During this stage of the operation, the anesthesiologist should monitor the vital signs closely; if there is too
much traction on the globe, the blood pressure will increase
and the heart rate will decrease.
An osteotomy is then made across the orbital oor, back
~1.5 cm from the orbital rim. This osteotomy, most easily
done via the lower lid incision, is extended across the medial orbital wall, nally connecting the osteotomy lines of
the medial orbital walls together. During this dissection it
is mandatory to protect the orbital contents with a malleable retractor. Outside the orbits the lateral cut continues
along the body of the zygoma and then proceeds medially,
below the infraorbital foramen and toward the pyriform
aperture of the nose. This osteotomy is done through an
intraoral approach. The process is repeated on the opposite
side. The nasal septum is cut after a submucosal dissection
has been performed. All the osteotomy lines should now be
connected. The thin bones are easily fractured as the orbitonasal unit is mobilized in a forward and slightly downward
direction. The spur that is developed in the inferior segment
is made to abut against the rim of the superior segment
after the lower part of the orbit is advanced. Autogenous
bone blocks obtained from the split calvaria (or split rib)
are impacted into the nasofrontal and zygomatic defects
and secured in place using miniplates, microplates, or wires
(Fig. 123). Split rib grafts or split calvarial grafts are then
laid over the gaps created over the maxilla and orbital oors.
Additional strips of either rib grafts or calvarial grafts are
placed and either wired or lag screwed to the supraorbital
rims. The pericranial ap is then placed over the superior
orbital rims and the bone grafts in this area.


Figure 122 Artistic reconstruction showing the various osteotomies and bone onlays.



80 Pediatric Neurosurgery

Figure 123 Artistic reconstruction showing the bone onlays from a lateral view.

It is especially important to use extreme care with the
globes and lacrimal system during the orbital dissection. If
not, a stula can occur. If proper care is not taken with the
medial osteotomies, the surgeon may inadvertently extend
into the cribriform plate and thus enter the anterior cranial
fossa. If this happens, there is the possibility of a dural tear
and subsequently a cerebrospinal uid (CSF) leak. It is often
useful at the end of the operation, prior to elevating the skin
ap, to have the anesthesiologist perform an extended Valsalva maneuver while the surgeon looks for a CSF leak. If a
CSF leak occurs in the postoperative period, a lumbar spinal
drain placed for 5 to 7 days usually corrects the problem. If
not, the patient has to be reexplored and the leak sealed.

Closure of the bicoronal incision begins with copious irrigation and hemostasis. Particular attention must be paid
to the skin folds and the gutter space that develop because
debris collects here and offers the most potential for infection. The temporalis muscles are then laid back in their
proper position and secured into place using Vicryl sutures.
These sutures are easily placed through the hole of the microplates or though holes made in the frontal bone. A drain
is placed and attached to light suction for 48 hours. Scalp
closure is accomplished using both galeal layer and the skin
layer. The lower eyelid skin incisions are closed in a single
layer using ne nylon (less than 50), which must be applied meticulously to reduce scarring. The mouth is again
irrigated with Betadine solution, and the transbuccal incision is closed using a single layer of absorbable sutures.




Figure 124 (AC) Postoperative views of the patient 1 year after surgery.

The tarsorrhaphy sutures are then removed and the eyes

irrigated with a balanced salt solution. Lacri-Lube (Allergan,
Irvine, CA) is placed on both eyes and a head dressing is
applied. Excellent results can be obtained using these aforementioned techniques (Figs. 124A124C).

Postoperative Management Including

Possible Complications
Intensive care monitoring for at least 48 hours postoperatively is highly recommended for these patients. Because
of the blood loss that can occur and the risk of epidural
hematoma formation, hemodynamic monitoring is mandatory.
The patient is kept on antibiotics for 3 days postoperatively. The patient takes a clear liquid diet for 3 days and

then advances to a soft diet, rinsing the mouth with saline

between meals throughout this period. After a period of 1
to 2 weeks, the patient usually may resume a regular diet.
There will be signicant postoperative periorbital swelling,
especially by the third postoperative day, but this resolves
gradually. It is often helpful to warn the parents and patient
of this swelling preoperatively.
The bone grafts may undergo resorption with time. If
this happens, the patient may require future contouring. As
mentioned previously, there is evidence that calvarial bone
undergoes less resorption than rib grafts.

Distraction Osteogenesis
Distraction osteogenesis utilizes application of a slow traction force on the osteotomeized bony segments, allowing
new bone formation in the gap. This technique not only

Figure 125 (A) Prole view of Crouzons patient with shallow forehead and retruded infraorbital rims. (B) At the conclusion of monoblock distraction with the internal device. (C) Eight months after the removal of the distraction device. The exorbitism and midface retrusion have improved.

82 Pediatric Neurosurgery
expands the skeleton, but also the surrounding soft tissue,
thereby avoiding the need for bone graft, and the relapse
rate is lower. Distraction osteogenesis is particularly useful for monoblock advancement because there is no immediate creation of dead space. And thus, the infection rate
is decreased. The distraction device can be either internal
or external. In general, the external device offers a better
three-dimensional control of the advancement.
Fig. 125A shows a 5-year-old child who has Crouzons
syndrome who presented with exorbitism. He has both at
forehead and shallow infraorbital rim in addition to retromaxillism. He is an ideal candidate for monoblock advancement. The frontal craniotomy is rst performed, followed
by Le Fort III osteotomies. It is critically important to completely mobilize the Le Fort III segment. Distraction is applied, and the frontal bone ap is xated to the distraction
device at a rate of 1 mm advancement per day (Fig. 125B).
When the supraorbital rim and infraorbital rim reach the
desired position, the distraction process is concluded. The

distraction pins are left for an additional 6 to 8 weeks for

the bone to heal, known as the consolidation phase. The
patient is then returned to the operating room to remove
the distraction device. With the advent of the resorbable
footplates, only the activation pins need to be removed at
the bedside (Fig. 125C).
This chapter is a revision of the chapter, Correction of Exorbitism by Constance M. Barone, M.D., Ravelo Argamaso,
M.D., David F. Jimenez, M.D., and James T. Goodrich, M.D.
The chapter appeared in the Neurosurgical Operative Atlas,
Volume # 2, edited by Setti S. Rengachary and Robert H.
Wilkins. The Neurosurgical Operative Atlas was published by
the American Association of Neurological Surgeons (AANS)
from 1991 to 2000.
We would like to acknowledge and thank Ravelo Argamaso, M.D., for his help and efforts on the original chapter
published in the rst edition of this work.

Depressed Skull Fracture in Infants
Marion L. Walker

Patient Selection
The depressed skull fracture in infants <1 year of age differs from depressed fractures in the older age groups. This
is due to the relative plasticity of the skull, which is not
yet fully ossied. These fractures, referred to as ping-pong
fractures because of their resemblance to an indented pingpong ball (Figs. 131A and 131B), may often be treated
by a conservative approach. Observation may be the only
treatment required. These fractures will frequently remold
within a short time, especially in newborns, making surgical
approaches unnecessary. A surgical procedure is usually not
indicated if remolding occurs, but if elevation of the fracture
is needed the surgery is much less invasive than the typical
depressed skull fracture requires. The surgery is generally
nonemergent and is usually done weeks following the injury, after there has been adequate time to document that
the fracture will not appropriately remold.
It has been reported that, in newborns, the bone will eventually remodel to a normal contour without any surgical intervention. This may indeed occur, but signs of remodeling
of the bone should begin within a few weeks. Prolonged
pressure of the depression on the cortex could potentially
generate a seizure focus, although this possibility is quite
rare. The potential risks of the surgery are infection, intracranial hematoma formation, and blood loss. However, with
good surgical technique the possibility of complication is
very low.

Preoperative Preparation
Anesthetic Considerations
Anesthetic considerations are very important in this young
age group. Current studies suggest that the child should
have nothing by mouth for at least 3 hours after drinking
clear uids and for 4 to 6 hours after ingesting formula or
milk. Good intravenous access is essential and blood should
be available. After the induction of anesthesia, prophylactic
antibiotics are given. Increased intracranial pressure usually
is not a problem. Narcotics should be avoided in this age
group. It is recommended that the arterial oxygen saturation levels be maintained at 90 to 93%. The young infant has

difculty maintaining body temperature due to decreased

insulation and the large amount of surface area of the head.
Heat loss should be a major concern and the room must
be kept warm. When the child is exposed, warming lights
should be used. A heating blanket is always placed on the
operating room table. Warming the anesthetic gases also
helps to prevent hypothermia.

Operative Procedure
The child is positioned on the operating table with a roll
placed behind the back to create a 45 degree angle between
the patient and the table (Fig. 131C). The head is turned to
lie directly lateral. The partial lateral position of the body
relieves any potential stress to the cervical musculature and
ensures that there will be no obstruction of venous return.
The incision is then outlined in a pattern extensive enough
to encompass a scalp incision surrounding the fracture
should this become necessary. When the child is draped, an
area large enough to accommodate a larger incision around
the fracture site should be allowed.
The initial incision should be ~1.5 cm in length along the
suture line nearest to the depressed fracture, most often the
coronal suture, and at least 2 to 3 cm from the midline (Fig.
132A). The periosteum is opened with a Bovie electrocautery. A periosteal elevator is used to free the periosteum and
delineate the bone edge at the suture line (Fig. 132B). The
periosteal elevator is then used to strip the dura from the
bone in a sweeping fashion until the depression of the bone
is reached (Fig. 132C). With consistent gentle upward pressure, the periosteal elevator is used to lever the bone from
a concave to a convex position (Fig. 133). In many cases,
the surgeon will feel a snap as the bone resumes its normal
contour. The bone can splinter, and gentle sweeping motions while continuing upward elevation may be necessary.
Care should be taken not to push the instrument through a
fracture line if one is present. Counter-pressure is obtained
by placing the thumb or index nger of the opposite hand
over the defect as pressure is exerted from below the fracture. This technique allows for greater control.
Once the bone is elevated, the wound and epidural space
are irrigated to ensure hemostasis. The galea is closed with
40 absorbable sutures in a simple interrupted fashion. The
scalp may be closed with sutures or an adhesive substance.


14535_C13.indd 83

4/23/08 3:14:43 PM

84 Pediatric Neurosurgery

Figure 131 (A,B) The depressed skull fracture in an infant resembles an indentation in a ping-pong ball. (C) For operative treatment the child is
placed supine, the shoulder is elevated 45 degrees, and the head is turned to a lateral position.

14535_C13.indd 84

4/23/08 3:14:44 PM

13 Depressed Skull Fracture in Infants

Figure 132 (A) The 1.5 cm incision is made along the suture line
nearest the fracture and a periosteal elevator is inserted. (B) The periosteal elevator is inserted through the suture line into the epidural space


beneath the fracture. (C) The periosteal elevator is used in a sweeping

motion, with upward pressure to elevate the fracture.


14535_C13.indd 85

4/23/08 3:14:45 PM

86 Pediatric Neurosurgery

Figure 133 (A) The periosteal elevator is passed through the suture, (B) is advanced into the epidural space beneath the fracture site, and (C) is
used to lever the depressed bone back into a normal convex position.


14535_C13.indd 86

4/23/08 3:14:48 PM

13 Depressed Skull Fracture in Infants

Postoperative Management Including

Possible Complications
The depressed skull fracture in the infant <1 year of age
can be treated safely and effectively by the surgical method
described if the fracture fails to remold within a reasonable
time. The hospital stay is usually 24 to 48 hours. The results
are excellent, and long-term sequelae may be avoided. The
child is exposed to very little risk while undergoing this

14535_C13.indd 87


This chapter is a revision of the chapter, Depressed Skull
Fracture in Infants by Lyn C. Wright, M.D., and Marion L.
Walker, M.D. The chapter appeared in the Neurosurgical
Operative Atlas, Volume # 2, edited by Setti S. Rengachary
and Robert H. Wilkins. The Neurosurgical Operative Atlas
was published by the American Association of Neurological
Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank Dr. Lyn Carey,
M.D., for her help and efforts on the original chapter published in the rst edition of this work.

4/23/08 3:14:49 PM

Orbital Hypertelorism and Orbital Dystopia
Constance M. Barone, David F. Jimenez, and James Tait Goodrich

Patient Selection

Preoperative Evaluation

Orbital hypertelorism is dened as true lateral displacement

of the orbits. Clinically, this is seen as an exaggerated distance between the orbits. Normal measurements between
the anterior lacrimal crests in Caucasian adults range from
18.5 to 29.9 mm for females and from 19.5 to 30.7 mm for
males. Orbital dystopia refers to the position of the orbits
aligned on the vertical axis. The 4-year-old girl presented
here had craniofrontonasal dysplasia with both orbital dystopia and orbital hypertelorism. Her interorbital distance
measured 34 mm preoperatively. Normal interorbital distance for a 4-year-old girl ranges from 13 to 21 mm, with
the 50th percentile being 17 mm.

At the Monteore Medical Center, all patients with craniofacial disorders are evaluated at the Center for Craniofacial
Disorders. The patient is evaluated by individuals from the
disciplines of pediatrics, pediatric neurology, pediatric neurosurgery, genetics, plastic surgery, ophthalmology, dentistry,
dentofacial surgery, otolaryngology, and speech therapy. All
patients have plain x-ray lms of the face and skull as well as
computed tomography (CT) scanning with bone windows in
axial and coronal views. We routinely use three-dimensional
CT scan reconstruction with a software program that allows
us to perform measurements and surgical osteotomies on
the reconstructed model preoperatively (Fig. 141).

Figure 141 Three-dimensional computed tomography reconstructions for preoperative planning. (A) The preoperative reconstruction
shows the bony detail of the dystopia and orbital hypertelorism; from
this image the amount of orbital movement required can be calculated.


(B) The orbital units are outlined. (C) The orbital repositioning necessary has been calculated and the bone units repositioned; the numbers
generated by the computer are very helpful in determining the requirements for adequate positioning.

14 Orbital Hypertelorism and Orbital Dystopia


procedure. For eye protection, all patients have bilateral

tarsorrhaphies performed or corneal shields placed prior to
formal draping. The patient is intubated orotracheally.

Operative Procedure
The patient is placed in the supine position with the head
on a cerebellar headrest. Rigid xation devices such as a
Mayeld head clamp are not used because the surgeon will
need to move the head (never more than 10 to 15 degrees,
however). Anesthesia equipment is on the left side of the
patient and the nurses Mayo stand is placed over the midabdomen. Overhead tables are not used because they restrict frontal access and visualization of the patients head.


Figure 142 A child with craniofrontonasal dysplasia manifested by

orbital hypertelorism and orbital dystopia.

In the patient presented here, the interorbital distance

between anterior lacrimal crests was 34 mm. The right supraorbital rim was 8 mm lower than the left supraorbital
rim, representing an orbital dystopia (Fig. 142). Multiple
cranial defects were noted from a prior cranial operation
at another institution. In addition, the nose was short, with
the right nasal ala being higher than the left nasal ala. Such
clinical ndings are typical of those seen in children with
craniofrontonasal dysplasia and require a complex craniofacial reconstruction.

Preoperative Preparation
We routinely start antibiotic therapy (to cover for both skin
and nasal ora) at the time of anesthesia induction in the
operating room. Because all surgical manipulations are extradural, anticonvulsant and steroid therapies are not used
routinely. With an extensive reconstruction in a 4-year-old
child, two units of packed red blood cells are made available
(preferably donor-directed). If available, a cell saver unit can
rescue up to 50% of the patients lost blood volume. Because
of the risk of extensive blood loss, all patients require at
least two large-bore intravenous lines of 16 gauge or larger.
An arterial line is placed during induction to permit serial
blood gas and hematocrit measurements, etc., during the

Because a bicoronal incision as well as bilateral infraorbital

incisions will be used, the head and the entire face down to
the nose must be prepared. The nares are cleansed with cotton swabs soaked in Betadine (providone-iodine) solution.
The head is shaved only along the proposed bicoronal incision for a width of ~1.5 cm. The face and head are cleaned
with Betadine scrub, cleansed with alcohol, and painted
with Betadine solution. In an adult, the bicoronal incision
site is then injected with 0.5% lidocaine and a 1:200,000
epinephrine solution. In a child (<7 years old) we use 0.25%
lidocaine and a 1:400,000 epinephrine solution. Sterile towels are placed below the posterior portion of the bicoronal
incision and are also stapled to the posterior aspect of the
scalp. A facial drape is then placed just below the nose and
over the upper lip and endotracheal tube. It is important
that the drapes are loose so that the head can be moved
during the procedure.
We routinely run all our suction lines, cautery cords, etc.,
toward the foot of the patient. Because our surgical team
routinely sits, this allows easy mobility of the chairs (i.e., the
chairs are not rolling over the cords and tubes).
Because the operative site is often irrigated during the
procedure, it is important to have waterproof outer drapes.
Some of the newer drape designs have large plastic bags for
uid collection: We have found these to be quite useful.

Skin Incisions
To gain access to the frontal bone, orbits, and nasal region,
a bicoronal scalp incision is used. This incision is carried
from tragus to tragus, well behind the hairline, both for a
cosmetic closure and to allow for a large pericranial ap that
can be used in the subsequent repair.

Flap Elevation
A full-thickness scalp ap is elevated following the standard
subgaleal plane, leaving the pericranium intact. The pericra-

90 Pediatric Neurosurgery
nium is then elevated as a second separate layer. The aps
are carried forward to the orbital rims, to the level of the
supraorbital nerve and artery on each side. These are frequently encased in a small notch of bone in the supraorbital
rim. This notch can be opened easily using a small osteotome or small Kerrison rongeur. An inverted V-shaped skin
incision is made on top of the glabelar region (Fig. 143) to
allow proper nasal lengthening as well as resection of the
excess skin after the orbits have been moved medially. If
there is no asymmetry in nasal length, a midline incision is
chosen. Particular care must be taken in elevating the scalp
ap if the patient has had a prior operative cranial procedure
(as did the patient presented here) because multiple bony
defects may be present that cause the dura to be adherent
to the overlying pericranium and galea. The scalp ap dissection must expose the belly of the temporalis muscle and
the zygomatic arch bilaterally. In the central facial region,
the nasal bones as well as the medial orbital wall should be
exposed fully. The temporalis muscle is elevated as a unit
using a monopolar needle tip electrocautery. To gain access
to the infraorbital region as well as the maxilla, a lower
eyelid incision is made between the cheek skin and the periorbital skin. This incision is made within a natural skin fold
and usually is located directly above the infraorbital rim. As
an alternative, a subciliary incision can be performed and
the skin elevated as a skin-muscle ap. However, this sometimes leads to an ectropion and/or to subciliary inversion, so
we are not as likely to use it.

Craniofacial Osteotomies
A frontal craniotomy is performed ~2.5 cm above the superior orbital rims (Fig. 144). The use of a high-speed
drill is preferred (e.g., Midas Rex drill with a 135 footplate
[Medtronic Inc., Fort Worth, Texas]), which allows speedy
bone removal in a single step and thus decreases blood loss.
The frontal lobes are then gently retracted extradurally, thus
exposing the orbital roof and anterior cranial fossa on each
side. Filaments from the olfactory bulbs are cut, and the dura
is oversewn if cerebrospinal uid (CSF) leakage is noted.
Subperiosteal elevation of the scalp ap continues over the
orbital rims into the orbits to a distance just past the equators of the globes. A horizontal osteotomy is performed 1.5
cm above the superior orbital rims (Figs. 144 and 145).
This leaves a horizontal bar or bandeau of bone attached to
the cranium. A block of bone below the bandeau, consisting
of the crista galli, the nasal bones, and the ethmoid bones,
is removed en bloc using the Midas Rex drill with a C-1 attachment and an osteotome.
The ethmoid sinuses are then exenterated carefully using
pituitary rongeurs. Care must be taken to remove all mucosa so that a mucocele does not develop in the future. The
cartilaginous nasal septum is identied between the septal
mucosa and is dissected free from the mucosa. A septal resection is performed, with removal of all of the septal cartilage between the medial orbital walls. This creates room
for the medial translocation of the orbitonasal complex and
maintains an open airway. Tears in the nasal mucosa are
repaired with chromic catgut sutures. A segment of the removed nasal bone is then placed in the midline opening at

Figure 143 An intraoperative view showing the V-Y correction to

shorten the nasal bridge and correct the excessive nasal skin from a
medial movement of orbits.

the anterior cranial base, thus separating the cranial cavity

from the nasal cavity.
The periorbita is completely freed circumferentially
around each globe. Using an osteotome, the lateral osteotomies are performed anterior to the zygoma, as approached
from the bicoronal incision. This osteotomy is continued via
the lower lid incision. In a 4-year-old girl, it is important to
keep the osteotomy line above the infraorbital foramen on
each side to avoid destroying future tooth buds. The lower
osteotomies are continued until they meet the medial osteotomies, which were performed previously (Figs. 144
and 145). Because the right superior orbital rim was 8 mm
lower than the left in our patient, an 8 mm segment of bone
is now removed from the superior aspect of the right orbit
to allow a superiorly directed repositioning of the right eye
(Fig. 144).
The osteotomy within the bony orbit is done circumferentially for a 360 degree cut, ~2 cm in from the orbital rim (Fig.
145). An assistant holds a malleable retractor to protect
and retract the globe. An osteotome is placed in the inferior
lid incision to perform the inferior and lower medial and
lateral wall osteotomies. The orbital roof and upper medial
and lateral orbital walls are divided through an intracranial
approach with an assistant placing gentle retraction on the

14 Orbital Hypertelorism and Orbital Dystopia


Figure 144 Artistic reconstruction showing the osteotomy lines and resection of the midline nasal complex and sinuses.

frontal lobe and another protecting and retracting the globe

with a malleable retractor (Fig. 145). A completed osteotomy is then conrmed using a 2 mm osteotome.
Medial canthopexies are performed by rst identifying the
medial canthal ligaments, passing a 26 gauge wire through
them, and then repositioning them through a new hole that
is drilled posterior and superior to the lacrimal crest (Fig.
146). The wires are passed through these holes and then
anteriorly over the medial orbital wall. The wires are tied
together, bringing the orbits toward the midline. In addition,
the right orbit is repositioned superiorly by 8 mm. These
bone units are secured to the bandeau using microplates.
Using Joseph scissors, the overlying nasal skin is separated from the underlying upper and lower lateral nasal
cartilages, creating a pocket for a cantilever bone graft. The

cantilever bone graft consists of a full-thickness piece of

calvarial bone, 8 mm wide and 5 cm long. The bone graft
is then placed in the created nasal pocket and is secured
in place using a microplate to the bandeau. This cantilever
technique stabilizes the bone graft to improve the previously at appearance of the nose. The distance between
the medial orbital walls now measures 15 mm. We usually
overcorrect the medial approximation of the orbital walls
because the orbits tend to drift laterally with time.
The pericranial ap is then reected between the nasal
and cranial cavities. The frontal bone plate is replaced and
secured into position using microplates. The temporalis
muscles are then rotated anteriorly and sutured in place using Vicryl sutures placed through holes in the microplates or
through drill holes made in the bandeau and frontal bone.

92 Pediatric Neurosurgery

Figure 145 Artistic reconstruction outlining
the osteotomy lines along the orbital walls.

Prior to closure it is extremely important to check for any CSF
leaks. Undetected leaks can hinder postoperative recovery.
The critical area for examination is around the cribriform
plate region. Any dural tears during the craniotomy have to
be repaired. We often have the anesthesiologist perform an
extended Valsalva maneuver to look for a CSF leak.
Closure of the bicoronal scalp incision is performed after copious irrigation with saline solution and meticulous
hemostasis. Debris commonly collects beneath the scalp
ap. These areas must be irrigated out thoroughly because

they are a potential source of infection. A drain is attached

to light suction for 48 hours. Scalp closure is a standard
two-layer closure with approximation of the galeal and skin
The lower eyelid incisions are closed in a single layer using a 50 or 60 nylon suture. The excess nasal skin is then
removed along the prior inverted V incision (after a V-Y
lengthening of the right side of the nose in the present case)
(Fig. 143). This incision is closed in two layers. The tarsorrhaphy sutures are removed. The pupils are checked and
the eyes irrigated with a balanced salt solution. Lacri-Lube
(Allergan, Irvine, CA) is then placed in each eye.

14 Orbital Hypertelorism and Orbital Dystopia

Figure 146 Artistic reconstruction showing the bone units repositioned. Bone grafts have been placed laterally, to the side of each orbit.
Because the right orbit had to be moved superiorly, a graft was placed in


the defect inferior to the orbit. The medial canthal ligaments have been
reattached in this schematic.


Postoperative Management Including

Possible Complications
The patient is continued on antibiotics (selected to cover
both skin and nasal ora) far 5 days postoperatively because
the nasal cavity was entered, which increases the risk of
infection. There is always signicant periorbital swelling
postoperatively as well as strabismus. The patient and family should be forewarned of this in the preoperative counseling sessions. In most patients the strabismus corrects to
its original pattern. Any surgical correction of a postopera-

tive strabismus should be delayed for at least 6 months, and

preferably for 1 year.
CSF leakage, meningitis, mucocele formation, bone graft
loss, epiphora from duct obstruction, visual loss, and death
have all been reported with these types of complex craniofacial repairs. With proper attention to detail, such complications can be reduced to the very minimum. As mentioned
above, dural tears must be repaired meticulously during
the operation. The use of the pericranial ap to separate
the nasal and cranial contents is also important in reducing
the incidence of infection. These patients must be followed
for the delayed development of mucoceles and meningitis.

94 Pediatric Neurosurgery

Figure 147 (A) Frontal view of a child with hypertelorism prior to surgery; (B) 8 months after surgery.


Patients undergoing canthopexies have the potential to develop lateral drifting of the medial canthopexies and may
require later secondary revisions.

Clinical Case
This 5-year-old child presents for orbital hypertelorism correction (Fig. 147A). Several alternative techniques are presented in this patients surgery. She underwent correction of
bid nasal tip previously with open rhinoplasty and unication of the domes of the lower lateral cartilages. Occasionally, silicone tubes may be inserted in the lacrimal punctae
at the time of surgery if osteotomies were to interrupt the

nasolacrimal drainage. No midline forehead-glabelar incision was planned, as the brow position can be modied later
with medial microplug hair grafting and lateral trimming.
A zig-zag or wavy coronal incision was made in the scalp,
followed by subperiosteal dissection to the superior orbital
rim. Laterally, the supercial layer of the deep temporal fascia was entered at the level of the frontozygomatic suture
to reach the zygomatic arch, to avoid injury to the frontal
branch. Anteriorly, subperiosteal dissection was continued
centrally caudal to the nasal bones. Care was taken to preserve the medial canthal attachments. Frontal craniotomies
were performed without leaving an intact frontal bandeau,
which facilitates osteomies in the anterior cranial fossa.
The olfactory nerve is preserved whenever possible. The

Figure 148 (A) Three-dimensional computed tomography scan before surgery; (B) 8 months after surgery.

14 Orbital Hypertelorism and Orbital Dystopia

orbital osteotomies were performed as described except
the osteomy at the infraorbital rim was performed from
the coronal incision; therefore, a lower lid incision was not
necessary. After medial translocation and xation of the
bony orbits with resorbable plates, a cantilever bone graft
was placed on the nasal bridge. The integrity of the medial
canthal tendon is checked, and if it is interrupted, then reattachment is performed. A pericranial ap was placed in the
oor of the anterior cranial fossa to isolate the nasal cavity.
Finally, at the conclusion of the procedure, paranasal bolster
dressing was applied using trasnasal sutures. The bolster
consists of inner Xeroform (Kendall Company, Manseld,
Massachusetts) and outer rigid support, such as aluminum
plates. We believe that this step is important to reestablish
the unique topography of the naso-orbital valley. Photo-


graphs (Fig. 147) and CT scans (Fig. 148) demonstrate

pre- and postoperative ndings.
This chapter is a revision of the chapter, Correction of Orbital Hypertelorism and Orbital Dystopia by Constance M.
Barone, M.D., Ravelo Argamaso, M.D., David F. Jimenez, M.D.,
and James T. Goodrich, M.D. The chapter appeared in the
Neurosurgical Operative Atlas, Volume # 2, edited by Setti S.
Rengachary and Robert H. Wilkins. The Neurosurgical Operative Atlas was published by the American Association of
Neurological Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank Ravelo Argamaso, M.D., for his help and efforts on the original chapter
published in the rst edition of this work.

Closure of the Myelomeningocele
David G. McLone

Repair of the myelomeningocele in a newborn is relatively

straightforward and based upon an understanding of the
developmental anatomy of neurulation and what has gone
wrong. Preservation of neurological function, detection of
associated anomalies, and prevention of postoperative
complications are the essential aims of this procedure.
Pre-operative concerns include preparation of the infant
for surgery and informing the family of the risks and advantages of surgery as well as what the future likely holds
for the child.

Patient Selection
Fortunately, prenatal ultrasound and serum AFP have allowed time for a decision about the future of the pregnancy
and, if the pregnancy is going to term, time for the family
to prepare for the birth of a child with a neural tube defect. When the birth of a child with a neural tube defect
is a surprise, it is important for the parents to have some
time to grapple with the many new issues facing them. It is
equally important to proceed with denitive treatment in a
timely fashion. Fortunately, most parents today are aware
that the child will be born with spina bida and have had
prior counseling. Unless the newborn is critically ill, repair
of the myelomeningocele should proceed. Signicant delays
increase both morbidity and mortality.
Early closure of the myelomeningocele remains an important part of initial management. Although studies have not
documented any increase in the decits in the survivors of
briey delayed closure, our experience with patients transferred late and requiring delayed closure in the face of infection has been a decrease in motor function in some and an
increased rate of ventriculitis (37%), as compared with that
in patients with early closure (7%).
The natural history of unrepaired newborns who are fed
but denied antibiotics would indicate that 40 to 60% will
survive, often much more signicantly impaired. If antibiotics are added to the care of the unrepaired children, the
mortality and morbidity fall to levels similar to those of
neonates repaired in the rst 24 hours.
Optimally, we operate on the neonate soon after birth,
preferably in the rst few postnatal days. Prenatal diagnosis makes this increasingly more possible. The opera-


tion may safely be deferred for up to 72 hours without an

increase in complications. This delay is particularly important for the unstable or critically ill newborns. These
babies can usually be stabilized within 72 hours, and this
time is usually well spent. A search for coexistent anomalies of other organ systems should be undertaken during
this time. Severe anomalies or absence of vital organs or
unrepairable cardiac defects may portend a poor outcome.
Renal anomalies are common but not usually life-threatening. Although the neonate may not produce signicant
amounts of urine during the rst 24 hours, the presence of
urine in the bladder implies the presence of functioning
kidneys. Ultrasonography can delineate most major renal
anomalies. Syndromes related to chromosomal anomalies
may not be obvious upon initial inspection but should be
sought out.
Although most coexisting anomalies are not immediately
life-threatening and may be dealt with without much difculty, it is important to remember that a few newborns
with myelomeningocele may have potentially fatal associated malformations and may not be saved. Intervention to
prolong the lives of these infants in the setting of a dismal
outlook makes little sense. They should be kept comfortable,
and their families should be supported.

Preoperative Preparation
The preparation of the neonate with a myelomeningocele
for surgery is usually not difcult. Most have a high hematocrit and an adequate intravascular volume, and uid
resuscitation is therefore usually not necessary. Common
perioperative complications include hypothermia and hypoglycemia, both of which are easily prevented through the
judicious use of heating and monitoring of serum glucose.
The placode may become desiccated with prolonged exposure to the air and should therefore be protected. Covering the placode with sterile, saline-soaked gauze is preferable. The dressing may be covered with plastic wrap to
prevent rapid evaporation of the saline. Substances that are
toxic to tissues and result in inhibition and delay of wound
healing should not be used directly on the malformation.
The use of perioperative antibiotics is left to the discretion
of the surgeon. We have tended to use them.


Closure of the Myelomeningocele

Operative Procedure

Missed Abnormalities

Preservation of Neurological Function

Split Cord (Diastematomyelia)

Preservation of Neural Tissue

Prior to closure, the infant should be kept on his or her
abdomen to reduce mechanical trauma to the neural tissue.
During the subsequent surgical repair, great care must be
taken to avoid drying, traction on neural elements, irrigation with hot saline, and excessive use of electrocautery.
With magnication and microinstrumentation, the opening
of signicant blood vessels can be avoided. Occasionally, the
entire closure can be performed without cautery.
It has clearly been shown that the exposed neural tissue is
functional. Movement of muscles subserved by spinal cord
segments involved in the placode as well as the presence
of somatosensory evoked potentials conducted through
the placode both point to the functional nature of this tissue. Even when the initial examination fails to demonstrate
movement of muscles innervated by the placode, the placode should still be considered functional because more
than one third of these patients subsequently gain motor
functions not previously detected. Therefore, all neural tissue must be preserved.

Preservation of Vascular Supply

Preservation of the vascular supply to the placode is essential if this tissue is to survive. Unlike the normal spinal cord,
the blood supply to the placode does not enter exclusively
through the vertebral foramina along the nerve roots. Many
large vessels pass directly through the laterally reected
dura mater and supply the myelomeningocele. Those supplying the junction between the neurulated spinal cord and
the placode seem to be at greatest risk. Rarely is it possible to preserve all of these vessels, and, fortunately, they
can sometimes be sacriced if necessary without apparent
injury to the placode. Nonetheless, great care must be exercised to preserve these vessels while mobilizing the dura
for closure (Fig. 151A).

Inclusion Dermoid
Great care should be exercised in separating the edge of the
placode from the contiguous cutaneous epithelium (Figs.
152A and 152B). Some pearls of epidermoid tumors may
already reside within the placode. Retained fragments, possibly even a single cell, could, if imbricated within the closure, produce an inclusion epidermoid tumor (Figs. 153A
and 153B). These inclusion dermoids produce not only tumors but associated desquamation debris, which may also
stimulate an intense arachnoiditis. Later in the childs life, a
tethered cord release in the face of the scar produced by this
inammatory process can be extremely difcult.


Both the rostral and caudal ends of the closure site should
be closely inspected prior to the closure of the placode to
identify associated tethering, bony spurs, or brous bands.
Cranially, removal of an additional lamina may be necessary
to adequately visualize the adjacent spinal cord (Figs. 153C
and 154). Hemimyelomeningoceles may also be readily
visualized by examining the adjacent spinal cord. The presence of an asymmetrical neurological decit preoperatively
should alert the surgeon to the possibility of a hemimyelomeningocele or an associated split cord malformation.

Thickened Filum Terminale

Caudal to the placode a thickened lum terminale may often
be present (Figs. 151B, 155A, and 156A). This should be
sectioned if present. Spinal cord tethering in these patients
may result as much from a missed thickened lum as from
adhesions to the placode.

Anatomical Reconstruction
The different types of myelomeningocele are best understood in terms of an archetypal anatomical deformity and
variation about that archetype. The basic deformity consists of an open neural placode, which represents the embryologic form of the caudal end of the spinal cord prior to
neurulation (Fig. 155A155C). A narrow groove passes
down the placode in the midline. This represents the primitive ventral sulcus, and it is directly continuous with the
central canal of the closed spinal cord above (and occasionally below) the neural placode. Cerebrospinal uid passes
down the central canal of the spinal cord and discharges
from a small pit at the upper end of the placode to bathe
the external surface of the exposed neural tissue. This uid
does not indicate rupture of subarachnoid space ventral to
the myelomeningocele.
The size of the sac on the babys back at the time of birth is
dependent upon the amount of spinal uid that is collected
ventral to the neural placode. The majority of lesions will be
ush with the babys back. A smaller number of placodes are
raised far above the surface of the back by marked expansion of the subarachnoid space. Generally, however, both
types are grouped under the myelomeningocele heading.
In most cases, the spinal cord rostral to the neural placode
is normal in gross form. Anomalies such as split cord and
absence of a segment of the spinal cord can exist above the
neural placode, however. Concurrent arteriovenous malformations and lipomas of the spinal cord are also possible.
Occasionally, the neural placode is in a totally disorganized
state. In these cases, the neural placode appears to have
undergone intrauterine infarction so that portions of it are
severely dysplastic and reduced to a simple membrane. This
would support the concept that myelomeningocele is indeed a progressive intrauterine disease.

98 Pediatric Neurosurgery

Figure 151 (A) Vessels entering the placode (arrows) are preserved
during reconstruction of the neural tube. (B) A drawing shows the free
edges of the dura being held open as the neural tube is reconstructed.
The arrow indicates the point where the thickened lum was cut. (C) A

photograph showing the beginning of reconstruction of the neural tube

at the rostral portion of the myelomeningocele. (D) The epidural space
lateral to the neural tissue is opened; it is important not to carry this
dissection ventral to the neural tissue.


Closure of the Myelomeningocele


Figure 152 (A) A drawing shows the incision being made at the junction of normal and abnormal thin skin. (B) A photograph at surgery shows
the abnormal thin skin being cut free at the junction (arrows) of the abnormal skin and the placode.


Figure 153 (A,B) Intraoperative photographs show inclusion dermoid tumors with marked arachnoiditis. (C) Photograph shows a split cord malformation (arrows) proximal to
the myelomeningocele.

100 Pediatric Neurosurgery

Figure 154 A contrast computed tomography scan shows a hemimyelomeningocele, a split cord malformation; the hemi cord on the
left had a myelomeningocele.

Figure 155 (A) A terminal myelomeningocele; the arrow indicates

the area of a thickened lum terminale. (B) A larger myelomeningocele shows the ventral sulcus in the midline and the arrow indicates
the entrance to the central canal of the adjacent normal cord. (C) An
infant with a large thoracolumbar myelomeningocele and a kyphotic
deformity at the junction of the thoracic and lumbar spines.

The functional motor and sensory levels are related to

but not always consistent with the anatomical level of the
lesion. Often, function is preserved below the anatomical
segments involved. Again, functional asymmetry occurs and
should raise the question of an additional lesion such as a
split cord.
The normal anatomical structures derived from the neural
tube are almost always present but are open in the midsagittal plane (unneurulated), and the dorsal roots are therefore displaced laterally. Because the neural crest is usually
involved in the defective neurulation, the dorsal roots are
often attenuated or absent. The exposed neural surface is
the ependymal surface of the neural placode and is continuous with the central canal of the spinal cord rostrally.
The lateral edges of the neural tissue are developmentally
the alar (sensory) plate with the dorsal root entry zones at
the lateral edge. The medial portion of the placode is basal
(motor) plate and contains the anterior motor horns. Ventral
to the placode along either side of the midline the motor
roots exist from the placode. The sensory roots enter the
cord at the periphery of the placode lateral to the motor
roots. The dorsal root ganglions are usually reduced or absent,
so the dorsal roots are small or absent. The ventral surface
of the placode is covered with pia-arachnoid, which is directly contiguous with the arachnoid membrane of the sac.
The sac usually encloses an intact subarachnoid space. An
understanding of this anatomy is essential to reconstituting
the spinal cord and its coverings.
Dissection of the myelomeningocele begins at the junction of the abnormal covering epithelium and the normal
skin near the rostral end of the placode (Figs. 152A and
152B). This junction should be incised around the entire
circumference of the myelomeningocele. Once this has
been completed, the dissection is carried toward the neural
placode. Dividing the epithelial junction from the neural
tissue requires care because on the one hand, this is the
region where dorsal roots and segmental vasculature enter
the neural placode, and on the other hand, any residual skin
elements may grow to become inclusive epidermoid tumors
(Figs. 153A and 153B). The use of magnication enables
one to dissect free any nerve roots that are adherent to surrounding tissues. When this has been completed, the neural
tissue may oat freely on an arachnoid enclosed sac of cerebrospinal uid. Once the neural tissue is freed, every attempt should be made during anatomical reconstruction of
the spinal cord to prevent later retethering of the placode.
Although pial-to-pial closure of the placode into a tubular
structure has not completely prevented retethering, it may
reduce the incidence of this complication (Figs. 151B and
151C). More importantly, it makes untethering of the spinal cord later considerably easier to perform. The reapproximated neural tube is usually adherent only along the dorsal
closure line (Figs. 157A157D). In contrast, leaving the
placode open allows the unclosed neural tissue to become
densely adherent over the entire exposed ependymal area
of the placode; the laterally displaced dorsal roots are usually caught in the scar and require tedious dissection to free
them (Figs. 157E157G).
The central canal is reconstructed throughout its entire
length so that the neural placode becomes a tube. Closing


Closure of the Myelomeningocele 101

Figure 156 (A) A drawing shows the suturing of the dura. We now
prefer a running locked suture rather than the interrupted suture. (B)
Dissection in the plane between the subcutaneous fat and the muscle


fascia allows the skin to be mobilized to aid closure. (C) A drawing shows
closure of the skin.

102 Pediatric Neurosurgery

Figure 157 (AD) A contrast computed tomography (CT) myelogram

shows a reconstructed neural tube that has retethered along the suture
line to the overlying dural closure. (FG) Contrast CT myelograms show10.1055/978-1-60406-039-3c015

the neural placode into a neural tube and folding the arachnoid sac around the tube encloses the cord within an envelope of cerebrospinal uid. By suspending the closed neural
tube in an intact cerebrospinal uid compartment, we hope
to decrease the possibility of scarring and adherent neural
elements that might later result in tethering of the spinal
cord as the child grows. Magnetic resonance imaging has
become the imaging modality of choice, and a postoperative
study can demonstrate the reconstructed neural tube (Figs.
158A158D), but not as clearly as a contrast computed
tomography myelogram.
The open edges of the dura mater attach to the underside
of the skin lateral to open skin edge. To ensure adequate
dura mater for closure, the most lateral extent of the dura
must be found and detached at that point (Fig. 151B). No
dissection in the epidural space ventral to the neural tissue should be attempted (Fig. 151D). The dura is usually
very thin under the spinal cord, and if torn is difcult to
repair. During dural closure, the neural tissue may become
included in the suture. Therefore, care should be taken to
avoid this preventable complication.
Once the dura mater is free, it is closed in the midline. This
layer should be closed watertight if possible. We prefer
a running locked nonabsorbable suture of 5 or 70 (Fig.
156A). The dural closure must not constrict the underlying neural elements or interfere with the blood supply to
the reconstructed cord. Potential recovery may be lost to
ischemia or infarction if dural or fascial coverings constrict
the underlying tissues.
Mobilization and midline approximation of lateral para
spinal muscle fascia are optional and not essential (Fig. 15
6B). It may not be easy to obtain signicant lateral tissues.
Muscle closure at the lumbosacral level is often difcult

ing neural placodes adherent to the overlying dural closure. (H) A contrast CT myelogram showing a free-oating reconstructed neural tube.

because the fascia of sacrum and ileum are densely adherent to the bones.
Thoracic and upper lumbar myelomeningoceles can be
difcult to repair if associated with a kyphotic deformity
(Fig. 155C). To allow skin closure without compression
may require a kyphectomy. It has the benets of making the
closure easier, giving the patient a at back, and converting
muscles from exors of the spine to extensors, which prevents progression of the deformity.
Closure of the skin should be performed in the midsagittal
plane whenever possible. Future untetherings or orthopedic
procedures will be facilitated by a simple midline closure
(Fig. 156C). Mobilization of the skin should also include
the subcutaneous fat layer because the vascular supply to
the skin comes through this layer. Blunt dissection in the
plane between the muscle and subcutaneous fat is the best
method to preserve the blood supply (Fig. 156B). Some
consideration of cosmesis should be given here, but this
is not a major consideration if it poses any added stress to
neural tissue.

The timing of shunt placement is a matter of some debate.
Approximately 20 to 30% of patients with myelomeningoceles do not need a shunt, and therefore we have advocated
delaying a shunt procedure until well after the initial closure. In the presence of obvious severe hydrocephalus at
birth, however, it would seem to make little sense to delay
and subject the patient to a second anesthetic. Placement of
the shunt at the time of initial closure in these cases is safe
and reduces the risk of cerebrospinal uid leakage or wound
breakdown postoperatively.


Closure of the Myelomeningocele 103

Figure 158 (AD) Four neonates with myelomeningoceles. Postoperative magnetic resonance imaging axial T2 scans show the reconstructed
neural tubes.

Postoperative Management
Management of the Closure Site
A variety of techniques have been employed to protect the
closure site postoperatively, including placing the patient
prone or suspending the patient from a sling. These maneuvers are of little value. We simply place the patient in a
bassinet postoperatively and allow him or her to be held
in the mothers arms without restrictions. We have not encountered any signicant problems using this regimen.

Care of the Patient

At present, the patient is given intravenous uids with
10% dextrose for the rst 24 hours and then is given to the

mother for feeding. Daily inspection of the closure is recommended for signs of infection, separation of the skin edges,
or leakage of cerebrospinal uid.
During the hospital stay, instruction is given to the parents to prepare them for caring for the baby in the home.
This is the ideal time for the parents to become familiar with
the team that will assist them in the care of the baby as an
outpatient. It is essential that the parents gain condence
in their own ability to care for the baby and are aware that
the team is always available for support.

Dandy-Walker Malformation
Arthur E. Marlin and Sarah J. Gaskill

The Dandy-Walker malformation is a congenital anomaly

of the posterior fossa composing 7% of cases of congenital
hydrocephalus. This entity (Fig. 161) was rst described in
1911, although Dandys original case of a 13-month-old with
occlusions of the foramina of Magendie and Luschka was
not published until 1914. The term Dandy-Walker syndrome
(DWS) was introduced by Banda in 1954. The malformation consists of: (1) defective development of the cerebellar vermis; (2) cystic enlargement of the fourth ventricle;
(3) hydrocephalus; (4) enlarged posterior fossa; (5) elevated
transverse sinuses; and (6) obstruction of the outlets of the
foramina of Magendie and Luschka (not seen in all cases). The
pathogenesis is unknown. Some cases are suggestive that the
enlarged fourth ventricle causes secondary aqueductal stenosis and dilatation of the lateral and third ventricles.

Treatment via Shunt Placement

Although attempts at early re-establishment of the patency
of the outlets of the fourth ventricle have been made, the
mainstay of treatment is cerebrospinal uid (CSF) diversion. Three options exist for ventricular shunt placement:
(1) fourth ventricle; (2) lateral ventricle; and (3) both. The
distal catheter is usually placed in the peritoneal cavity. If
there are contraindications to peritoneal placement, such
as peritonitis or multiple adhesions, the right atrium can be
used. The correct type of diversion has been a controversial issue, with advocates and opponents for each option. A
fourth ventricular shunt, especially early in the course of
the disorder, may be all that is necessary, according to its
advocates. However, if secondary aqueductal stenosis has
occurred or there is not free communication between the
supratentorial and infratentorial ventricular systems, this
will not decompress the lateral ventricles and downward
herniation can occur. The same type of argument exists for a
lateral ventricular shunt; the posterior fossa cyst may not be
adequately decompressed and upward herniation can occur.
Preoperatively, ventricular communication can be assessed
by radionuclide ventriculography or computed tomography
(CT) positive contrast ventriculography.
Recently there have been case reports of endoscopic third
ventriculostomy with shunting of the fourth ventricle and
even endoscopic third ventriculostomy and placement of an
aqueductal stent.

The authors prefer a fourth and lateral ventriculoperitoneal

shunt in DWS with panventricular enlargement. Two ventricular (lateral and fourth ventricle) catheters are connected
proximally to a valve with a three-way connector and then to
the peritoneal catheter (Fig. 162). This allows equalization
of pressures between the intracranial compartments.

Selecting Shunt Equipment

An extensive variety of shunt systems is available, suggesting that no ideal system exists. The authors prefer straight
ventricular catheters, a proximal valve with a reservoir (to
allow access for CSF sampling), and a peritoneal catheter
with a distal open end and three series of side slits just
proximal to the ending (offering no resistance). The selection of valve pressure is another area of controversy. In
general, a low-pressure valve is appropriate for infants. Higher
pressure valves should be used when the calvarium is closed.
Variable resistance valves can also be used.

Patient Selection
Patients usually present at birth or shortly thereafter. Not
infrequently the diagnosis can be made prenatally by ultrasonography. Symptoms are referable to increased intracranial pressure. Macrocephaly at birth or, later, a rapidly
enlarging head with a circumference crossing growth percentile lines, irritability, and vomiting are the usual modes
of presentation. Examination reveals macrocephaly with a
bulging fontanelle, suture diastasis, and, often, engorged
scalp veins. The posterior fossa is commonly large and occipital transillumination is positive. The diagnosis can be
made by ultrasonography, CT, or magnetic resonance imaging (MRI) (Fig. 163). CT and MRI are denitive.

Preoperative Preparation
The issue of prophylactic antibiotics is unresolved for lack
of a large, properly designed series. The authors use prophylactic anticoagulase-negative Staphylococcus agents based
on recent hospital sensitivities. These are given just prior to
the surgical incision. Vancomycin is not used in situations
of increased intracranial pressure because of its potential


14535_C16.indd 104

4/11/08 11:36:42 AM

16 Dandy-Walker Malformation 105

Figure 161 An artists depiction of the Dandy-Walker malformation in the midsagittal plane.

histamine release and the possibility of causing further increases in intracranial pressure.
Attention to detail in the operating room is mandatory.
Operating room personnel should be aware that infection is
the most serious complication of any shunting procedure.
Trafc in the operating room should be kept to a minimum
once the instruments are open. Only necessary personnel
should be present in the operating suite in an effort to keep
airborne particles to a minimum. Movement within the operating room should also be kept to a minimum.
If no intravenous access is present, an inhalational agent
is given by mask until this is obtained. Anesthesia is accomplished after thiopental induction and endotracheal intubation with fentanyl and Versed (midazolam), or rectal Brevital (methohexial). A short-acting, nondepolarizing muscle
relaxant such as mivacurium can be used for intubation.

14535_C16.indd 105

Operative Procedure
Patient Positioning
The patient is positioned supine with the head turned laterally (to the left for a right-sided shunt). A roll is placed under the shoulders. The size of the roll depends on the size of
the patient and the head. The roll is placed so that the posterior auricular region, neck, thorax, and abdomen are on
the same plane to ensure an easy pass of the shunt passer.
If the head is very large, the body may need to be elevated
with towels to accomplish this (Fig. 164). The head should
be positioned at the edge of the table close to the surgeon.
The patients bladder is emptied by a Cred maneuver to
minimize the risk of bladder perforation during peritoneal
catheter placement.

4/11/08 11:36:42 AM

106 Pediatric Neurosurgery

Figure 162 Operative shunt positioning.


Surgical Technique
The landmarks for the lateral ventricle are identied. The
cranium should be perforated supratentorially 4 to 6 cm
above the inion and 2.5 to 4 cm lateral to the midline. The
line from the projection of the anterior fontanelle (~1.5 cm
above the eyebrow) is determined to establish the trajectory
of the lateral ventricular catheter. As these landmarks will be
obscured by the surgical drapes, external guidelines in the
room can be used to ensure the proper trajectory. Enough
hair is shaved to allow an adequate border for draping. A
curvilinear incision is marked with an inferior limb of sufcient length to allow for a fourth ventricular catheter (Fig.
164). The skin is to be reected laterally so that the shunt
apparatus is not directly under the incision. A paraumbili-

cal incision is also marked. At this time, the length of the

lateral ventricular catheter should be determined. The tip
should be anterior to the choroid plexus; thus, the external
landmark is anterior to the coronal suture.
Thorough skin preparation is essential. This is accomplished by a 10 minute preparation with providone-iodine
soap and solution followed by Betadine gel. Before draping, landmarks are again noted. Sterile towels are applied
in a fashion to allow minimal skin exposure. Excess gel is
removed with sponges and an Ioban-impregnated barrier
drape is applied to hold the towels in place. Liquid-impenetrable drapes are then used to complete the draping of the
We use the surgical isolation bubble if available. This is
placed over the patient and the operative procedure is per-

Figure 163 (A) Sagittal and (B) axial magnetic resonance imaging views (T1-weighted images) of a patient with a Dandy-Walker malformation.

14535_C16.indd 106

4/11/08 11:36:50 AM

16 Dandy-Walker Malformation 107

Figure 164 Patient positioning showing skin incisions (dashed lines).


formed within the bubble (Fig. 165). All necessary instruments and supplies are also placed in the bubble. The shunt
components (doubly wrapped) are kept in their second wrapper until inside the bubble, where they are opened and left to
soak in a bacitracin solution until placement. If the bubble is
not used, the shunt equipment is not opened until draping is
complete. The three-way connector is attached with a double
30 silk tie to the ushing device (Fig. 166).
The cranial incision is made rst. Hemostasis is obtained
with bipolar coagulation. A subcutaneous pocket is made
with blunt dissection caudal to the area of the bur hole to
house the ushing device. Some medial subcutaneous dissection is also done for the three-way connector and ventricular catheters. Retraction is accomplished with a small

self-retaining retractor that is also hemostatic. The calvarium is perforated supratentorially.

The suboccipital muscles are split and the occipital
squama identied. A twist drill hole is placed and enlarged
with a rongeur so the dura is well visualized. In the newborn
or young infant the dura may have extensive venous lakes.
It can also appear bluish because of the underlying cyst. It is
thus aspirated with a 22 gauge needle to ensure that there
is not a venous lake. If no area can be found without blood
return, a 40 silk purse-string suture is placed (Fig. 167).
When placing the stitch, it is important to penetrate both
leaves of the dura.
The cranial incision is covered with bacitracin-soaked
gauze while the paraumbilical incision is made. The shunt

Figure 165 The surgical isolation bubble.


14535_C16.indd 107

4/11/08 11:36:54 AM

108 Pediatric Neurosurgery

Figure 166 A detailed depiction of the shunt connections.


tube passer is then tunneled from the abdomen to the cranial

incision so the distal tube can be passed from the cranium to
the abdomen. Although this can be done in either direction,
tunneling the passer from the abdomen to the cranium has
less risk of penetrating the pleural space. The distal tubing
is connected to the ushing device. The ushing device is
seated in the subcutaneous pocket without kinks.
The supratentorial dura is coagulated and a small opening is made with a no. 11 blade. The ventricular catheter on
a stylet is passed and CSF is taken for culture, cell count,
and glucose and protein determinations. The proper length
of catheter should be determined preoperatively so that it
will be anterior to the coronal suture and thus anterior to
the choroid plexus. A good ow of CSF should be obtained
and the catheter connected to one limb of the three-way
connector with a double-tied 30 silk suture.
Next, a small hole is made in the dura of the posterior
fossa in the manner described above. The ventricular catheter is passed in ~3 to 4 cm; again, a good ow of CSF should

14535_C16.indd 108

be obtained. The surgeon should be careful to avoid placing

the catheter so far into the cyst that it impinges on the brain
stem. Cerebellar reconstitution is unlikely to occur, so the
catheter need not be placed in a precise location within the
cyst. The catheter here is secured gently to the muscle with
a 30 absorbable suture to prevent kinking. A gentle curve is
made with the tubing subcutaneously and it is brought rostrally to be secured to the remaining limb of the three-way
connector, again with a double-tied 30 silk suture (Figs.
163 and Fig. 166). This region is now covered with a bacitracin-soaked gauze.
Attention is turned to the abdominal incision. The peritoneal catheter should be draining CSF spontaneously. Not infrequently, the distal catheter will need to be aspirated with
a blunt needle to remove an air lock and ensure spontaneous ow. If there are no contraindications such as previous
abdominal surgery, peritonitis, or an enlarged or surgically
augmented bladder, a trocar is placed through the rectus
sheath into the peritoneal cavity, directed medially toward

4/11/08 11:36:55 AM

16 Dandy-Walker Malformation 109

Figure 167 Placement of the ventricular catheter through a dural venous lake.

the umbilicus. Simultaneously with trocar placement, the

anesthesiologist gives the patient a Valsalva maneuver. The
catheter is passed through the trocar. A full length (90 cm)
of peritoneal catheter is placed as there are no associated
complications, and the need for an elective shunt lengthening is eliminated. If the use of the trocar is contraindicated,
then the rectus sheath is split. The peritoneum is identied and a small opening is made; the surgeon takes care to
lift the peritoneum away from the underlying bowel. After
function of the shunt catheter is conrmed by spontaneous ow of CSF, it is placed into the peritoneal cavity under
direct vision. The rectus sheath is closed with absorbable

Postoperative Management Including

Possible Complications
In an infant, the head of the bed is elevated 30 degrees and
feedings are begun as soon as bowel sounds are audible. In
a child with a xed cranium and very large ventricles (i.e., a
small cortical mantle), the head of the bed is kept at for the
rst postoperative night and then is elevated progressively.
In this situation a variable resistance valve or antisiphon
valve such as a Delta valve can be used and the patient mobilized the rst postoperative day.

Skin Closure
Wounds are now copiously irrigated with a bacitracin solution, 100,000 units in 500 ml of normal saline. The subcutaneous tissue of the abdomen and the galea of the scalp
are closed with 30 or 40 absorbable sutures. The skin is
closed with a subcuticular stitch of 40 or 50 absorbable
sutures and Steri-Strips.

14535_C16.indd 109

Volumes can be written on shunt complications, but with

meticulous technique and attention to detail these should
be minimal. Infection rates should not be more than 8% and
may approach 1 to 2%. The treatment of infection involves
sterilization of the CSF and, usually, shunt replacement. The
postoperative x-ray lms will forewarn of obstruction due
to poor shunt positioning.

4/11/08 11:36:55 AM

110 Pediatric Neurosurgery

CSF leakage around the cranial catheter can occur because
of a lack of cortical mantle to seal the dural opening around
the catheter or because of too large an opening. If this is
not eliminated by head elevation, a bur-hole type device
or Rickham reservoir may need to be placed instead of a
straight ventricular catheter. Subdural effusions or hematomas are treated expectantly if asymptomatic.

14535_C16.indd 110

Intra-abdominal injury is rare. Intestinal perforation may

present by shunt infection with gram-negative or multiple organisms. This can usually be treated by removing the
catheter and treating the infection, followed by replacement
of the shunt. Perforation of the bladder can be managed by
catheter removal and Foley drainage of the bladder.

4/11/08 11:36:59 AM

Surgical Management of Chiari I
Malformations and Syringomyelia
Elizabeth C. Tyler-Kabara, Richard B. Morawetz, and W. Jerry Oakes
This chapter outlines our approach to the adult patient with
a Chiari I malformation and/or syringomyelia. As evidenced
by the multiplicity of procedures used in the treatment of
these patients, there is no proven optimal approach. In fact,
there is no proof that a patient with an incidentally identied Chiari I malformation or syrinx is destined to become
neurologically debilitated, because the natural history of
this condition has never been described satisfactorily. It has
become possible only in the era of magnetic resonance imaging (MRI) to elucidate the long-term efcacy of various
surgical procedures in the treatment of Chiari I malformations and syringomyelia. Unfortunately, this information
will not become available to us for another 5 to 15 years,
and the follow-up of these patients can exceed the length
of a single neurosurgeons career.

Patient Selection
Abnormal neurological signs referable to a Chiari I malformation or syrinx, or progressive syrinx enlargement, on serial imaging studies are indications for surgical intervention.
Weakness and dissociated sensory loss are commonly found
in patients with syringomyelia, and bulbar signs often accompany syringobulbia.
However, it is not uncommon to encounter patients with
impressive abnormalities seen on imaging studies but without denite abnormal neurological signs (e.g., in a setting
of pain following a work-related injury, or hemihypesthesia
from head to toe, including the sacrum). In these cases a
conservative approach could be adopted, with follow-up
MRI studies at appropriate intervals. If the syrinx remains
unchanged in size and the patient has not developed progressive neurological abnormalities, no intervention is indicated. Similarly, a period of observation is warranted for
the patient with a Chiari I malformation who presents with
headache alone. If the headache seems related to the Chiari
I malformation, as evidenced by worsening with cough or
Valsalva maneuver, we recommend surgical intervention
but emphasize to the patient that the goal of surgical intervention is not to relieve the headache but to minimize
any chance of neurological deterioration. Cine MRI may be
helpful to identify patients with abnormal cerebrospinal
uid (CSF) ow at the craniocervical junction who are most
likely to benet from decompression.

In contrast, surgical intervention is warranted in patients

with documented progressive syrinx enlargement, even in
the absence of abnormal neurological signs, because these
patients are at risk for development of progressive neurological dysfunction at a later date. At the other end of the
spectrum, the potential benets of surgical intervention in
the wheelchair-bound, debilitated patient with syringomyelia must be weighed against the risk of acute respiratory
failure and long-term ventilator dependence in the perioperative period. There are no proven guidelines to follow in
the selection of surgical candidates in this setting.

Preoperative Preparation
Preoperative assessment of these patients includes a computed tomography (CT) scan or MRI of the head to determine
the presence or absence of hydrocephalus. Hydrocephalus,
if present, should be treated by a shunting procedure prior
to consideration of other surgical interventions. An MRI of
the entire spinal cord should be obtained to assess the presence and/or extent of a syrinx. Electrophysiological studies are not routinely performed, because the results do not
usually inuence surgical decision making. It is important
to emphasize that the goal of the procedure is to halt the
progression of neurological signs and symptoms rather than
to ameliorate the symptoms.

Operative Procedure
Chiari I Malformation
For the patient with a Chiari I malformation without syringomyelia, a suboccipital decompression and upper cervical
laminectomy with a generous duraplasty are performed.
No attempt is made to resect the tonsils, plug the obex, or
shunt the fourth ventricle, because these interventions have
not been demonstrated to improve results beyond those
achieved by bony and dural decompression alone. The patient is positioned prone using a head pin xation device.
The neck is exed and the head of bed is raised. Preoperative
antibiotics are administered as anesthesia is induced.
Fig. 171A illustrates the position of a vertical incision made from the inion to the mid-cervical region. The


112 Pediatric Neurosurgery

Figure 171 (A) Schematic of incision placement and bone work for operative exposure of a Chiari I malformation. (B) Operative exposure of a
Chiari I malformation with the oor of the fourth ventricle exposed.

paraspinous muscles are dissected using a subperiosteal

technique (Fig. 171B). A suboccipital craniectomy and C-1
laminectomy are performed with an emphasis on achieving
foramen magnum decompression rather than on a large
lateral suboccipital craniectomy. The technique during this
portion of the procedure should be quite gentle, as there is
little subarachnoid space under the dura and there is risk
of injury to the brain stem and upper cervical cord. A C-2
laminectomy may be performed to permit the decompression to be carried inferiorly, below the tonsils. Excessive
laminectomies should be avoided to prevent the increased
risk of postlaminectomy kyphosis. The dura is then opened
in a vertical direction. The patients own pericranium may
be used as a graft. Some surgeons prefer cadaveric or articial dural substitutes. An elliptical graft is fashioned and
sutured in place to provide a generous decompression of
the cerebellar tonsils. The wound is closed in layers. We do
not routinely image the patient in the immediate postoperative period unless there is an indication for the study.
An occasional pseudomeningocele will develop but usually
resolves. For the symptomatic patient with a Chiari I malformation and an associated syrinx, we perform the same

For the patient with a large symptomatic syrinx that fails
to respond to adequate craniocervical decompression, and
for the management of posttraumatic syringomyelia, a syringoperitoneal shunting operation may be performed. Entrance into the syrinx in the midthoracic region is preferable
so that any sensory decit is localized to the chest wall. The

goal is to eliminate the pressure gradient across the wall of

the syrinx and completely decompress it.
The patient is placed in the lateral position with the left
side down to allow access to the right upper quadrant of the
peritoneal cavity (Fig. 172). A two-level laminectomy over
the largest portion of the syrinx is performed. The dura is
opened in a vertical fashion off the midline toward the side
of the dorsal root entry zone to be entered. The proximal
tip of a James peritoneal shunt is directed cephalad into the
syrinx and a few of the drainage holes are also placed in the
subarachnoid space, thereby creating a syringosubarachnoid as well as a syringoperitoneal shunt (Fig. 173). The
dura is closed, and the shunt is secured to the dura using a
purse-string suture. The distal portion of the shunt is tunneled subcutaneously to the abdominal wall and is placed
in the peritoneal cavity.

Postoperative Management Including

Possible Complications
Follow-up MRI scans are obtained 6 months after operation
to assess whether or not the syrinx has collapsed. Further
intervention, however, is reserved only for the patient who
worsens neurologically.

Difcult Cases
Occasionally we are referred patients who have undergone
one or more procedures for syringomyelia but who continue
to deteriorate neurologically. In this setting, the imaging

17 Surgical Management of Chiari I Malformations and Syringomyelia 113

Figure 172 Positioning of a patient for a syringoperitoneal shunting operation. The location of the skin incision is shown as a dashed line.

Figure 173 The proximal end of a syringosubarachnoid/syringoperitoneal shunt is being inserted into the syrinx.

114 Pediatric Neurosurgery


Figure 174 Magnetic resonance imaging scan of

a 33-year-old man with a previously treated multiloculated syrinx.

studies often demonstrate a persistent uncollapsed syrinx

containing multiple loculations. In the patient with a Chiari
I malformation who has undergone previous syringoperitoneal shunting, we believe that a craniocervical decompression is the next step. For the patient with a large syrinx but
with no craniocervical abnormality, or in a patient who has
previously undergone both shunting and craniocervical decompression, a craniocervical decompression is warranted.
As a nal option or in posttraumatic syringomyelia, a second
syringoperitoneal shunt insertion, done with care to ensure
syringosubarachnoid shunting, is performed. Caution is advised because potential damage to the spinal cord (particularly the dorsal columns) is high with repetitive shunt placements and with attempts to fenestrate syrinx loculations.
Although septations in the syrinx are usually incomplete,
occasionally a localized syrinx collapses without collapse
of the more remote portion of the syrinx following a shunting operation. It is in this setting that treatment failure is
usually seen.
Fig. 174 illustrates the MRI of a 33-year-old male who
had previously undergone two syringoperitoneal shunt
placements for right hand weakness and numbness. His
symptoms continued to progress, and MRI follow-up
studies demonstrated that the cervical syrinx, which had
multiple loculations, had increased. He subsequently underwent a suboccipital decompression and upper cervical duraplasty. Fig. 175 shows the continued presence of
a large syrinx with septations despite multiple operative

Figure 175 Continued presence of a multiloculated syrinx despite

multiple operations.

17 Surgical Management of Chiari I Malformations and Syringomyelia 115

The natural history of Chiari I malformations and syringomyelia is not fully understood, and the relative benets
of various treatment options remain unproven. We do not
operate on asymptomatic patients unless imaging studies
demonstrate a large or progressive syrinx enlargement,
because there is no certain evidence that these patients
are destined to develop a neurological abnormality. In the
symptomatic patient with a Chiari I malformation, with or

without a syrinx, we perform a suboccipital decompression

and upper cervical laminectomy with duraplasty. In patients
who harbor a large persistent symptomatic or posttraumatic
syrinx, a syringoperitoneal shunt is performed. Unfortunately, there exists a group of patients who present difcult
management problems in that they continue to deteriorate
neurologically despite surgical therapy. No uniformly satisfactory approach has been demonstrated for the treatment
of these patients.

Split Cord Malformations
Dachling Pang

Recently, Pang et al described a unied theory of embryogenesis that holds that all double spinal cord malformations
arise from a common embryogenetic error: a failure of prospective notochordal cells to achieve midline integration
following their ingress through Hensens node, allowing
the simultaneously lengthening ectoderm and endoderm
to form adhesions across this central fenestration in the
notochord. The subsequent incorporation of multipotential
mesoderm (mesenchyme) into this adhesion constitutes
the endomesenchymal tract, which not only permanently
bisects the notochord but also forces each overlying hemineural plate to neurulate against its own heminotochord
in a severely compromised manner. The basic malformation, therefore, consists of two heminotochords and two
hemineural plates separated by a midline tract containing
ectoderm, mesenchyme, and endoderm. Further evolution
of this basic form into the full-grown malformation depends
on four factors: (1) the ability of the heminotochords and
hemicords to achieve midline healing; (2) the interaction
between each heminotochord and hemineural plate during
neurulation; (3) the persistence of the endomesenmchymal
tract; and (4) the developmental fates of the three germ
Variable healing of the notochord results in the spectrum
of associated vertebral anomalies ranging from bid vertebral bodies (buttery vertebrae), to widened bodies
with midline tracts, to plain widened bodies. Partial healing of the hemineural plates results in the so-called cleft
cord, a single cord with double central canals and a deeply
indented midsection. Abnormal neurulation of the hemineural plate, hinged to the cutaneous ectoderm on only
one side and receiving mechanical and inductive inuence
from only one (lateral) set of paraxial mesoderm, results in
a misshapened hemicord with unpredictable internal cytoarchitecture varying from four healthy gray horns to a
single rudimentary gray column. Complete inability of one
or both hemineural plates to neurulate, perhaps due to an
untenable relationship with the heminotochord(s), results
in an associated hemimyelocele or myelomeningocele, respectively. Persistence of the dorsal (ectodermal) portion of
the endomesenchymal tract causes a patent dermal sinus
tract to maintain continuity with the midline septum; the
tract sometimes encysts to form a dermoid between the
hemicords. Persistence of the ventral (endodermal) portion
of the mesenchymal tract and its connection with the em-

bryonic gut explains the associations of split spinal cords

with intestinal duplication and malrotation. Finally, and in
some respects most importantly, the developmental fates of
the germ elements within the midline endomesenchymal
tract determine the state of the meningeal investment of
the hemicords, the nature of the mature septum, the presence of ganglion cells and nerve roots bridging between a
hemicord and the septum, the tethering of the hemicords to
the dorsal dura by broneurovascular bands (myelomeningocele manqu), and the rare occurrence of an enterogenous
(neurenteric) cyst in the midline cleft.
Based on their proposed theory, Pang et al suggested replacing the terms diastematomyelia and diplomyelia (which
would imply different embryogenetic mechanisms for the
two types of double cord malformations) with the nomenclature split cord malformation (SCM). They introduced a
new classication of double spinal cords founded on two
easily identiable features: the dural arrangements of the
hemicords and the nature of the midline septum.
If specialized mesodermal cells destined to form dura and
neural arch (from the meninx primitiva), normally found in
the region between the notochord and neural tube, are incorporated into the endomesenchymal tract, a median dural
layer forms next to the medial aspect of the hemicord and
joins the dura that normally grows around the lateral aspect
of the hemicord to complete a separate dural tube for each
hemicord. Additionally, in accordance with their sclerogenic
function, the meninx primativa cells within the endomesenchymal tract facing away from the hemicords also form
a midline bone spur between the two median dural walls,
continuous with the bone of the developing vertebral centrum. This conguration, called type I SCM, therefore consists of two hemicords, each contained within its own dural
tube, separated by a dura-sheathed rigid osseocartilaginous
median septum. Inasmuch as the endomesenchymal tract
frequently reaches the neural arches, the median bone spur
bisects the spinal canal into two separate compartments.
The spinal cord is transxed solidly to the spinal canal by
the bony and dural septa. The sclerogenic effect of the meninx primativa cells when these admix with cells of the
developing neural arches accounts for the often massively
hypertrophic fusion of several adjacent laminae at the level
of a type I SCM (Fig. 181).
In contrast, the endomesenchymal tract in a type II SCM
does not recruit meninx primativa cells. A thin brous sep-


14535_C18.indd 116

4/23/08 3:14:14 PM


Split Cord Malformations 117


Figure 181 Type I SCM. (A) The typical features

are depicted. (B) Axial view from computed tomographic myelogram showing a sagittal bony
septum (BS), double dural sac, and hypertrophic
neural arches.

tum, texturally different from dura, will form from the ordinary mesenchyme in the space between the hemicords.
Here also, no arachnoid, bone, or cartilage will form. Both
hemicords will lie within a single arachnoid and dural tube
inside a noncompartmentalized spinal canal, separated
by a brous rather than a rigid osseocartilaginous median
septum (Fig. 182). However, this brous septum is always
adherent to the medial aspect of the hemicords, and by virtue of its rm peripheral attachment to the ventral and/or

14535_C18.indd 117

dorsal dural wall, it is as real a tethering lesion as the bone

spur of a type I SCM.
The determining features of this classication do not
overlap between the two types; there is never a type I SCM
with dual dural sacs that does not have a rigid midline bone
or cartilage within the median dural cleft, nor is there ever
a type II SCM with a single dural sac but a naked piece of
bone or cartilage unlined by dura. Typing is thus easily made
with high-quality neuroimaging studies, an important pre-

4/23/08 3:14:14 PM

118 Pediatric Neurosurgery

Figure 182 Type II SCM. (A )The typical features are depicted. (B) A midline lling defect just ventral to the hemicords, representing an axial view
from a computed tomographic myelogram showing a single dural sac and a small ventral brous septum.

operative step because the surgical techniques are different

for each of the two types of SCM. The other commonly associated features of SCM such as paramedian nerve roots,
myelomeningocele manqu, dermoids, dermal sinus tracts,
centromedian blood vessels, thickened lums, and intestinal anomalies occur not only in both types of SCM but also
in relatively similar frequencies.

14535_C18.indd 118

Patient Selection
Thus, both types of SCM are tethering lesions. As with all
other tethering lesions, the mere presence of an SCM in a
child is sufcient indication for surgical release of the cord
because neurologial deterioration is very common in these
children, and because lost function is seldom reclaimable

4/23/08 3:14:15 PM

when treatment is rendered late. In both types of SCM, the
operation aims at removing the median septum and all
other associated bands, dermoids, lipomas, and enterogenous
cysts that might be attaching or otherwise anchoring the
hemicords to the surrounding dura. Unlike the obvious bony
dural septum in a type I lesion, the thin brous septum in a
type II SCM may not show up on preoperative neuroimaging studies. This should not deter an exploration, because
in every case of type II SCM the author has explored, a taut
brous or brovascular septum has been found to tether
the hemicords. In addition, most SCMs located in the low
thoracic or lumbosacral region have at least one associated
lesion tethering the tip of the conus, which must also be
removed during the same procedure.
In contrast to children with SCM, the evidence to support
prophylactic surgery in asymptomatic adults with SCM is
much less convincing, and most adults have been operated
on for symptoms and/or progressive decits. There are currently no available data on the natural history of tethered
cord syndrome in asymptomatic adults, but it is known that
neurological deterioration can be precipitous after a fall or
strenuous exercise. I therefore recommend operating on
asymptomatic adults who are otherwise healthy and lead a
physically vigorous life, but managing conservatively those
who are old or inrm, or who have a sedentary lifestyle.

Preoperative Preparation
Preoperative Neuroimaging Studies
Magnetic resonance imaging (MRI) is an excellent screening
test but will miss the details of structures within the median
cleft. Computed tomographic myelography (CTM) with iohexol is more sensitive than MRI for displaying ne, soft tissue bands and associated myelomeningocele manqu, and
it also shows the bony anatomy (such as the neural arches)
to great advantage. It is superior to MRI in providing important information for precise localization of the septum and
for delineating the size, obliquity, and relationships of the
type I median septum. CTM is strongly recommended as the
surgical roadmap for all SCMs.

Intraopearative Neurophysiological Monitoring

Real-time intraoperative somatosensory evoked potentials
(SEPs) are measured on all patients using standard needle
electrodes overlying the common peroneal, posterior tibial,
and median nerves. For monitoring the lower sacral sensory nerves and the S-2 to S-4 segments, pudendal SEPs are
also obtained using pairs of disk or needle electrodes on
either side of the penis in the male, and on the preclitoral
skin and labia majora in the female. Occasionally, the lower
sacral motor roots and the ventral conus segments need
to be evaluated separately from the corresponding sensory
pathways; for this, external urethral sphincter electromyography and external anal sphincter manometry using an anal
balloon are also used during surgery. The latter techniques
are especially useful to distinguish functioning lower sacral
roots from taut brous bands that tether the conus.

14535_C18.indd 119

Split Cord Malformations 119

Operative Procedure
Patient Positioning
All patients are placed in the full prone position with parallel
trunk rolls resting against the clavicles and anterior superior
iliac spines. The abdomen must hang freely between the
rolls to minimize venous bleeding. All electrophysiological
stimulating electrodes and sphincter pressure monitors are
checked for optimal functioning after the patient is settled
in the nal operative position but before operative drapes
are applied.

Surgical Technique
The aim of surgery is to relieve the cord of the tethering
effect of the median mesenchymal septum. Because the relationship between septum and hemicords is so drastically
different between type I and type II SCMs, the surgical techniques for the two lesions are also very different.

Type I SCM
Planning the skin incision requires knowledge of the exact
vertebral level of the median septum. This may be determined accurately from the preoperative CTM, then localized
to the patients spine using a preoperative plain x-ray lm
and a surface metal marker, and nally conrmed at operation with an intraoperative x-ray lm after exposure of the
spinous processes. A linear midline skin incision is made
to span at least two laminar levels above and two below
the laminae bearing the septum. In a type I SCM, the bony
septum is always extradural, being completely surrounded
and excluded from CSF by the medial walls of the double
dural tubes. The medial dural walls thus form a complete
dural sleeve for the bone in the sagittal midline. The septum itself is frequently fused with, and thus hidden under,
the neural arches so that it is not immediately visible after
the subperiosteal exposure of the posterior bony structures.
Noting the peculiar bony anatomy of the adjacent neural
arches such as a bind state, eccentric spinous processes,
exostoses, and abnormal fusion helps to guide the surgeon
to the right level. Another useful hint is that the septum is
often located where the spinal canal is widest, or where the
neural arches and spinous processes are hypertrophic and
fused with adjacent laminae into a knobby mass.
The extent of the laminectomy should include at least one
level rostral and one caudal to the septum-bearing laminae.
The hypertrophic laminae are rongeured away piecemeal
around the attachment of the septum until only a small
island of lamina is left attached to the dorsal end of the
septum (Fig. 183A). This affords a circumferential view of
the bone spur still within its dural sleeve so that its dural
attachment can be safely dissected off the bone deep within
the cleft. However, it must be understood that once the dorsal support of the septum (by the laminae) is eliminated,
the septum is no longer anchored rigidly at both ends and
might be pushed from side to side depending on its ventral
anchorage. Excessive lateral movement of the septum thus
may injure the subjacent hemicords and must be avoided.

4/23/08 3:14:16 PM

120 Pediatric Neurosurgery

A Woodson dental elevator with a thin, angulated, sharp
edge is well suited to peel off the dura with minimal lateral
wedging motions (Fig. 183B). In the majority of cases,
the broader attachment of the septum is dorsal; its ventral
junction with the vertebral body is usually narrow or even
brocartilaginous rather than bony, which makes it easy to
avulse the septum from its deep attachment (Fig. 183C).
If the ventral attachment happens to be broad and bony,
the ventral stub of the septum may be removed by a small
pituitary rongeur or microdrill. Time is well spent during
this part of the operation because complete extradural removal of the bone spur greatly facilitates later resection of
the dural sleeve.
In most type I lesions, the rigid septum is either purely
bony or partly bony and partly brocartilaginous. Purely
cartilaginous septa are rare. Bony or cartilaginous, the septum always encloses one or more prominent central blood
vessels, which can give rise to brisk bleeding if torn. A quick
plunge with a small piece of bone wax on a cotton patty
deep into the cleft should handle the bleeding adequately.
The dura is opened on both sides of the now-empty dural
cleft to isolate the sagittal dural sleeve (Fig. 184A). The
medial aspect of each hemicord is often tightly adherent
to the dural sleeve by brous bands that must be cut (Fig.
184B). Paramedian dorsal nerve roots, when present in a
type I lesion, typically stretch from the dorsomedial aspect of
the hemicords to end blindly within the median dural sleeve.
These are nonfunctional and must be cut prior to resection of
the dural sleeve. The dural sleeve is always wedged against
the caudal reunion site of the hemicords, and any free part
of the hemicords not closely apposed to medial mesenchymal
structures would be rostral to the septum. In a widely split
cord, this free area is considerable and constitutes a safe area
to begin resection of the dural sleeve. Proceeding caudally
from the rostral margin of the sleeve where the hemicords
are least adherent, the surgeon cauterizes the ventral attachment of the sleeve to seal the central vessels and then cuts
it ush with the ventral dural wall (Fig. 185A). Not having
the enclosed bone spur greatly simplies the deep resection
of the sleeve. The most hazardous part of this undertaking
is at the caudal end where the hemicords reunite and hug
tightly against the caudal margin of the sleeve, where the
taut pressure on the cord is readily felt and where slight upward migration of the cord might sometimes be seen right
after the whole sleeve has been resected.
Complete resection of the dural sleeve exposes the ventral
extradural space in the sagittal midline (Fig. 185B). If a
prominent bony stump is left from the extradural removal
of the septum, it can be trimmed down easily with a small
ophthalmic rongeur until it is no longer in contact with the
ventral surface of the hemicords. Closure of the anterior dural defect is unnecessary because of the abundant adhesions
of the ventral dura to the posterior longitudinal ligament
that would naturally prevent CSF leakage. Anterior dural
closure may actually be undesirable because the anterior
suture line potentially increases the likelihood of anterior
cord tethering. Posterior dural closure ultimately converts
the double dural tubes into a single sac.
Occasionally, a broneurovascular stalk containing paramedian dorsal nerve roots, brous bands, and large blood

Figure 183 Exposure of a type I bony septum. (A) After the laminae
rostral and caudal to the midline septum are resected, the septumbearing laminae are removed carefully around the dorsal stump of
the septum where it is attached to the ventral surface of the hypertrophic neural arches. (B) The surgeon then performs subperiosteal
separation of the median dural sleeve from the bony septum with a
small dental elevator. (C) The bony septum is then avulsed from its
ventral attachment, which is often slender.

14535_C18.indd 120

4/23/08 3:14:16 PM


Figure 184 Intradural exposure of a type I SCM. (A) A dural opening is

made from rostral to caudal along the medial edge of the median dural
sleeve (dashed line). (B) After lateral retraction of the dura, the median

Split Cord Malformations 121

dural sleeve is isolated except ventrally. Note the paramedian dorsal

roots attaching the medial aspect of each hemicord dural sleeve.


vessels tautly tethers the hemicords to the dorsal dura (myelomeningocele manqu). These bands always penetrate
the dura at a level more caudal than their origin from the
hemicords, and they often form an exuberant tuft of vascular broadipose tissue clinging tenaciously to the outer
surface of the dura. These bands must be cut ush with the
hemicords to complete the untethering process. The tuft of
extradural broadipose tissue, which cannot be swept away
with the normal extradural fat, is a clue to the presence of
the myelomeningocele manqu underneath.

In all cases of type II SCM, some form of brous (mesenchymal) septum is found within the midline cleft. The aim of
surgery is to remove the tethering effect by resecting the
brous septum and any associated bands such as those of a
myelomeningocele manqu.
Three patterns of such nonrigid median septa are found in
type II lesions: (1) The least common is a complete brous
septum stretching between the ventral and dorsal surfaces
of the dural sac. The septum is entirely intrathecal. Except for
this feature and the fact that it is nonosseous, the complete
brous septum transxes the hemicords to the surrounding

14535_C18.indd 121

In 5 to 10% of type I SCMs, the bony septum is oblique. It arises
from the midline posterior surface of the vertebral body but
then immediately reaches diagonally across the spinal canal
to divide it into two asymmetrical compartments. Without
exception, the hemicord contained within the larger compartment (major hemicord) is much larger than the hemicord in
the smaller compartment (minor hemicord), sometimes by a
factor of 2 or 3. Moreover, the larger hemicord frequently possesses one set of lateral ventral roots but two sets of dorsal
roots, whereas the smaller hemicord gives off only a single set
of ventral roots.
In these unusual cases, the exposure of the minor hemicord
is hampered because it is partly sheltered by the overhanging oblique bone spur as well as being ventrally rotated away
from the surgeons view. Moreover, the smaller hemicord is
extremely delicate. It can thus be injured inadvertently during the removal of the bone spur. This unusual pattern of
asymmetric splitting must be recognized through preoperative imaging as a signature of heightened risk, so that the
surgeon can make every effort to avoid jarring the delicate
minor hemicord while dealing with the blind underside of the
oblique bone spur.

4/23/08 3:14:18 PM

122 Pediatric Neurosurgery

Figure 185 Resection of the median dural sleeve of a type I SCM. (A)
The dural sleeve is resected ush with the ventral dural surface, proceeding from the rostral free part to the caudal end. Note the thickened lum terminale. (B) A ventral dural defect is left after complete

resection of the dural sleeve ush with the vertebral body. Note the sectioning of the thickened lum after it has been cauterized. The conus is
now completely untethered.


dura in the same manner as does the type I bony septum. (2)
Slightly more common is the purely ventral brous septum.
Its intimate adherence to the ventromedial aspects of the
hemicords in effect anchors the cord ventrally where the incomplete septum fuses with or penetrates the ventral dura.
(3) The most prevalent kind is the purely dorsal septum that
attaches the dorsomedial aspects of the hemicords to the
dorsal dura. A tuft of brovascular tissue in the extradural
space is sometimes found connected to the septum through
a small defect in the dorsal dura.
Hypertrophic and fused laminae, common in type I lesions,
are seldom found in type II SCMs. In fact, the neural arches
of type II lesions are often attenuated or even bid. Laminectomy for these SCMs is technically easy and safe. However,

14535_C18.indd 122

except where there is a tuft of brovascular tissue clinging

tenaciously to the dura, marking the site of a dorsal septum,
the exact location and extent of the purely intradural type II
brous septum is usually not apparent after dural exposure.
This underscores the importance of pinpointing the precise
location of the type II septum before laminectomy by carefully studying the bony anatomy on the CTM.
A midline dural opening immediately exposes the purely
dorsal septum and the dorsal portion of a complete septum,
to one side of the dural incision (Fig. 186A). A purely ventral septum and the ventral part of a complete septum have
to be sought, either between the hemicords or by gently rotating the hemicords to one side (Fig. 186B). Like the type
I bony septa, type II brous septa are found near the caudal

4/23/08 3:14:19 PM


Split Cord Malformations 123

Figure 186 Treatment of a type II SCM. (A) Shown is a type II SCM
with a purely dorsal brous septum attached to the medial aspects of
the hemicords. Note the direction of the septum, which points caudally
toward the dorsal dura. The large broneurovascular stalk of a myelomeningocele manqu is just caudal to the midline septum. Ex = an extradural broadipose tuft attached to the myelomeningocele manqu;

end of the split. However, the length of split in type II lesions

is, by comparison, much shorter than that of the type I SCM,
and, because brous septa are thin, the hemicords in type
II SCM are apposed much closer together, with very little
free part. Intra-cleft exploration in some type II lesions,
therefore, should not be attempted. Fortunately, the spinal
canal at the site of the split is usually wider than usual and
thus will allow the hemicords to be rolled gently to one side
for ventral exploration.

14535_C18.indd 123

MM = myelomeningocele manqu; N = a paramedian dorsal nerve root

within the septum. (B) A type II SCM with a purely ventral brous septum. Note the anterior marginal artery at the edge of the septum, and
the direction of the septum, pointing caudally toward its ventral dural

The shape of the type II septum varies from a broad rectangular or trapezoid sheet to a narrow triangular sail, but
one invariant feature is that the point of attachment between hemicords and septum is usually rostral to the point
of attachment between dura and septum (Fig. 186A). This
is true of all three kinds of brous septa, giving the appearance that the brous septum is dragged upward by rostral movement of the cord occurring after formation of the
primordium of the septum. This upward dragging converts

4/23/08 3:14:20 PM

124 Pediatric Neurosurgery

the incomplete (ventral or dorsal) septum into a backwardpointing oblique sheet and the complete septum into a Vshaped sheet with the apex pointing rostrally.
Centromedian blood vessels are always seen with the type
II septa, either as a marginal artery skirting the rostral or
caudal edge of the septum (Fig. 186B), or as a leash of parasagittal vessels loosely incorporated with the septum on
either side. Very rarely, a true arteriovenous malformation
is seen, with large vessels weaving in and out of the median
cleft. Unless these large vessels are contributing to the tethering of the hemicords, they should be left alone.
Paramedian dorsal nerve roots and myelomeningocele
manqu are commonly found in type II SCMs (Fig. 186A).
Such roots invariably course dorsally after emerging from

Figure 187 A composite SCM (type I-type II-type I combination). (A)

An axial view from a computed tomographic myelogram (CTM) shows a
type I SCM at T-12 with an oblique bony septum and double dural tubes.

the dorsomedial aspect of the hemicords. The puny nerve

roots either end blindly in the septum or penetrate the
dorsal dura with the septum. The more robust broneurovascular stalks of the myelomeningocele manqu always
penetrate the dorsal dura at a level more caudal than their
hemicord origin. These bands contribute to the tethering
and must be cut ush with the cord surface.
Untethering is completed by simply cauterizing the central vessels and excising the median brous septum. In some
cases of complete sepia, the brous and vascular elements
are seen coming through a small defect in the ventral dura,
presumably where the original endomesenchymal tract
arose. This small defect never leaks CSF and does not need
to be repaired.

(B) This CTM view shows a type II SCM at L-1 with an oblique brous
septum within a single dural sac. (C) This CTM view shows a type I SCM
at L-4. All three septa are coplanar in obliquity.


14535_C18.indd 124

4/23/08 3:14:22 PM

Special Circumstances
Associated Distal Tethering Lesions
Associated tethering lesions not directly connected with the
SCM are common. In my recently reported series of 39 cases,
all SCMs located below T-7 were associated with at least one
additional lesion tethering the conus, whereas the majority
of SCMs above T-7 had no other cord anomaly. The most
common conus lesion is a thickened lum, followed in prevalence by terminal and dorsal lipomas, dermal sinus tracts,
and limited dorsal myeloschisis. Thus, the entire neuraxis

Split Cord Malformations 125

must be screened carefully by preoperative neuroimaging.

The untethering is not complete unless these secondary lesions are also treated (Fig. 185).

Composite SCMs and Multiple SCMs

A composite SCM consists of two or more SCMs of differing
types occurring in tandem, with no normal cord between
each individual SCM. The most common composition of a
composite SCM is a type I-type II-type I combination (Fig.
187). Each individual component is typical of its kind: Each

Figure 188 Intraoperative exposure of a composite SCM (same lesions. The brous septum has been resected, which pro-case as shown
in Fig. 187, after the two type I bony septa have vides room in the middle; the dural sleeves now can be resected been removed. (A) a midline

type II brous septum lls the safely from a middle free space toward
the respective .crotches. interval between the median dural sleeves of
the two type I of the split cord.


14535_C18.indd 125

4/23/08 3:14:23 PM

126 Pediatric Neurosurgery

of the type I lesions has an extradural osseocartilaginous
spur, median dural sleeve, and hypertrophic neural arches,
and the middle type II lesion has a median brous septum
within a single dural tube. The three septa, although of two
different textures, are invariably continuous, suggesting
that the entire lesion results from a single (but very large)
endomesenchymal tract in which meninx primitiva precursor cells have been included at both ends to cause the type I
conguration, but they have not been included in the middle where the median septum remains brous. The total
length of the septa can be very long, sometimes spanning as
many as seven vertebral levels. The split cord is coextensive
with the septa so that the rostral and caudal reunion sites
of the hemicords hug tightly to the rostral margin of the
rostral bone spur and the caudal margin of the caudal bone
spur, respectively.
Exposure, laminectomy, and extradural resection of the
two type I bone spurs are as described above. The dural
opening begins rostral to the rostral dural sleeve, skirts
around it, returns to the midline over the type II lesion, and
finally skirts around the second dural sleeve (Fig. 188A).
Resection of the middle type II fibrous septum should then
be done to gain a free area within the midportion of the
median cleft. This will provide working room to begin the
resection of the type I dural sleeves above and below; for
the rostral dural sleeve, resection goes from caudal to rostral, whereas resection goes from rostral to caudal for the
caudal dural sleeve. This way, the manipulation of the dural sleeve is always directed toward and never away from
the tight reunion site of the hemicords at either end (Fig.
188B). Central vessels and paramedian nerve roots are
scattered throughout the extent of the long row of midline septa and all have to be divided. Release of the cord
is accomplished with total excision of all three septa and
reconstruction of a single dural sac for the entire affected
length of the cord.
If two or more SCMs occur in the same patient but are
separated by an interval of normal spinal cord, they are
true multiple SCMs. These are rare because they result from
multiple endomesenchymal tracts, that is, from multiple
embryological errors in the same neural tube. The individual
SCMs may be all type I or type II, or of each type. Their surgical treatment is as described.

Associated Dermal Sinus Tract and Dermoid Cyst

A dermal sinus tract is formed when the original connection
between the endomesenchymal tract and the cutaneous ectoderm is retained. Because of this embryological relationship, the deep end of the sinus tract is always in continuity
with the mesenchymal median septum regardless of the
type of SCM, but because the median septum is either extra- or intrathecal depending on the lesion type, the clinical
signicance of the retained dermal sinus tract depends on
the type of SCM it is associated with.
For a type I lesion, the dermal sinus tract can be traced
all the way from the cutaneous opening (the pit) through
midline defects in the lumbosacral fascia, muscles, and neural arches to the bone spur (Fig. 189A). Thus, the tract is

14535_C18.indd 126

always excluded from the CSF, and except in rare situations

where it penetrates one wall of the median dural sleeve to
involve the hemicord, it does not contribute to the tethering.
It is removed together with the bone spur before the dura
is ever opened. Even if the tract may occasionally encyst
to form a dermoid cyst, it lies outside the dural sac and
seldom becomes large enough to cause compression of the
In a type II SCM, however, the dermal sinus tract is of
necessity intradural where it retains connection with the
median brous septum. In its intradural course, the sinus
tract is often densely adherent to the hemicords or the cord
caudal to the caudal reunion site, thereby exerting a separate tethering effect on the cord. In addition, more than
50% of dermal sinus tracts in a type II SCM will develop a
dermoid cyst within the dura, often large enough to cause
cord compression (Fig. 189B). The entire intradural sinus
tract and cyst must be excised to eliminate the tethering
effect and prevent recurrence. The cyst is rst collapsed by
intracapsular evacuation of its cheesy content, and the
cyst wall is then carefully peeled off the pial surface of the
cord. Its deep end is removed with resection of the brous
median septum.

Associated Myelomeningocele and Hemimyelocele

Approximately 25 to 35% of SCMs have an associated, and
adjacent, open neural tube defect. Depending on whether
one or both hemicords are involved in the dysraphic sac,
the lesion may contain a hemimyelocele or a full-blown
myelomeningocele, respectively.
In most myelomeningoceles, the open neural placode is
terminal and thus is caudal to the SCM. Less commonly, the
open placode is segmental and may therefore be either rostral or caudal to the SCM. Because the open dysraphic sac is
usually treated at birth, by the time the SCM is diagnosed
by later neuroimaging studies and explored via a second
surgical procedure, the original neural placode would have
already developed dense adhesions to the dorsal dura.
Because the placode is always adjacent to the split cord
and may well be contributing to the tethering, it is always
freed from the dura when the SCM is being treated. The
most caudal set of intact laminae above the open spinal
defect are removed to expose normal, nonadherent dura.
Depending on whether the placode is rostral or caudal to
the SCM, it is carefully detached from the dorsal dura by
sharp dissection with microscissors either before or after
the median septum is excised.
Most hemimyeloceles are segmental. The unaffected
hemicord is usually hidden from view by the median
septum during the original sac closure, when the hemiplacode is mistaken to be the whole lesion. During the
definitive procedure for the SCM, the entire dural sac
(double or single) must be exposed to give access to both
hemicords. This often requires cutting into dense scar
tissue. Again, removing the adjacent normal laminae
helps to define the full width of the dura rostral to the
scar. After identifying and removing the median septum,
the hemiplacode is detached from the dorsal dura with
sharp dissection.

4/23/08 3:14:24 PM


Split Cord Malformations 127

Figure 189 SCM and associated dermal sinus tract. (A) In a type I SCM,
the dermal sinus tract is continuous with its bony septum and median
dural sleeve but is entirely extradural. (B) In a type II SCM, the dermal

Postoperative Management Including

Possible Complications

sinus tract is intradural. A large dermoid cyst within this tract maintains
a connection with the median brous septum (the original endomesenchymal tract).

Neurological Injury

the bone spur. This is especially true when the bony septum
is oblique and the delicate minor hemicord is tucked under
the overhanging septum within the acute angle made by the
septum and the vertebral body. The risk of hemicord injury
is minimized by the following precautions:

Worsening of neurological function occurs in <5% of patients following surgery for SCM. In general, the surgical
morbidity is higher among patients with a type I SCM, probably as a result of injury to the hemicords during removal of

1. Accurate preoperative depiction of the peculiar angulation of the bony septum and the relationship of the minor hemicord to the bony overhang.
2. Wide laminectomy at the site of the septum to improve

14535_C18.indd 127

4/23/08 3:14:25 PM

128 Pediatric Neurosurgery

exposure of both dural sacs. Without wide exposure, the
oblique bone spur may be mistaken for the lateral wail of
the spinal canal.
3. Extremely careful piecemeal resection of the dorsal end
of the septum. As more of this end of the septum is bitten away, more of the minor dural tube is exposed, which
in turn will permit further peeling of the median dural
sleeve off the underside of the septum to prepare for
more bone removal.

Cerebrospinal Fluid Leakage

Wound complications are seldom encountered in patients
with SCM except when there is a previously treated open
neural tube defect adjacent to the split cord. These children often have poor myofascial coverage as well as tenuous, scar-ridden skin over the dural sac and may develop
CSF leakage through the incision. This underscores the importance of a watertight dural closure. In addition, these
children should be managed in the fully prone position for
5 to 7 days after surgery similar to the way a newborn with
a myelomeningocele is positioned after sac closure. Heavy
sedation may be necessary to prevent an infant from wiggling too much and pulling too vigorously at the incision.

14535_C18.indd 128

Postoperative Bladder Management

Transient detrusor weakness and urinary retention is not
uncommon after the resection of a type I SCM. It can last
from a few days to several months and tends to be more
common in adults than in children. Permanent worsening
in bladder function occurs in ~3% of patients.
A bladder catheter is left in the rst 2 postoperative days
when the patient is conned to complete bedrest. The
catheter is removed on the third postoperative day as the
patient is encouraged to ambulate and use bathroom facilities. If there is obvious difculty in micturition or if a
large postvoid residual urine volume is obtained repeatedly,
the patient is given a cholinergic agent such as bethanechol
chloride to strengthen detrusor contraction, or an a-sympathetic blocker such as prazosin hydrochloride to encourage
relaxation of the internal urethral sphincter. If the postvoid
residual urine volume remains high on medication, an intermittent catheterization program is started in the hospital
and is continued on an outpatient basis. After 4 to 6 weeks,
a cystometrogram is obtained to evaluate bladder function
and to determine whether therapy can be discontinued.

4/23/08 3:14:26 PM

Tethered Cord Syndrome Secondary to
Previous Repair of a Myelomeningocele
Timothy A. Strait
Until a satisfactory technique for primary repair of a myelomeningocele in a newborn is devised, retethering with
the potential for delayed neurological deterioration will
remain a vexing problem for the pediatric neurosurgeon.
Invariably, scarring occurs between the neural placode and
the overlying dura, which xates the end of the spinal cord.
Progressive longitudinal tension builds up within the cord
because its growth is exceeded by that of the spinal column.
The greatest amount of tension exists within the cord segment between the attachments of the last pair of dentate
ligaments at T-12 and the adherent neural placode. Cord
damage probably occurs from a combination of repeated insults from both stretching with movement and compression
by the overlying most caudal intact lamina. These mechanical forces have been shown to produce vascular changes
that impair oxidative metabolism. Surgical untethering can
reverse the vascular abnormalities and prevent permanent
cord damage.
This chapter will focus on the operative management of
patients with a tethered cord syndrome arising from a previous myelomeningocele repair in the lumbar region. The
fundamental principle of the surgical procedure is a rostral-to-caudal approach, working from normal to disturbed
anatomy. This method allows for easy identication of the
important anatomical landmarks, which safely leads to the
area of tether and reduces the risk of inadvertent injury to
viable neural elements.

is to halt the progression of the neurological decit. When

a tethered cord is present on MRI in a child who has not yet
reached a major growth spurt, I generally favor intervention even in the absence of symptoms because of the risk of
deterioration. Finally, most children should undergo elective
untethering before corrective surgery for scoliosis.
Once a candidate for surgery has been chosen, the preoperative neurological examination provides insight regarding
the lowest level of functioning spinal cord.

Preoperative Preparation
This operation is performed under general endotracheal anesthesia using inhalational agents. Muscle paralysis is not
used because nerve stimulation is often necessary. Prophylactic vancomycin is administered at least 45 minutes prior
to making the incision.

In most cases, latex precautions are undertaken. This includes avoidance of all items made of latex. Most importantly, cases are scheduled as rst for the room because it
has been shown that circulating latex particles are largely
responsible for eliciting allergic reactions.

Patient Selection
Children and young adults with a previous myelomeningocele repair who develop symptomatic retethering present
commonly with progressive pain, leg weakness, and scoliosis. Physical examination is frequently difcult to interpret because the vast majority of patients have a preexisting neurological decit. The detection of a tethered cord
has become extremely accurate with magnetic resonance
imaging (MRI). Typically, an elongated spinal cord is seen
attached to the previous repair site. The terminal portion
of the spinal cord is displaced dorsally and often contains a
syringohydromyelic cavity (Fig. 191).
Patients with a symptomatic tethered cord are generally
offered operative intervention. The principal goal of surgery

Operative Procedure
The patient is placed prone on the operating table with
foam chest rolls to allow adequate respiratory excursion
and reduce intra-abdominal pressure. Older children are
placed in the knee-chest position on an Andrews spine
frame. All pressure points are padded well. The head is
turned to one side and rests on a comfortable headrest
(Fig. 192A).
The anesthesiologist, anesthesia equipment, and monitoring devices are all at the head of the operating table. The sur-


14535_C19.indd 129

4/11/08 11:36:16 AM

130 Pediatric Neurosurgery

Figure 191 Sagittal magnetic resonance imaging scan of a 4-year-old

child with a previous L-5 myelomeningocele repair. An elongated spinal
cord is seen with the terminal portion tethered to the previous operative

geon and assistant are opposite each other at the patients

sides and the scrub nurse is next to the surgeon toward the
patients feet. The operating microscope is mounted with
the assistants observation port and is draped with commercially available plastic drapes. The microscope stand is
positioned near the head of the table and is brought in from
the side opposite the surgeon.
The skin is shaved and adhesive plastic drapes are applied
around the proposed incision. When plastic reconstruction
of the previous skin incision is planned, an expanded surgical eld is required. The lumbar area is scrubbed with
providone-iodine soap and then painted with a 10% solution
of providone-iodine. The surgical eld is draped using rst
towels, then an iodinated adhesive plastic sheet, and nally
disposable sheets. The proposed incision is inltrated with
lidocaine with 1:200,000 epinephrine for hemostasis.

14535_C19.indd 130

site. The cord is displaced posteriorly and contains a syringohydromyelic


Incision and Exposure

The skin incision is in the midline, beginning over the most
caudal intact lamina; it is carried down to the fascia. A selfretaining Weitlaner retractor is used to expose the underlying fascia. With the electrocautery, the fascia is incised and
the paraspinal musculature is dissected from the spinous
process and laminae in the subperiosteal plane. Retraction
of the muscles is performed with a bilaterally bladed retractor. A complete laminectomy is performed with the Midas
Rex pneumatic drill (Medtronic, Inc., Fridley, MN) to expose
the dural tube several centimeters cephalad to the previous
myelomeningocele repair. Once the dural plane has been
established, the skin incision is then extended caudally and
carried down to the level of the dura overlying the neural
placode (Fig. 192B).

4/11/08 11:36:17 AM


Tethered Cord Syndrome Secondary to Previous Repair of a Myelomeningocele 131

Figure 192 (A) The patient is placed prone on the operating table with foam chest rolls. (B) Line drawings of sagittal and posterior views depict
the surgical anatomy of a previously repaired lumbar myelomeningocele. The cord is tethered by the neural placode scarred to the dura.

14535_C19.indd 131

4/11/08 11:36:19 AM

132 Pediatric Neurosurgery

Figure 193 Following a laminectomy of the lowermost lamina, (A) the dura cephalad to the previous repair site is opened in the midline and (B)
this opening is extended caudally to the neural placode. (C) The arachnoid is opened and (D) retracted bilaterally.

14535_C19.indd 132

4/11/08 11:36:23 AM


Tethered Cord Syndrome Secondary to Previous Repair of a Myelomeningocele 133

At this point, the microscope is brought into the operative

eld. The dura cephalad to the repair site is opened in the
midline with care to preserve the arachnoid (Figs. 193A and
193B). Dural tack-up sutures are placed and cottonoid strips
are laid over the exposed dura. Tension on the tack-up sutures
provides better exposure and relative hemostasis of bleeding
from the epidural space. The dural incision is extended caudally to the neural placode (Fig. 193B and 193C).
The dorsally displaced spinal cord is visible immediately
beneath the arachnoid. The arachnoid is opened carefully by
sharp dissection and its edges are secured to the dura with
Weck clips (Fig. 193D). Arachnoid dissection is directed

caudally to permit identication of the spinal cordneural

placode junction. Dense adhesions between the neural placode and the overlying dura are encountered. It is usually
unnecessary to dissect the entire neural placode from the
overlying dura because it ordinarily contains no functional
tissue. However, nerve roots exiting at this segment are
stimulated with a stimulator to determine whether useful
function is present (Fig. 194A). Functioning nerve roots
are preserved carefully and mandate transection of the cord
below the level at which they exit (Fig. 194B). The spinal
cord generally is transected at the cephalad aspect of the
neural placode. Usually the terminal portion of the spinal

Figure 194 (A) The arachnoid dissection is directed caudally to the
spinal cordneural placode junction; the lowest functional nerve roots
are identied. (B) The cord is transected at the cephalad aspect of the

neural placode. (C) The cord retracts rostrally and resides in an arachnoid enclosure.


14535_C19.indd 133

4/11/08 11:36:30 AM

134 Pediatric Neurosurgery

cord retracts rostrally and resides in a more ventral position
within the arachnoid enclosure (Fig. 194C). The conguration of the supercial vasculature changes from linear to
coiled after the cord tension is released. Another nding
indicating successful untethering is that the exiting nerve
roots from the lower portion of the cord follow a downward
oblique course rather than a transverse or upward course
(Fig. 194C).

A large bulbous neural placode complex or a dermal inclusion cyst requires total excision to facilitate closure. Closure
begins with a meticulous reapproximation of the previously
opened arachnoid with 100 interrupted sutures (Fig. 19
5A). This layer will prevent the transected end of the cord
from retethering to the overlying dura. If a shallow dural

Figure 195 (A) Closure of the arachnoid layer prevents the retracted end of the cord from retethering to the overlying dura. (B) A duraplasty is
performed with cadaver dura. (C) The rest of the wound is then closed in anatomical layers.

14535_C19.indd 134

4/11/08 11:36:33 AM


Tethered Cord Syndrome Secondary to Previous Repair of a Myelomeningocele 135

tube exists, duraplasty with a cadaver dural graft is performed (Fig. 195B). Otherwise, the dura is closed with interrupted 40 sutures. A paravertebral fascial layer closure
is accomplished after bilateral relaxing incisions are made.
The subcutaneous layer and nally the skin edges are closed
with sutures and staples, respectively (Fig. 195C). A sterile
dressing is applied.

Postoperative Management Including

Possible Complications

gery. Patients without an indwelling shunt are susceptible

to postoperative cerebrospinal uid leakage in the event
that a watertight dural closure is not achieved. This problem
is handled either by reoperation and suture closure or by
inserting a temporary external lumbar drain and keeping
the patient recumbent for several days. Wound dehiscence
is prevented by excising broad dermal scars in an elliptical
fashion and closing the skin without undue tension after
widely undermining the subcutaneous layer from the fascia.
The risk of retethering is reduced by a meticulous arachnoid

A postoperative low-pressure headache is avoided by keeping the patient recumbent for the rst 24 hours after sur-

14535_C19.indd 135

4/11/08 11:36:35 AM

Sectioning of the Filum Terminale
Frederick B. Harris, Naina L. Gross, and Frederick A. Boop

The tethered cord syndrome is a complex of neurological symptoms and orthopedic deformities associated with
a low-lying conus medullaris (below L23). Embryologically, abnormal retrogressive differentiation of caudal cell
mass presumably gives rise to the hypertrophied filum
terminale leading to a tethered cord. Recent radiological
advances have improved our knowledge and understanding of the pathogenesis and associated pathologies. It is a
common feature in many congenital malformations such
as spinal lipoma, diastematomyelia or split cord malformation, myelomeningocele, cloacal malformations, and
tight filum terminale. Symptomatic tethering may also
occur after longitudinal growth in patients with scarring
after myelomeningocele repair, spinal tumor resection,
or trauma. The tethered cord syndrome may be seen at
any age; however, symptoms usually begin in childhood
without gender predilection. Clinical signs and symptoms
of the tethered cord syndrome may vary and commonly
include spina bifida occulta, lower extremity weakness
with gait difficulty, muscle atrophy, a short limb, ankle
abnormality, perineal or lower extremity sensory deficits,
bladder dysfunction, pain (back, leg, or foot arches), and
kyphoscoliosis. Occult tethering may have delayed onset
of symptoms into adulthood, with a peak incidence in the
fourth decade of life.
The advent of magnetic resonance imaging (MRI) has
led to increased awareness of the tethered cord by pediatricians, orthopedists, and urologists, allowing for more
frequent diagnosis. Many cases of tethered cord are incidential ndings in neurologically normal patients due to
imaging studies performed for other reasons. Surgery for
the release of a tethered cord has become one of the most
common operations done by pediatric neurosurgeons.
Ninety percent of pediatric patients presenting with pain
have complete relief or improvement following surgical
sectioning of the lum terminale. Seventy ve percent of
patients presenting with motor symptoms have improvement in motor function following surgery. However, only
50% presenting with bowel and bladder dysfunction improve following surgery. Patients with progressive scoliosis
may experience stabilization or improvement of their scoliotic curvature with early untethering; therefore, release
of a tethered cord should generally be considered prior to
scolioisis correction.

Patient Selection
The diagnosis of a tethered cord in the newborn usually occurs after the recognition of a midline cutaneous anamoly
such as an intergluteal sinus or dimple, tuft of hair, hemangioma, cutis aplasia, or subcutaneous lipoma. The diagnosis
is also often made when infants with cloacal malformations
(cloacal exstrophy, anal atresia, omphalocoeles) are studied
with spinal ultrasound. Beyond the rst few weeks of life,
ossication of the dorsal elements may limit the use of ultrasound, therefore requiring MRI.
Loss of continence after toilet training and an associated history of constipation are the most common symptoms beyond

Figure 201 A sagittal T1-weighted magnetic resonance image

through the lumbar region demonstrates a thickened lum terminale.
The hyperintensity (arrows) indicates fatty inltration of the lum.


14535_C20.indd 136

4/11/08 11:23:34 AM

infancy. Complaints of intermittent back and leg pain, especially evoked by exertion, may be associated. These patients
most commonly have a normal neurological exam, but physical ndings may include pes cavous deformity, calf wasting,
atrophy of a gluteus muscle, asymmetry of foot size, lower
sacral hypesthesia, and mild scoliosis. Spinal radiographs may
show spina bida occulta and urodynamic studies may reveal
a neurogenic bladder. Myelography is usually not necessary
and MRI is the diagnostic study of choice (Fig. 201).
Children symptomatic for 6 months or less can generally
expect improvement in their symptoms following surgery.
On the other hand, those symptomatic for 1 year or more
may often have stabilization of their symptoms or decits
but a lower chance of neurological improvement. Once a
child becomes symptomatic, the natural history is usually
one of symptomatic progression. The best management for
asymptomatic, neurologically normal children with incidental ndings on imaging remains controversial. It is the
responsibility of the surgeon to discuss with the family the
information available on the natural history and chances of
recovery once the child becomes symptomatic. Fat in the
lum is a frequent incidental MRI nding, and if the conus
is at a normal level and there are no clinical indications of
a tethered cord, surgery is usually not recommended; however, these children should be followed clinically.

Preoperative Preparation
Careful evaluation of the axial MRI through the lumbar spine
is necessary preoperatively to document any associated pathology such as a dermal sinus tract. When this is noted, the
lum may not traverse the length of the spinal canal but attach to the thecal sac dorsally. MRI is helpful in determining
the appropriate level of the laminotomies.

Sectioning of the Filum Terminale 137

Patients are instructed to bathe well the night prior to

surgery and receive perioperative antibiotics. Patients who
already have an active infection of a sinus tract or meningitis should be treated adequately for such infection prior to
any surgical intervention.

Anesthesia Considerations
It is recommended that the anesthesiologist use a shortacting muscle relaxant during induction to allow for intraoperative stimulation of nerve roots, if necessary, and to
allow for the recognition of motor response with inadvertent stimulation of nerve roots. Intraoperative spinal cord
monitoring is not routine for a simple sectioning of a thickened lum terminale; however, it may be useful in other
pathologies such as repair of a diastematomyelia.

Operative Procedure
After securing the endotracheal tube and placement of a
Foley catheter, if warranted, the patient is placed in the
prone position on chest rolls extending from chest to the
iliac crest (Fig. 202). This prevents abdominal compression
and secondary distension of the epidural venous plexus,
reducing venous bleeding. Arms are placed on arm boards
for anesthesia access, and all pressure points are padded.
The patient is then placed in a slight Trendelenburg position
to reduce loss of cerebrospinal uid during the procedure,
which can be especially problematic in a patient with ventriculomegaly.

Figure 202 The patient is positioned prone on bolsters with padding of all extremities. A slight Trendelenburg position prevents the overdrainage
of cerebrospinal uid.

14535_C20.indd 137

4/11/08 11:23:35 AM

138 Pediatric Neurosurgery

Surgical Technique
An extended caudal-to-rostral sterile preparation and draping should be performed to allow for extension of the incision as needed. The planned incision is marked on the skin
in the midline, and the dermis is injected with lidocaine
with epinephrine. An incision is made and dissection is carried down to the level of the thoracolumbar fascia incorporating any existing dimple or sinus tract by ellipsing around
the opening in the skin (Fig. 203). Hemostasis is obtained
with the bovie and bipolar cautery. The surgeon must keep
in mind that in spina bida patients and infants that the
lamina and spinous processes may not be fully developed
or ossied. Care must be taken not to inadvertently pass
instruments into the spinal canal. The paraspinous muscles
are dissected away from the lamina in the subperiosteal
plane out to the facets, leaving the facet joint capsule intact.
A self-retaining retractor is then placed.
The partial laminectomy is then performed removing both
the inferior half of the superior lamina and the superior half
of the inferior lamina at the level of pathology. Following

the laminotomy, the dura is cleared of ligamentum avum

and epidural fat using sharp dissection and bipolar cautery.
A 2 cm segment of exposed dura is adequate for simple
lum sectioning. At this point, hemostasis is imperative to
prevent excessive bleeding from epidural veins once the
dura is opened and the thecal sac decompressed.
A midline durotomy is next performed using a sharp
scapel and hook. 40 Surgilon suture is then placed in the
dura to retract the edges (Fig. 204A). Next, microscissors
are used to open the arachnoid and expose the lum. The
lum is located in a dorsal midline position and has a different texture and color (purple hue) in comparison with
nerve roots due to fatty inltration and associated veins.
Upon identication of the lum, it is elevated with a nerve
hook and the undersurface inspected for possible adherent
nerve rootlets (Fig. 204B). The lum is carefully isolated
from all nerve roots.
The lum may be stimulated with low electrical current
if desired. Higher current is avoided because current spread
may stimulate other roots and give a false impression. Once
it is clearly identied, the lum is coagulated with bipo-

Figure 203 The surgeons view demonstrates the skin incision (straight dashed line) and the underlying laminectomy (dashed rectangle).

14535_C20.indd 138

4/11/08 11:23:36 AM


Sectioning of the Filum Terminale 139

Figure 204 (A) The dura is opened longitudinally and the lum terminale is exposed. (B) The thickened lum is elevated using a sharp
nerve hook. It is inspected carefully for adherent nerve rootlets. Elec-

tricical stimulation may be necessary to differentiate this structure

from a nerve root.


14535_C20.indd 139

4/11/08 11:23:39 AM

140 Pediatric Neurosurgery

Figure 204 (cont'd) (C) Bipolar coagulation prior to sectioning prevents bleeding from the retracted ends of the lum following transection. (D) The dura is repaired in a watertight fashion with 40 braided

nylon or silk. Note the rostral migration of the conus medullaris following sectioning of the lum.


14535_C20.indd 140

4/11/08 11:23:42 AM

lar forceps and sectioned with microscissors (Fig. 204C).
After sectioning, the ends of the lum will retract up into
the spinal canal, so one must be sure of hemostasis prior to
completing the sectioning.

Prior to closure of the dura, the intradural space is thoroughly irrigated with sterile saline to wash out any blood.
Once hemostasis is obtained and the irrigant is clear, the
dura is closed with 40 Surgilon sutures (Fig. 204D). A
Valsalva maneuver is performed to verify the integrity of the
dural closure. The epidural space is irrigated with bacitracin
in saline. Gelfoam is then placed in the epidural space to
serve as a blood patch to prevent spinal headaches. A dural
sealant such as Tisseal or DuraSeal (Conuent Surgical Inc.,
Waltham, MA) placed in the epidural space may allow earlier mobilization of the child.
Quarter percent bupivicaine can then be inltrated in the
paraspinous muscles and wound edges for postoperative
pain control. The paraspinous muscles and fascia are then
reapproximated with 20 or 30 interrupted Vicryl sutures.
The subcutaneous tissues are reapproximated, and the skin
is closed with subcuticular Monocryl sutures.

Postoperative Management Including

Possible Complications
Dermabond (Ethicon Inc., Somerville, NJ) skin glue is applied
to the wound to allow for early bathing without wound contamination. The patient is also left horizontal for up to 8
hours, thereby preventing spinal headaches. The patients
diet is advanced as tolerated and uids are encouraged. Patients are generally discharged home on the rst postoperative day once ambulatory and voiding.

14535_C20.indd 141

Sectioning of the Filum Terminale 141

Complications and Prognosis

Low-grade fevers or symptoms of meningismus may occur secondary to intradural blood. Occasionally children
may develop difculty with urination requiring intermittent catheterization from manipulation of the lower nerve
roots. All of these symptoms may prolong hospitalization;
however, most resolve spontaneously within a few days.
There is a very low risk of infection and also minimal blood
loss occurs during this procedure. Spinal uid leak is a risk
but uncommon with good watertight wound closure.
Souweidane and Drake have reported two cases in which
the sectioned lum terminale scarred into the wall of the
dura and, over time as the children grew, led to recurrent
tethering. Aside from these two cases, it is generally felt that
sectioning of a symptomatic lum is curative. Orthopedic
and motor decits should stabilize and sensory symptoms
are likely to resolve. However, follow-up MRI imaging will
always demonstrate a low-lying conus and, in the absence
of clinical history, will be interpreted by the radiologist as
showing a tethered spinal cord.
This chapter is a revision of the chapter, Sectioning of the
Filum Terminale by Frederik A. Boop, M.D., and William M.
Chadduck, M.D. The chapter appeared in the Neurosurgical
Operative Atlas, Volume # 3, edited by Setti S. Rengachary
and Robert H. Wilkins. The Neurosurgical Operative Atlas
was published by the American Association of Neurological
Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank William M.
Chadduck, M.D., for his help and efforts on the original
chapter published in the rst edition of this work.

4/11/08 11:23:43 AM

Frederick B. Harris, Naina L. Gross, and Frederick A. Boop

Diastematomyelia, or split cord malformation (SCM) type

I, is a congenital spinal anomaly in which the spinal cord
is divided longitudinally by a brous or bony septum. It
has been proposed that this malformation originates from
one basic error that occurs at the time of closure of the
primitive neurenteric canal. Formation or persistence of an
accessory neurenteric canal between the yolk sac and the
amnion occurs, creating an endomesenchymal tract that
splits the notochord and neural plate. This defect has been
reproduced in salamanders in the study of Emura, conrming the embryogenesis.
There are two types of SCM. Type I SCM consists of two
hemicords, each with its own dural tube and separated by
a dura-ensheathed rigid bony (osseocartilaginous) median
septum. Type II SCM consists of two hemicords within the
same dural tube separated by a nonrigid, brous median
septum. Type I SCMs are usually associated with cutaneous
abnormalities at the level of the split, such as a tuft of hair,
nevus, or hemangioma. Type II SCMs usually have no cutaneous abnormality at the level of the split, but spina bida
occulta is usually located in the lumbosacral region.
In many instances, syringomyelia will be recognized on
the preoperative magnetic resonance imaging (MRI) in association with the diastematomyelia. We generally plan to
fenestrate the syrinx at the time of untethering if it is large
and in the proximity of the diastematomyelia. Otherwise, it
is recommended that the syrinx be followed with interval
MRI scans. The syrinx rarely requires shunting and usually
does not progress if the untethering is successful.

Patient Selection
The diagnosis of SCM in the newborn usually occurs after the
recognition of a midline cutaneous anomaly such as an intergluteal sinus or dimple, tuft of hair, hemangioma, cutis aplasia, or subcutaneous lipoma. It may also be seen in tandem
with a myelomeningocele. The recognition of an asymmetrical lower extremity motor exam in the neonate born with a
myelomeningocele should alert the clinician to the possibility of an associated SCM. These infants are often studied with
spinal ultrasound, which may show a low-lying conus. After
the rst few weeks of life, ossication of the dorsal elements
limits the use of ultrasound, therefore necessitating an MRI.

Loss of continence after toilet training, an associated history of constipation, and scoliosis are the most common
symptoms beyond infancy. Complaints of intermittent
back or leg pain evoked by exertion may be associated.
These patients may have a normal neurological exam, but
physical findings such as pes cavus deformity, calf wasting, atrophy of a gluteus muscle, asymmetry of foot size,
lower sacral hypesthesia, and mild scoliosis may be common. Spinal radiographs may show spina bifida occulta and
urodynamic studies should show evidence of a neurogenic
If the children have been symptomatic for 6 months or
less one can generally expect improvement in their symptoms. On the other hand, those symptomatic for 1 year or
more may often have stabilization of their symptoms or deficits but may have less chance of neurological improvement.
Once a child becomes symptomatic, the natural history is
one of symptomatic progression. The best management for
asymptomatic, neurologically normal children with incidental ndings on imaging remains controversial. Whether
these children undergo surgical correction or not, lifelong
clinical followup is advised.

Preoperative Preparation
Spinal radiographs show spina bifida occulta and widened
pedicles and may define the bony spicule. Urodynamic
studies commonly reveal a neurogenic bladder. Axial computed tomography (CT) scanning is important for the preoperative understanding of the bony anatomy. MRI may
identify a thickened filum terminale and an associated syringomyelia, which will allow the surgeon to plan for an
incision long enough for simultaneous sectioning of the
filum and/or fenestration of the syrinx if it is large and in
the proximity of the SCM. Careful evaluation of the imaging studies preoperatively is necessary to determine the
SCM type and location of the separation of the hemicords
and the bony septum. CT myelography may be necessary to
visualize a meningocele manqu or a fibrous band between
the two hemicords.
All children should receive an antimicrobial bath the night
prior to surgery. Perioperative antibiotics are given prior to
the incision.


14535_C21.indd 142

4/11/08 11:35:59 AM

21 Diastematomyelia 143
Anesthesia Considerations

Surgical Technique

The anesthesiologist should use a short-acting muscle relaxant during induction to allow for intraoperative stimulation of nerve roots if necessary and recognition of a motor
response in inadvertent stimulation of the nerve roots. Intraoperative spinal cord monitoring may be of value during
this repair.

Extensive preparation and draping should be performed to

allow for extension of the incision as needed or placement
of a second incision if the lum terminale is located a distance from the SCM. A midline incision is made down to the
level of the thoracolumbar fascia (Fig. 211). The laminae
at this level are always malformed and often incomplete.
Careful dissection through the fascia and exposure of the
bony anatomy is required to avoid inadvertent penetration
through an incomplete arch and damage to the spinal cord.
Laminectomy at the level of the malformation, one level
above and one level below, is often adequate for the repair.
The dura of both hemicords is then exposed (Fig. 212A).
The canal at this level is normally quite wide. The bony
septum of the SCM can be quite vascular, which is easily
controlled with bone wax and Gelfoam. The bony septum is
then removed with a small rongeur (Fig. 212B) or a highspeed drill (Fig. 212C) down to the posterior longitudinal
ligament. A diamond bur helps to minimize bleeding from
the septum. If there is a thickened lum causing tethering
of the spinal cord, it should be sectioned at the same setting. This may require a separate incision and laminectomy
depending on the location.

Operative Procedure
After securing the endotracheal tube and placement of a
Foley catheter, the patient is placed in the prone position
on chest rolls extending from the chest to the iliac crest.
This prevents abdominal compression and secondary distension of the epidural venous plexus, thus decreasing venous bleeding. The arms are placed on arm boards and all
pressure points are padded. The patient is then placed in a
slight Trendelenburgs position to prevent excessive loss of
cerebrospinal uid during the procedure.

Figure 211 A midline skin incision (dashed line) is made over the level of the diastematomyelia. Note the accompanying thickened lum terminale,
which may require a second incision.

14535_C21.indd 143

4/11/08 11:36:00 AM

144 Pediatric Neurosurgery

Figure 212 (A) Removal of the dysmorphic laminae allows exposure
of the diastematomyelia. Periosteal elevators and other dissecting instruments should be used with care because the laminar arches may

Next, the dura is opened in the midline above the level of

the malformation, ellipsing the defect, leading into a midline
dural incision below the level of the defect (Fig. 212D). The
ventral segment of the dura separating the two hemicords
is approximated, if possible, with ne prolene sutures and
inverted knots. This is done in an effort to recreate the oor
of the spinal canal (Fig. 212E) and prevent the regrowth of
the bony septum in a young child. Next, the hemicords are

14535_C21.indd 144

be incomplete. (B) The dorsal portion of the diastematomyelia septum

is removed with a ne rongeur. (C) A high-speed air drill is used to bur
away the residual septum of bone. (continued)

explored for tethering adhesions ventral and dorsal, which

are taken down with sharp dissection using microscissors
and/or a knife. Strict hemostasis is obtained, the intradural
space is irrigated with saline, and the dura is closed with
40 Gore-Tex suture (Fig. 212F). This creates one thecal sac
containing both hemicords (Fig. 212G).
The epidural space is irrigated with bacitracin in saline.
The epidural veins are coagulated as necessary for hemo-

4/11/08 11:36:03 AM

21 Diastematomyelia 145

Figure 212 (continued) (D) Once the septum has been removed, the
dura is opened in an elliptical fashion. (E) The ventral dura is closed primarily when possible. (F) Completion of dorsal dural closure. (G) After

the ventral and dorsal incisions are closed, the two hemicords are contained within a single thecal sac.


14535_C21.indd 145

4/11/08 11:36:06 AM

146 Pediatric Neurosurgery

stasis and a Valsalva maneuver is performed to ensure watertight closure of the dural repair. A dural sealant may be used
as well. Next, layered closure of the paraspinous muscle and
fascia is performed with interrupted absorbable sutures.
The subcutaneous tissue is approximated with absorbable
sutures. The skin is closed with absorbable suture and covered with Dermabond (Ethicon Inc., Somerville, NJ).

neurological status. It is important to remember to place

all patients on a stool softener as constipation secondary to
narcotic medications is a common complication. Once postoperative muscle spasm has resolved in the child presenting with scoliosis, a repeat x-ray of the spine will generally
show several degrees of improvement in the scoliosis. This
improvement may be signicant enough to obviate the need
for spinal instrumentation.

Postoperative Management Including

Possible Complications

This chapter is a revision of the chapter, Repair of Diastematomyelia by Frederick A. Boop, M.D., and William M. Chadduck, M.D. The chapter appeared in the Neurosurgical Operative Atlas, Volume # 3, edited by Setti S. Rengachary and
Robert H. Wilkins. The Neurosurgical Operative Atlas was
published by the American Association of Neurological Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank William M.
Chadduck, M.D., for his help and efforts on the original
chapter published in the rst edition of this work.

If anesthesia support is available, an epidural catheter may

be placed under direct visualization at the time of closure to
allow for postoperative pain control and potentially earlier
mobilization. This catheter may then be removed after 24 to
48 hours, once pain is controlled by oral and/or intravenous
medications. The patient is kept in the horizontal position
for 24 to 48 hours postoperatively, thereby preventing spinal headaches. The patient is discharged once at baseline

14535_C21.indd 146

4/11/08 11:36:10 AM

Frederick B. Harris, Naina L. Gross, and Frederick A. Boop

Patient Selection

Anesthesia Considerations

A lipomyelomeningocele is a subcutaneous lipoma that

passes through a midline defect in the lumbar fascia, vertebral neural arch, dura, and meninges associated with an
abnormally low spinal cord. It is believed to be caused by a
failure of dysjunction of the cutaneous and neural ectodermal layers during secondary neurogenesis. More than 50% of
patients will present with a midline cutaneous abnormality
overlying the defect, 30% with bladder problems, and 10%
with orthopedic deformities and/or neurological decits in
the lower extremities. Most patients are diagnosed at or before birth. Spinal ultrasonography or magnetic resonance
imaging (MRI) should be performed on newborns with a
subcutaneous lipoma or midline cutaneous birthmark. In infants with normal neurological function, MRI can be delayed
until just prior to elective surgery. The appropriate timing of
surgery is still debated; however, it has been said that all children by the age of 6 years will have some neurological decit
if the lipomyelomeningocele is not repaired. This remains
contentious. It is clear that prophylactic surgery should be
entertained only if the risk of surgical complications is small.
For those who remain unrepaired, close neurological and
urological follow-up is mandatory. It is not uncommon to see
newly diagnosed patients present with neurological deterioration in their fourth or fth decade of life.

The anesthesiologist should use a short-acting muscle relaxant during induction and no muscle relaxant during the surgery to allow for intraoperative stimulation of nerve roots.
This also allows for the recognition of motor responses in
inadvertent stimulation of nerve roots. Intraoperative monitoring of spinal cord, bladder, or rectal sphincter should be
considered during the procedure.

Preoperative Preparation
MRI of the entire spine should be performed to rule out
other associated abnormalities of the spine and better dene the levels of exposure needed, as intradural extension
of the lipoma may progress several segments rostral to the
point of dural penetration. Although ultrasound is a good
screening tool, it does not dene the needed anatomical
detail in preparing for surgery. Plain lms of the spine will
likely show spina bida, fusion abnormalities, and sacral
defects in most cases. However, computed tomography (CT)
and plain lms add little additional information to that supplied by high-resolution MRI. All patients should have a detailed preoperative urological evaluation to document any
voiding decits. This is true whether or not there is a clinical
history of voiding dysfunction. There is no longer a role for
myelography in the evaluation of these children.

Operative Procedure
After securing the endotracheal tube and placement of a
Foley catheter, the patient is placed in the prone position
on rolls extending from the chest to the iliac crest. This prevents abdominal compression and secondary distension
of the epidural venous plexus, thereby decreasing venous
bleeding. The arms are placed on arm boards and all pressure points are padded. The patient is then placed in slight
Trendelenburgs position to prevent excessive loss of cerebro-spinal uid (CSF) during the procedure.

Surgical Technique
Following sterile preparation, the skin is covered with an
iodine-impregnated drape allowing for extensive caudal to
rostral exposure. The skin incision begins midline and cephalad to the subcutaneous lipoma continuing inferiorly,
ellipsing the mass. This allows for removal of excessive skin
and adipose tissue (Fig. 221). The incision returns to midline caudal to the mass and is extended low enough to expose the cauda equina below the lipoma. Dissection of the
subcutaneous lipoma is then performed down to the lumbar
fascia (Fig. 222A). It is important during the dissection to
provide only light retraction on the lipoma as the mass is
attached to the dorsum of the spinal cord. The lipomyelomeningocele is then amputated at its area of fascial penetration (Fig. 222B). Occasionally, neural elements may extend
up into the lipoma, as is seen in a lipomyelocystocele. Therefore, care should be taken to study the preoperative MRI and
during the dissection to avoid amputating neural elements
should they extend above the level of the fascia.


14535_C22.indd 147

4/11/08 11:29:06 AM

148 Pediatric Neurosurgery

Figure 221 An elliptical skin incision (dashed lines) allows for easy removal of the subcutaneous lipoma and later skin closure without redundant
tissue. 10.1055/978-1-60406-039-1c022_f001

Understanding that neural arches will be malformed or

incomplete, the lumbar fascial attachments to the lipoma
are carefully dissected free and the fascia is opened in the
midline caudal to the mass. The lamina is then exposed
with meticulous dissection, avoiding inadvertently passing
an instrument through a defect. A laminectomy is then performed both rostral and caudal to the dural defect created
by the lipoma. Once normal-appearing dura is encountered,
the dissection continues in the epidural plane to the area of
abnormality. Similarly, normal dura can be exposed caudally
and dissected up to the defect. Once the epidural space both
rostral and caudal to the lipoma is dened, the dura is then
opened with a scalpel in the midline rostral to the lipoma.
Dural sutures are then placed for retraction of the dura. The
dural opening is then extended to the stalk and dura-lipoma
interface is dened with microdissection.
In the case of a terminal lipoma, the stalk will appear as
a fatty extension of the spinal cord, taking the place of the
lum terminale. In such cases, the nerve roots are usually
free both ventral and lateral to the lipoma. Once the nerve
roots are clearly identied, the lipoma may be surgically
removed at its exit from the dorsal defect in the thecal sac.
During the dissection the use of a nerve stimulator may

14535_C22.indd 148

prove valuable in differentiating aberrant nerve roots from

scar tissue. Rectal sphincter electromyography (EMG) and
pudendal nerve monitoring may also be of use.
In other cases, the so-called dorsal lipoma usually attaches
to the dorsum of the spinal cord, just rostral to the conus
medullaris. Dorsal lipomas may occur anywhere along the
spinal cord and have been seen in the cervical and thoracic
regions as well. The nerve roots will typically exit ventral to
the lipoma. It is not unusual for nerve roots to inltrate the
lipoma or be adherent to the undersurface of the lipoma.
Again, intraoperative neurophysiological monitoring can
be of help in distorted anatomy. Remembering that one is
working over the spinal cord; the use of intraoperative somatosensory evoked potentials may be useful as well. These
lipomas may at times extend several centimeters rostrally
or distally within the pia of the normal spinal cord. In such
cases, their expansion may act like an intramedullary spinal
cord tumor, causing injury by compression. The recognition
of this pathology may require one to perform a myelotomy,
following the lipoma and debulking it with the laser to completely decompress the involved spinal cord.
The untethering requires meticulous dissection of the
nerve roots from the mass with the use of microscissors and

4/11/08 11:29:06 AM


Lipomyelomeningoceles 149

Figure 222 (A) After the skin incision has been made, the subcutaneous lipoma is dissected free from the underlying thoracolumbar fascia.
Caution should be used in retracting upon the subcutaneous lipoma
because it is attached to the spinal cord. (B) Once the penetration of

irrigating bipolar forceps (Fig. 222C). An ultrasonic aspirator

or contact yttrium-aluminum-garnet (YAG) laser is useful for
careful and meticulous debulking of the lipomatous remnants
(Fig. 222D). Once again, caution must be taken as the tran-

14535_C22.indd 149

the lipoma through the fascial defect has been dened, the subcutaneous mass can be removed and the remaining stalk followed through the
fascial and dorsal dural defects. (continued)

sition from lipoma to neural tissue is indistinct; therefore,

conservative partial resection is preferred (Fig. 222E). Care
must be taken to completely free the lipoma from the dural
attachments to allow for successful untethering.

4/11/08 11:29:07 AM

150 Pediatric Neurosurgery

Figure 222 (continued) (C) Microscissors and bipolar coagulation are
used to debulk the intradural mass. (D) Because the residual lipoma
attaches to the neural elements, the CO2 laser is used to vaporize the

remaining fatty tissue. (E) There is an indistinct blending of brous and

neural tissue at this juncture, and conservatism at this point is wise.


14535_C22.indd 150

4/11/08 11:29:11 AM

Transitional lipomas are a mixture of the dorsal and terminal variants with the rostral portion mimicking the dorsal
variant with a transition ventrally as one moves caudally in
the canal. In such cases, the conus is often bulbous or misshapen. These are particularly difcult as the nerve roots exiting the malformed conus often travel through the lipoma
and are intermingled with the fatty tissue. They may be long
on one side and short on the other. In some instances, these
malformed roots may even serve as a source of tethering.
Transitional lipomas at times may be too extensive to allow for safe untethering of the patient. In the patient with
minimal decits, one must at times terminate the procedure
rather than risk injuring the child.
Once the lipoma has been debulked, careful inspection
of the cauda equina is performed. This allows for lysis and
removal of adhesions that could potentially cause tethering. Sectioning of the lum terminale should be performed
if it appears to have an abnormal or thickened appearance.
Meticulous hemostasis is achieved and irrigation of the intradural space performed. The dura is closed primarily if
there is redundant tissue. Otherwise, a dural patch should
be placed with interrupted 40 Gore-Tex sutures.
After watertight closure of the dura, irrigation of the epidural space is performed with bacitracin in saline. A Valsalva
maneuver is performed to assess the dural closure. A dural
sealant may then be placed over the suture line to lessen
the risk of a spinal uid leak. The fascia is then closed with
interrupted absorbable sutures. The subcutaneous tissues
are also closed in a watertight fashion with interrupted Vicryl sutures. Finally, the skin is closed with interrupted 40
Surgilon sutures or 40 Vicryl Rapide. Dermabond (Ethicon
Inc., Somerville, NJ) skin glue is placed over the incision to
prevent soilage.

14535_C22.indd 151

Lipomyelomeningoceles 151

Postoperative Management Including

Possible Complications
If the dural closure is tenuous, the patient should be nursed
in the horizontal position for 48 to 72 hours postoperatively.
This will theoretically avoid spinal headaches and will allow
the suture lines to begin healing without any additional
stress. The Foley catheter is left in place while the patient remains at. Once removed, the patient may require intermittent catheterization if voiding is difcult to initiate. Bladder
function almost always returns to normal within 1 week
of surgery. It is important to begin stool softeners postoperatively to avoid constipation and straining at stool. Some
patients may have paresthesias or dysesthesias in the lower
extremities during the rst few postoperative days, which
generally resolve in a few days in the pediatric population.
In adults, this neuropathic pain tends to persist and can be
very problematic.
This chapter is a revision of the chapter, Repair of a Lipomyelomeningocele by Frederick A. Boop, M.D., and William M.
Chadduck, M.D. The chapter appeared in the Neurosurgical
Operative Atlas, Volume # 3, edited by Setti S. Rengachary
and Robert H. Wilkins. The Neurosurgical Operative Atlas
was published by the American Association of Neurological
Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank William M.
Chadduck, M.D., for his help and efforts on the original
chapter published in the rst edition of this work.

4/11/08 11:29:18 AM

Untethering of the Spinal Cord after a
Previous Myelomeningocele Repair
Frederick B. Harris, Naina L. Gross, and Frederick A. Boop

Patient Selection
The optimal timing of surgery for untethering of the spinal
cord following the repair of a myelomeningocele in infancy
is unclear. However, the development of severe pain, progressive neurological decit, changes in bowel and bladder
function, formation of a syrinx, and progressive scoliosis
are specic reasons to investigate possible spinal cord tethering. It is agreed that the chance for functional recovery
following untethering is best correlated with the duration
of symptoms prior to surgery. Patients with a few months
durations of symptoms can expect functional improvement
following untethering. On the other hand, those with more
long-term symptoms tend to have only stabilization of their
disease following surgery. Therefore, it is important that
these patients be enrolled in a multidisciplinary program
utilizing the services of neurosurgery, orthopedics, urology,
developmental pediatrics, and rehabilitation to allow for
early detection of symptomatic and functional change.
One must also keep in mind that the above symptoms
of a tethered cord may also be caused by a nonfunctioning
ventriculoperitoneal shunt. This is particularly true of the
child who develops syringomyelia. In such cases, the central
canal of the spinal cord may dilate preferentially over the
ventricle of the brain. Performing an untethering operation
on a patient with a nonfunctioning shunt carries the risk
of brain herniation. The functionality of a shunt must be
conrmed prior to any spinal surgery. The computed tomography (CT) scan of such patients can be deceiving in that
slit ventricles do not always signify a properly functioning
shunt. Therefore, at times, surgical exploration of a shunt
may be prudent.

Preoperative Preparation
In addition to evaluating shunt function, we will generally
review a high-resolution magnetic resonance imaging (MRI)
scan of the entire spine prior to untethering the lumbar
spine. Several children with spina bida cystica will have
tandem lesions such as a split cord malformation higher up
in the spine. If this has not been evaluated, one may nd that
the more rostral lesion is actually the cause of the tethering
rather than the myelomeningocele. Likewise, a good set of
plain spine x-rays is essential to evaluate which laminar

arches are intact and which are dysplastic. A standing spine

lm will also allow measurement of scoliotic curves, which
can be compared following untethering to determine if the
curves are improved.
At most pediatric institutions, latex precautions are now
routine for spina bida patients. This is critical even for children with no real history of a reaction to latex. In days past
the author has had patients anaphylax and die from this
exposure. As such, it is standard ward and operating room
precaution to avoid latex in all children with spina bida,
with a history of latex sensitivity, and who manage their
bladders with clean intermittent catheterization.

Operative Procedure
The patient is placed in the prone position on chest rolls
with all pressure points padded, paying particular attention
to areas of denervated skin to avoid formation of pressure
sores. Patients with contractures are positioned with extra
bolstering as needed. The use of latex precautions has become routine. Once again, the patient is placed in a slight
Trendelenburgs position to avoid excessive drainage of cerebrospinal uid (CSF). The back is prepped and draped in a
normal sterile fashion. The previous surgical scar is identied. A preoperative MRI will delineate the anatomy.
The repaired neural placode is usually adhered to the dorsum of the spinal canal just beneath the skin. Therefore,
the previous incision is opened in a very delicate and meticulous fashion to avoid injury to any neural elements (Fig.
231). For this reason, the skin should not be inltrated
with lidocaine and epinephrine. After opening the skin incision, dissection begins at the rst spinal level rostral to the
placode. Either a laminectomy or a partial laminectomy is
performed above the defect to expose normal dura. Dissection then proceeds distally in the epidural plane to the area
of malformation. In some cases, the dura cephalad to the
placode may need to be opened due to difculty identifying the epidural space. This is performed to identify normal
spinal cord (Fig. 232A). Care should be taken when opening the dura in this location because the spinal cord is likely
adherent to the dorsum of the thecal sac.
Once the dura is entered cephalad, the neural elements
can be dissected intradurally inferior to the region of the
previous surgical repair. Microsurgical technique and care-


14535_C23.indd 152

4/11/08 11:23:22 AM


Untethering of the Spinal Cord after a Previous Myelomeningocele Repair 153

Figure 231 The scar from the previous myelomeningocele repair may either be revised or reopened to the level of the subcutaneous fat. In a thin
patient, care should be exercised at the time of incision of the skin because the neural elements may be just beneath it.

ful hemostasis must be used when performing this dissection (Fig. 232B). The area over the dorsum of the placode
is the most difcult area of the dissection; however, it is important that a dural layer and a fascial layer be identied to
later facilitate a watertight closure. The nerve roots will be
located lateral and ventral to the placode during dissection.
Mobilization and lysis of adhesions of the adherent roots are
often required. Large radicular vessels that may supply the
placode and/or nerve roots should be spared.
Careful inspection of the placode must be performed for
the identication of any dermal inclusions. These are often
present and may not be clearly identied on the preoperative MRI. When these dermal inclusions are present, they
should be removed. If the inclusion is adherent to the placode, use of a contact yttrium-aluminum-garnet (YAG) laser
may facilitate removal.
The dissection is completed once the placode is dissected
free and drops to the ventral aspect of the spinal canal (Fig.
232C). After the placode has been released, attention must
then be focused on the thorough inspection and lysis of
adhesions from the cauda equina. The lum terminale may
also require sectioning for complete detethering.
Thorough irrigation of the intradural space is performed
with saline, and the wound is examined for hemostasis. The
dura is closed primarily or a duroplasty may be required
if the canal is compromised (Fig. 232D). It is important

14535_C23.indd 153

that the canal be patulent to avoid repeated scarring and

tethering. The dura or dural substitute is closed with 40
Gore-Tex sutures. A dural sealant (DuraSeal) may also be
applied to the suture line. The epidural space is irrigated
with bacitracin in saline. The remaining layers are closed in
a watertight fashion with interrupted sutures and the skin
is closed with 40 Vicryl Rapide sutures.

Postoperative Management Including

Possible Complications
The patient is nursed in the horizontal position for 48 to 72
hours to prevent a spinal uid leak through the wound and
minimize the formation of a pseudomeningocele. Careful attention should be made to padding and protecting pressure
points in these patients, who often have delicate skin.
In patients with shunted communicating hydrocephalus, decompression of the thecal sac may lead to a shunt
malfunction in the perioperative period; therefore, signs
and symptoms of a malfunctioning shunt such as headache,
nausea, vomiting, or lethargy should not be attributed to
low spinal uid pressure, postoperative pain, or narcotics.
A spinal uid leak through the previous scar tissue rarely
occurs in the presence of a functioning shunt. If one cant
be sure on clinical grounds, one should have a low threshold

4/11/08 11:23:22 AM

154 Pediatric Neurosurgery

Figure 232 (A) The dissection is begun by dening normal dura and
spinal cord at the level above that of the previous repair. This allows
dissection of scar tissue from neural elements in dened planes. (B) The

for ordering a CT of the head to evaluate the shunt before

increasing the pain medications. Likewise, patients with
Chiari malformations have abnormal brain stem nuclei. In
some patients, small amounts of narcotics may be sufcient
to cause respiratory compromise or aspiration.

14535_C23.indd 154

right side of the placode has been dissected away from the dura, and
attention is now turned to the left side. Microinstruments prove very
useful in this dissection. (continued)

Low-grade fevers and meningismus postoperatively may

be explained by contamination of the spinal uid with
blood. However, high fevers should be evaluated by shunt
tap to rule out bacterial meningitis. The Foley catheter is
not removed until the patient is mobile. The patient should

4/11/08 11:23:24 AM


Untethering of the Spinal Cord after a Previous Myelomeningocele Repair 155

Figure 232 (continued) (C) Once dissected free, the neural placode will
drop into the ventral aspect of the spinal canal, signaling that the dissection is complete. One must then inspect the cauda equina and section
a thickened lum terminale if it exists. Meticulous bipolar hemostasis is
achieved prior to dural closure. (D) A patulous dural closure may serve

to prevent repeated tethering of the neural elements. The authors prefer

to use nonreinforced Silastic sheeting (Dow Corning, Midland, MI) for
the closure because of its inert properties. The use of Dacron-reinforced
Silastic for duraplasty should be avoided.


14535_C23.indd 155

4/11/08 11:23:27 AM

156 Pediatric Neurosurgery

return to his/her home catheritization schedule and bowel
This chapter is a revision of the chapter, Untethering of the
Spinal Cord after a Previous Myelomeningocele Repair by Frederick A. Boop, M.D., and William M. Chadduck, M.D. The chapter

14535_C23.indd 156

appeared in the Neurosurgical Operative Atlas, Volume # 3, edited

by Setti S. Rengachary and Robert H. Wilkins. The Neurosurgical
Operative Atlas was published by the American Association of
Neurological Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank William M.
Chadduck, M.D., for his help and efforts on the original
chapter published in the rst edition of this work.

4/11/08 11:23:30 AM

Brain Abscesses
Darric E. Baty, Eli M. Baron, and Christopher M. Loftus

Brain abscesses, also termed cerebral abscesses, are focal infections of brain tissue that begin as early cerebritis
and progress through four different stages, each with its
own distinct histological characteristics. These abscesses
are managed both medically and surgically; this chapter,
however, will focus on the surgical management of brain

The epidemiology of brain abscesses varies with location
and socioeconomic status. In the United States, there are
~1500 to 2500 cases per year; in developing countries, the
incidence is much higher. Some authorities feel that the
overall incidence is increasing because of a larger population of immunosuppressed patients (e.g., AIDS, transplant
recipients) who are living longer and acquiring opportunistic infections. The male to female ratio is 3:1 to 3:2. The
reason for the higher occurrence in males is not known.

Risk Factors and Vectors

Risk factors include congenital cyanotic heart disease
(CCHD); pulmonary abnormality, from arteriovenous stula
to infection; AIDS or other immunocompromised state; penetrating head injury; bacterial endocarditis; gastrointestinal
infection; dental abscess or recent odontogenic manipulation; previous infarction or ischemia in the setting of septic
emboli in the brain; meningitis; sinusitis; and neurosurgical procedure, including placement of intracranial pressure
monitor, ventriculostomy, and halo traction devices.
In the pediatric population, the increased hematocrit and
diminished arterial partial pressure of oxygen (pO2) associated with CCHD, especially tetralogy of Fallot, put the patient
at an estimated risk of brain abscess that ranges from 4 to
7%. In addition, those with left-to-right shunts do not have
the advantage of the ltering effects of the lungs and are
more susceptible to cerebral abscesses. Patients with RenduOsler-Webers syndrome often have pulmonary arteriovenous
stulae. The lifetime risk of developing a brain abscess in patients with pulmonary arteriovenous stulae is ~5%.
In adults, the most common source for brain abscess is
the lung. Specic entities include pulmonary abscess, bronchiectasis, and empyema. Acute endocarditis is more likely

to predispose to a cerebral abscess than the subacute form,

but brain abscess in the setting of endocarditis is overall a
rare nding. An odontogenic source is usually considered
only in patients who have had procedures in the previous 4
months or who are found to have dental abscesses.
There are three primary vectors for cerebral abscess: hematogenous spread, contiguous spread, and development
following cranial trauma or a neurosurgical procedure (Fig.
241). With the advent of improved antibiotics, the most
common vector has changed from the contiguous route
(prior to 1980) to the hematogenous one. No source can
be identied in up to 25% of cases, and dental sources may
spread via the hematogenous or the contiguous route.
Hematogenous dissemination to the brain and subsequent
abscess formation result in multiple intracranial abscesses
in ~10 to 50% of cases. In AIDS patients, the occurrence of
pyogenic cerebral infection is relatively rare, and the same
organisms affect these patients as the general population.
Contiguous spread of infection into the brain usually
originates from purulent sinusitis. Local osteomyelitis and
emissary vein thrombophlebitis play a pivotal role in allowing transmission from the inside of the sinus cavities to
the cerebral surface. Abscesses disseminated via this route
are practically always singular, and they are uncommon in
infants because they do not have aerated paranasal or mastoid sinuses. Temporal lobe and cerebellar abscesses typically come from mastoid air sinus and middle ear infections.
Chronic otitis media in an adult portends a 1:10,000 risk of
brain abscess formation per year, which can be signicant
over the persons lifetime if left untreated. Nasal sinusitis
leads to frontal lobe abscesses, whereas sphenoid sinus infection may spread to the cavernous sinus, with potentially
devastating results.
Neurosurgical procedures traversing air sinuses have an
increased incidence of intracranial abscess formation, as
do penetrating traumatic injuries, particularly with persistent cerebrospinal fluid (CSF) leaks not repaired surgically. Open skull fractures are associated with a 3 to 7%
risk of cerebral abscess development. Simple aspiration is
usually inadequate for abscesses that follow penetrating
injuriesopen debridement and removal of dead tissue is
required; bone fragments may serve as niduses of infection in this setting. Even simple neurosurgical procedures,
such as halo ring placement, may be complicated by brain
abscess formation.


14535_C24.indd 157

4/11/08 11:35:38 AM

158 Pediatric Neurosurgery


Figure 241 Brain abscesses have several

possible sources. Direct extension via mastoid, middle ear, or paranasal sinus infection
and hematogenous spread from septic emboli and periodontal abscesses are recognized methods of dissemination.

The organisms most commonly isolated from brain abscess
cultures are Streptococcus species; many of these (33 to 50%)
may be of the anaerobic or microaerophilic variety. No organism is cultured in up to 25% of cases. In 10 to 30% of cases,
the cultures may grow multiple organisms, typically including anaerobes such as Bacteroides species. Common clinical
sources for polymicrobial culture results are otitis media,
mastoiditis, and lung abscess. When abscess formation is secondary to frontal or ethmoidal sinusitis, Streptococcus milleri
and Streptococcus anginosus are frequently found.
Following trauma, the most common organisms demonstrated are Staphylococcus aureus and Proteus species. In immunocompromised hosts, fungal infections from Aspergillus
fumigatus pulmonary infections are commonly implicated,
and infections from Toxoplasma gondii, Nocardia species, and
Mycobacterium tuberculosis are also seen. Gram-negative or-

14535_C24.indd 158

ganisms are frequently seen in infants because IgM doesnt

cross the placenta.

Patient Selection
In newborns, seizures, irritability, increasing occipitofrontal circumference secondary to patent sutures, failure to
thrive, and meningitis are common presenting conditions;
most newborns are not febrile but do not tend to do well.
Papilledema is rare prior to the age of 2. In adults, signs
or symptoms of increased intracranial pressure, including
headache, nausea, emesis, and lethargy, are common but
nonspecic for brain abscess. Edema surrounding the lesion often is responsible for, or at least contributory to, the
symptomatology. High fevers are unusual, but low-grade

4/11/08 11:35:38 AM

24 Brain Abscesses 159

fevers can be seen. Altered mental status, from confusion
to coma, can occur. Seizures and hemiparesis develop in 30
to 50% of cases.

Blood work may be abnormal but is typically nonspecic
for cerebral abscess. The peripheral white blood cell count
(WBC) is signicantly elevated in only 30 to 40% of patients
and blood cultures are usually negative. Erythrocyte sedimentation rate (ESR) may be normal, especially in CCHD
where the ESR is lowered by polycythemia. C-reactive protein (CRP) is also nonspecic, and levels can be elevated in
brain tumors and abscesses elsewhere in the body.
CSF analysis was abnormal in >90% of cases where lumbar
puncture (LP) had been performed in patients with brain
abscesses. The CSF ndings are not diagnostic for abscess,
and an organism is identied in only 6 to 22% of cases. The
opening pressure is typically elevated, and there may be
elevation of the WBC and protein. In the setting of known
cerebral abscess, an LP is relatively contraindicated due the
risk of transtentorial herniation.
Computed tomography (CT) scanning approaches 100%
sensitivity for detection of brain abscesses. The four characteristic stages are: early cerebritis (stage I) during the rst
3 days; late cerebritis (stage II) during days 4 to 9; early
capsule formation (stage II) from days 10 to 13; and late
capsule formation (stage IV) from 2 weeks and onward.
Steroids tend to prolong the evolution of this process and
reduce the degree of contrast enhancement. Late cerebritis
and early capsule formation can be difcult to distinguish.
Late cerebritis tends to be more ill-dened and to have thick
ring enhancement, whereas early capsule formation has a
faint rim on precontrast studies and thin ring enhancement.
Delayed scans taken 30 to 60 minutes after the contrast
bolus is administered can also aid in the determination;
lack of decay of enhancement is indicative of late cerebritis,
whereas decay of enhancement is more typical of capsule
Magnetic resonance imaging (MRI) can also be used to image cerebral abscesses (Fig. 242). In the cerebritis stages,
the affected region is hypointense on T1-weighted imaging
(T1WI) and hyperintense on T2-weighted imaging (T2WI).
After capsule formation, T1WI is usually hypointense in the
center and in the perilesional edematous region, and mildly
hyperintense along the rim; the T2WI is typically isointense
or hypointense in the center with a well-dened rim, and
hyperintense in the perilesional edematous region. Many
times it can be difcult to radiographically distinguish cerebral abscess from a necrotic neoplasm. Brain abscesses usually have less edema associated with them than high-grade
necrotic neoplasms such as glio-blastoma multiforme. Spectroscopic evaluation reveals abundant lactate and mobile
lipids, as well as several amino acid signatures that are not
seen with necrotic neoplasms. Diffusion-weighted imaging
(DWI) may be even more useful: Abscesses have restricted
diffusion, a hyperintense signal, and a low apparent diffusion coefcient (ADC), whereas necrotic tumors have a
hypointense signal and an elevated ADC.

14535_C24.indd 159

In the rare instances where contrast-enhanced CT and

MRI are nondiagnostic, or cannot be performed because of
contraindications, WBC scintigraphy with technetium-99m
(99mTc) is an excellent option, with a reported sensitivity and
specicity of 100%. This is a tool that may also be useful for
detecting recurrent abscesses, especially following surgery
when the CT may be more obscured than usual with postoperative changes.

Medical versus Surgical Treatment

Almost every patient should receive the benet of combined
medical and surgical treatment modalities. If possible, cultures should be obtained before the administration of antibiotics to improve the chances of identifying the organism
and subsequently tailoring the therapies appropriately. Certainly, patients with lesions >2.5 cm should undergo surgery. Even poor surgical candidates with normal bleeding
parameters can usually tolerate a stereotactic biopsy under
local anesthesia.

Preoperative Preparation
In cases where surgery is delayed and empiric antibiotics must
be started, there is no one single correct treatment regimen,
but it is advisable to provide broad coverage. The combination
of vancomycin, a third-generation cephalosporin such as cefotaxime, and metronidazole (or chloramphenicol) is a reasonable
initial treatment strategy. Metronidazole (or chloramphenicol)
may be substituted with a one-time dose of rifampin for posttraumatic abscesses. In patients with AIDS, empiric treatment
should incorporate sulfadiazine and pyrimethamine to include
Toxoplasma gondii in the targeted spectrum of organisms.
The use of steroids in the treatment of intracranial abscesses remains controversial. In general, if the patient exhibits signs or symptoms of elevated intracranial pressure
and has a mass lesion on CT, steroids are administered even
though they may inhibit host defenses and reduce penetration of antibiotics into the abscess cavity.
With up to half of patients with brain abscesses experiencing seizures, prophylactic anticonvulsants are recommended. Phenytoin is generally loaded and administered to
maintain therapeutic levels, unless the patient has a known
contraindication. Phenobarbital may be used in a similar
fashion, but it is typically more sedating, which is, of course,
not desirable in the setting of an intracranial lesion.

Operative Procedure
The two primary surgical options are needle aspiration and
surgical excision. Needle aspiration can be performed using frame-based or frameless stereotaxy, or intraoperative
ultrasound guidance (Fig. 243). Stereotactic methods tend
to be less invasive, usually necessitating only a twist drill
hole or a bur hole, whereas ultrasound guidance is typically
performed with a generous-sized bur hole or a small craniotomy. There is a high reoperation rate, approaching 70% in

4/11/08 11:35:39 AM

160 Pediatric Neurosurgery

Figure 242 A left apical molar abscess was the infectious source for the
left temporal brain abscess in this male patient. The patient underwent a
left posterior temporal craniotomy for excision of the abscess; microaerophilic streptococci grew from the intraoperative culture. (A) Preoperative axial T1-weighted magnetic resonance image demonstrating a left

14535_C24.indd 160

posterior temporal lobe cystic lesion. (B) Gadolinium enhancement was

seen at the periphery of the lesion. (C) The T2-weighted image showed
marked edema surrounding the lesion. (D) Postoperative imaging, including this post-gadolinium T1-weighted scan, demonstrated postoperative changes and resolution of the abscess.

4/11/08 11:35:39 AM

24 Brain Abscesses 161

Figure 243 Ultrasound guidance allows needle aspiration of the abscess, permitting removal of purulent material for culture and aiding in decompression of the mass, which facilitates dissection of the capsule.

some series, for lesions treated with needle aspiration. Excision is reserved for well-encapsulated, solitary lesions in
accessible locations, Nocardia abscesses, and posttraumatic
abscesses associated with a foreign body; the length of antibiotics can be shortened, to about 3 days in some cases of
total removal. Appropriately resected lesions have a much
lower recurrence.
General anesthesia is administered with typical intracranial precautions, unless the patients medical condition
mandates use of local anesthetics exclusively. Hyperventilation and mannitol should be available for use as needed.
Intraoperative steroids can be administered, as well as additional anticonvulsant, if the preoperative level is low or
not available.
The location of the abscess and the technique employed
largely dictate the patients specic positioning. If feasible,
the head of the table should be elevated; the patient is regularly placed in the Mayeld skull clamp (Integra LifeSciences
Corporation, Cincinnati, Ohio), unless frame-based stereotaxy is being employed, permitting rigid immobilization
and an attachment site for the Greenberg retractor system.
Appropriate maneuvers are performed to ensure adequate
venous drainage. The patients operative site is shaved,
prepped, and draped in the usual sterile fashion.
The skin incision and bone work are tailored to the
planned procedure (e.g., a small stab incision for a frame-

14535_C24.indd 161

based stereotactic surgery or a multicentimeter curvilinear

incision for a small craniotomy). If ultrasound guidance is
utilized, the dura is ooded with irrigation and the region
is scanned with the probe to ascertain the center point for
the dural incision, directly over the abscess (Fig. 244). After
reecting the dura and localizing the lesion once again with
ultrasound (Fig. 245), a brain needle is passed into the abscess cavity and the contents are aspirated, decompressing
the lesion to aid in further dissection while also acquiring
a specimen for cultures and stains (Fig. 246). If complete
resection is indicated, the procedure is carried forward, as
below; otherwise, the operative eld is irrigated copiously,
hemostasis is obtained, the dura is closed, the bone ap is
secured, and the galea, subcutaneous tissues, and skin are
When complete resection is indicated, a corticectomy
is made directly over the abscess capsule, utilizing ultrasound guidance. Gentle suction and retraction are applied
until the capsule is identied. Cottonoids are used to help
dene the plane between the capsule wall and the brain
and to protect the brain surface (Fig. 247). The capsule is
removed in its entirety following circumferential dissection and is sent for pathological and microbiological examination (Fig. 248). The resection cavity is irrigated, and
all foreign material is removed. The closing is completed in
routine fashion.

4/11/08 11:35:44 AM

162 Pediatric Neurosurgery

Figure 244 Ultrasound guidance allows accurate planning of
the dural ap after the craniotomy has been performed.

Postoperative Management and

Antibiotic coverage must be adjusted to reflect new stain
and culture data. Unless gross total resection of the abscess
and capsule was performed, intravenous antibiotics are
typically continued for 4 to 12 weeks (most commonly 6).
A peripherally inserted central venous catheter or Hickman
catheter is usually necessary. Steroid medication is quickly
tapered. Anticonvulsants are frequently continued for 1
to 2 years. Routine head CT should be acquired at 1 and
2 weeks following surgery, and emergently if the patient
deteriorates. As CT findings lag behind clinical resolution,

head CT scans need to be followed every 2 to 4 weeks after the full course of antibiotic therapy until radiographic
resolution, then every 2 to 4 months for 1 year. Emergent
CT is obtained for any patient presenting with neurological

Five to 20% of abscesses recur within 6 weeks of discontinuation of antibiotic therapy. The modern mortality rate for
brain abscess is ~10%, but for fungal abscesses in transplant
recipients the rate rises to almost 100%. Neurological disability is present in 45% of patients, and 29% have residual

Figure 245 The ultrasound probe is again used to locate the abscess and plan the corticectomy. Note that
the medial reection of the dural ap protects the sagittal sinus whenever craniotomies close to the midline are

14535_C24.indd 162

4/11/08 11:35:46 AM

24 Brain Abscesses 163

Figure 246 Internal decompression of the mass is performed
with a brain needle. The purulent material is sent to the laboratory for cultures and sensitivity testing.

Figure 247 The abscess capsule is prepared for removal using
a combination of gentle suction, dissection, and progressive retraction, following the corticectomy. The neighboring cortex is
protected with cottonoids.

14535_C24.indd 163

4/11/08 11:35:50 AM

164 Pediatric Neurosurgery

Figure 248 Excision of the capsule is performed following circumferential dissection by gently elevating and then freeing it from the
underlying cortex.

hemiparesis. Late focal or generalized seizures are present

in 27% of patients.

Patients with brain abscesses have focal brain infections
that progress through four histological stages and portend
a relatively poor prognosis for full neurological recovery.
Aggressive medical and surgical therapies are needed to
provide the best patient care, and close persistent follow-up
is required to nd and treat recurrences, even after successful initial management.

14535_C24.indd 164

This chapter is a revision of the chapter, Surgical Management of Brain Abscess by Timothy C. Ryken, M.D. and
Christopher M. Loftus, M.D. The chapter appeared in the
Neurosurgical Operative Atlas, Volume # 3, edited by Setti S.
Rengachary and Robert H. Wilkins. The Neurosurgical Operative Atlas was published by the American Association of
Neurological Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank Timothy C.
Ryken, M.D., for his help and efforts on the original chapter
published in the rst edition of this work.

4/11/08 11:35:53 AM

Unilateral Coronal Synostosis (Plagiocephaly)
James Tait Goodrich and David L. Staffenberg

The surgical treatment of plagiocephaly (unilateral coronal synostosis) has undergone several technical changes
over the last 50 years. Several different techniques, including strip craniectomies, lateral canthal advancement, facial
augmentations, and others have been developed to treat this
congenital disorder. This chapter will address the treatment
of plagiocephaly using the bandeau/forehead reconstruction techniques.

Patient Selection
Plagiocephaly is a common and easily recognized disorder
caused by a premature fusion/sclerosis of one coronal suture.
As a result of this premature fusion/sclerosis, the patient
typically develops a attening over the side of the involved
suture with a compensatory frontal bossing of the contralateral forehead (Fig. 251). As a result of unequal growing
planes, the child develops an orbital dystopia. When viewing the child frontally the orbit on the affected side is wider,
whereas the orbit on the other side is narrower. This occurs
by the frontal bone being pushed forward and down by the
brain. In other words, when viewed frontally, the child appears to have one eye wide open and the other partially
closed. To address the problem of plagiocephaly requires
correcting the orbital dystopia, reducing the compensatory
frontal bossing, and providing a symmetrical alignment of
the forehead.

Preoperative Preparation
Children with plagiocephaly are commonly diagnosed at
birth or at the least by 3 months of age. It is not uncommon
for the pediatrician to typically follow these children for 3
to 6 months before calling upon the craniofacial team. Once
the diagnosis of craniosynostosis is recognized, a series of
diagnostic studies can be helpful in clarifying the diagnosis.
A routine skull series, in most cases, will identify the sclerosed suture. A three-dimensional reformatted computed
tomography (CT) scan is extremely helpful in the preoperative planning, plus it also documents the extent of suture
synostosis. The head circumference is measured and closely
followed to assess head growth. It is uncommon to have restricted head growth in a single suture closure, though it has
been reported to occur in up to 10% of cases. It is common to
see areas of digital markings on the inner table of the skull
on CT indicating localized areas of increased intracranial
pressure. These areas of pressure rarely become clinically
signicant in the younger child (i.e., <1year of age). As part
of the preoperative evaluation, the patient and family are
seen by all members of the craniofacial team, which includes the following: neurosurgeon, plastic surgeon, pediatrician, pediatric neurologist, geneticist, social worker, and
the child life specialist. After each team member has seen
the patient and family, a conference is held and the surgical
recommendations to the family are planned.

Timing of Surgery
Few subjects provoke as much discussion among craniofacial surgeons as the one on timing of the surgery. Initially
we felt these children should be operated on as soon as
possible, even as early as 2 to 3 weeks of age. We now
prefer to wait until the child is 4 to 6 months of age. This allows for a more mature hematological system and a larger
blood volume for the anesthesiologist. In addition, the calvarial bone is rmer with a more mature matrix allowing
better contouring and placement. The 6 to 12 month age
period is critical in the developing child. This is a period
of rapid head growth, and this growth assists in the nal
remodeling of the child's head and face; therefore, it is
critical to use this period of rapid growth to assist in the

A child with an isolated case of plagiocephaly is typically
healthy and rarely has medical problems associated with a
single suture synostosis. We routinely have the pediatrician
do a well-baby checkup a day or two before the child is to
be admitted. If the child comes in with a fever and workup
shows no active bacterial infection (normal white blood cell
count with no shift), we will proceed with surgery. If the
child has a upper respiratory infection but no signs of congestion, we will also proceed. However, if the child appears
septic or an active bacterial source is identied, then the
surgery is canceled and rescheduled. Before we instituted
this protocol nearly one-quarter of our patients were canceled on the day of surgery because of an unknown fever


14535_C25.indd 165

4/11/08 11:28:50 AM

166 Pediatric Neurosurgery

Figure 251 Artist's rendering of a child with right side
coronal craniosynostois (plagiocephaly). The area over the
right coronal suture is attened; a compensatory frontal
bossing has occurred in the left forehead. The sphenoid
wing also becomes deformed, giving rise to the harlequin eye seen on a anterioposterior skull x-ray.

source. We have reduced this gure to <5% with the above

The parents are counseled on providing donor-directed
blood from the family should they want to take advantage of
this option. Due to potential blood loss all patients require at
least two large-bore intravenous lines of 20 gauge or larger.
Arterial lines are placed for monitoring blood gases, hematocrit, electrolytes, etc., during the procedure. A Foley
catheter is placed to monitor urine output. We do not routinely use steroids or anticonvulsants. Antibiotics are used
(oxacillin 50 mg/kg) beginning with a preoperative dose and
performed for 24 hours.

parallel to the patient's side at foot level. The nursing team

is placed on the opposite side at the foot of the patient. A
small Mayo stand is placed over the patient's abdomen. A
second mobile stand is placed off to the side for surgical
trays. As multiple teams are involved in a staged fashion,
several surgical trays are needed. Our nursing team has
found it benecial to keep a large table in the background
with the various tray setups. As each surgical team comes
into the eld, their instruments are placed on the mobile
table, and the working instruments needed are placed on
the Mayo stand.

Skin Incision and Flap Elevation

Operative Procedure
The patient is placed in a supine position with the head
resting in a horseshoe headrest. This headrest allows some
movement and repositioning of the head during the case.
The draping is done so that the head is fully exposed from
the nasal tip to vertex. A 180 degree access to the head and
facial region is required so no stands are placed to either
side of the head of the patient. The anesthesia team is placed

14535_C25.indd 166

A bicoronal incision is performed in a zig-zag fashion and

carried from ear to ear well behind the hairline. In most
children it is not necessary to shave the hair; rather, the incision can be carried through a parted hairline. The skin ap is
elevated separately from the pericranium and carried down
to the orbital rims bilaterally. Both orbitozygomatic sutures
have to be exposed. The pericranium is then elevated as a
second layer and also carried down to the orbital rims. The
neurovascular bundle is identied at the supraorbital notch
and elevated. There is often a thin rim of bone at the notch
that can be easily opened with a ne osteotome. The dis-

4/11/08 11:28:51 AM

25 Unilateral Coronal Synostosis (Plagiocephaly) 167

section is then carried further around and under the orbital
rims. At the completion of the skin incision and forehead
ap, both orbital rims and zygomatic sutures are exposed.
The temporalis muscles are elevated from their insertion
point at the temporal line down to the level of the zygoma.
The belly of the muscle is not incised as this will cause atrophy and later, a cosmetic deformity. The temporalis muscle
is elevated with monopolar needle tip cautery on a lowcurrent setting and winged out on its base.
The orbital rims are further dissected until the nasion
suture is exposed. Laterally the orbital rims are dissected
to the attachment of the lateral canthal ligament. We rarely
detach this ligament except in severe cases of orbital dystopia. If this ligament is to be detached, an identication
suture is placed through the ligament and then cut on the
side closest to the orbital wall. This suture is key to locating
and reattaching the ligament at the end of the operation.

Craniotomy and Craniofacial Reconstruction

The two types of reconstruction we use in surgically correcting plagiocephaly are presented. In the younger child (<1
year) with malleable bone, we elevate the original orbital
bandeau and then reshape it prior to replacement (Fig. 25
2). In those cases where the bandeau is too deformed, a new
bandeau is developed from over the calvarial vertex. (Figs.
253 and 254). In a plagiocephalic child, the forehead unit
is almost always markedly deformed; therefore, a new forehead is marked out with a Marchac template. The osteotomy
cuts are outlined (Fig. 253) and intraoperatively will be
marked out with methylene blue. The team rst makes the
decision as to which type of craniotomy is to be performed,
and then the appropriate marks are mapped out.

Figure 252 (A) Technique for elevating the original bandeau and harvesting a new forehead from the right parietal region. (B) The bandeau
and forehead have been plated (using absorbable miniplates) to make
the tiara reconstruction. In this illustration the absorbable plates are

Fig. 252 illustrates a case where the original bandeau

was elevated and a new forehead elevated from the right
parietal region. The technique for elevating these units requires a round bur (e.g., Midas Rex M-33; Medtronic Inc.,
Ft. Worth, Texas) and a foot-plated craniotome (e.g., Midas
Rex B-5). The new forehead is marked and elevated rst.
Superior cuts along the orbital rim are then made. The remaining original forehead unit is now elevated as a single
piece taking care not to injury the sagittal sinus. Once this
forehead unit is off, the frontal lobes are gently retracted
exposing both orbital roofs and bilateral sphenoid wings.
Osteotomies are made just behind the bandeau and carried
over both orbital roofs and in front of the cribiform plate.
The zygomatic arch is cut at the suture line. The nasal bone
is cut at the nasion suture. Once the bandeau is freed up it
is handed off to the plastic surgery team who construct a
tiara bandeau and forehead (see insert on Fig. 252B).
Typically one to three absorbable miniplates are used to
reconstruct this unit; we now place these plates inside of
the bone rather than on the outside. We have converted to
this technique of placement to reduce the palpable contour of the plates under the skin. The neurosurgery team
uses this interval to inspect for any dural or sinus tears,
which must be meticulously closed. The bandeau and forehead are now replaced in a symmetrical fashion and tongue
and groove to the temporal/squamosal bone. If there is an
asymmetry in the squamosal bone, radial cuts can be made
and the bone greensticked and outfractured. A bone strut
is placed from the top of the tiara to the skull (Figs. 255
and 256). The two open areas, lateral to the strut, are now
reconstructed with the remaining bone in a mosaic fashion.
A Tessier rib bender is helpful in remodeling the bone. Split
lamellar bone grafts are harvested from the inner table to

on the outside of the bone. We now place the plates on the inside
of the bone except at the lateral tongue-and-groove xation points. By
doing this we have reduced the postoperative complaints from parents,
who can feel these plates through the skin.


14535_C25.indd 167

4/11/08 11:28:55 AM

168 Pediatric Neurosurgery

Figure 253 An intraoperative view showing the methylene blue

markings for the osteostomies. The lack of a coronal suture can be
appreciated on the right side, as well as the resultant deformity in the
anterior fontanelle and metopic suture. A Marchac template has been
used to mark out the new forehead, which is being taken from just
above the bandeau.

patch any remaining defects. In our experience, bone defects >1.5 to 2.0 cm will not close, particularly in the child
>1 year of age. With the use of split-thickness bone grafts
we now rarely leave any open bone defects. The bone units
are secured into position with either 40 Nurolon or 30
Vicryl. We no longer use wire because of the risk of skin
protrusion. Metal miniplates are no longer used in the child
<3 years of age. We have moved away from the use of wire
and metal miniplates because of unacceptable migration
patterns in the growing child. The recently developed absorbable miniplates can be quite helpful, but again are used
only in areas where extra structural support is needed. As
can be seen in Fig. 255 we use absorbable sutures mostly
to stablize the aps. This technique has worked well for the
last 10 years with no cases of bone displacement requiring
further surgery.
An alternative craniofacial technique is used when the
original bandeau is too deformed to be reshaped. A new
bandeau is harvested from high over the parietal convexity
(Fig. 254). In this case the new forehead unit comes from
just behind the bandeau. The dissection techniques and osteotomy cuts are the same as explained above. The original
bandeau that is removed is cut up and used in the mosaic

Closure Technique
Once the bone units have been stabilized, the operative
eld is copiously irrigated to removed bone dust, debris,
and other potential sources of infection. The gutters formed

14535_C25.indd 168

Figure 254 (A) Artist's rendering showing the technique for making
a new bandeau, here harvested over the convexity. The new forehead
is coming from the vertex, just behind the bandeau. Also shown in
outline are the deformed skull sutures and anterior fontanelle. (B) An
intraoperative view showing the frontal bandeau and new Marchac
forehead. The orbital rims can be appreciated in the bandeau, which
has been recontoured to make it more symmetrical. The forehead was
harvested using a Marchac metal template (see Fig. 253).

where the skin aps have been folded over need particular
attention, as a good deal of debris collects here.
The pericranium is reelevated and tacked into position
with several absorbable sutures. The temporalis muscles
are reattached to the pericranium. This technique is an

4/11/08 11:28:55 AM

25 Unilateral Coronal Synostosis (Plagiocephaly) 169

Figure 255 Intraoperative photograph after the reconstruction (see Fig. 256). In this case we have used mostly absorbable sutures instead of wire and metal plates. We also now use
only absorbable miniplates in cases where extra stabilization is
needed, and they are placed inside of the craniotomy.

Figure 256 The reconstruction completed with the tiara replaced. A strut unit of bone is placed from the top of the tiara
to the skull; this is a key structural unit for stabilization. The
lateral sides of the strut are then lled in with the remaining
bone in a mosaic fashion. These bone units are held in position with absorbable sutures. In this illustration the plates are
seen on the outside; we now place these plates on the inside
with the exception of the tongue-and-groove xation on the
lateral sides.

14535_C25.indd 169

4/11/08 11:28:58 AM

170 Pediatric Neurosurgery

extremely important maneuver, for if this muscle is not
adequately repositioned, bitemporal dimpling will result
postoperatively providing an unacceptable hourglass conguration to the face.
The skin ap is reapproximated with a subgaleal closure
using an absorbable suture such as 30 Vicryl. We no longer
place any form of subgaleal drain, as doing this has reduced
postoperative infection and blood loss. Skin closure is done
with a 40 Monocyrl suture in a subcuticular fashion. SteriStrips are applied to the skin edges. A snug but nonbinding
uffy dressing is placed over the entire calvaria.

Specialized Instrumentation
The use of absorbable miniplates in craniofacial surgery has
been quite helpful. These low-prole plates allow instantaneous xation and stabilization. The use of a high-speed
drill system (e.g., Midas Rex) that has small cutting bits,
such as the C-1, helps immensely in the bone harvesting
and splitting plus decreases the overall operating time. A
reciprocating saw and thin-cutting osteotomes are helpful
in splitting the calvarial bone.

Postoperative Management Including

Possible Complications
The child is cared for in the pediatric intensive care unit for
a 24 hour postoperative period. Appropriate analgesia is
provided to reduce stress for the child. After 24 hours the
child is fed by mouth. The child is positioned by the nursing
team to lie only on the back so that no undue pressure is
applied to the forehead. Occasionally restraints are needed
to assist in keeping the child appropriately positioned. On
postoperative day 3 to 4 the child not uncommonly develops
fever in the 38 to 39C range. A routine fever workup is done
but rarely is a source found, and it is thought these fevers are
due to the blood resorption that occurs under the scalp ap.
The child is typically discharged on day 4 or 5, assuming he
or she is eating and has resumed normal behavior.

astating to the child, as all of the infected bone will need to

be removed. Fortunately this complication is very low, <2%
in most series. It cannot be emphasized enough the attention that must be paid to sterility, in particular the irrigation
done at the end of the case to remove debris and bone dust
that can be a source of potential infection.
Risk of injury to the neural structures remains very low in
all reported craniofacial series. It is the responsibility of the
neurosurgical team that no direct injury occurs to the brain
while the calvarial remodeling is underway. On our service
a member of the neurosurgical team is present throughout
the entire surgery and participates in the closure.
Signicant blood loss, despite careful attention to detail,
is always a possibility because of the size of these children.
We routinely request from the family donor-directed blood,
which is obtained 1 week before surgery. Two units are more
than adequate in 95% of cases. The single most common
cause of excessive blood loss is a tear in the sagittal sinus.
When recognized these tears have to be repaired by the neurosurgical team as quickly as possible. The team must also
be prepared for air embolism if a major tear occurs though
a clinically signicant air embolism is very unusual.
Cerebrospinal uid (CSF) leakage and the potential meningitis can be devastating to the patient. For this reason we
are extremely meticulous in the closure, always checking for
CSF leaks. It is often helpful to have the anesthesiologist provide a sustained Valsalva maneuver prior to the craniofacial
reconstruction to check for dura tears and potential leaks.
This chapter is a revision of the chapter, Unilateral Coronal Synostosis by James T. Goodrich, M.D., and Ravelo Argamaso, M.D. The chapter appeared in the Neurosurgical
Operative Atlas, Volume # 4, edited by Setti S. Rengachary
and Robert H. Wilkins. The Neurosurgical Operative Atlas
was published by the American Association of Neurological
Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank Ravelo Argamaso, M.D., for his help and efforts on the original chapter
published in the rst edition of this work.

The single most feared complication is infection, particularly infection to the harvested bone. Osteomyelitis is dev-

14535_C25.indd 170

4/11/08 11:29:02 AM

Moyamoya Syndrome in Children with
Pial Synangiosis
R. Michael Scott and Edward R. Smith
Moyamoya syndrome is a chronic cerebrovascular disorder
of unknown etiology characterized by a progressive stenosis
of the intracranial internal carotid arteries and their distal
branches. There is a compensatory enlargement of the collateral vasculature to the brain, specically, the small vessels
near the carotid apex supplying the cavernous sinus, optic
apparatus, etc.; the vessels on the cortical surface and in
the leptomeninges; and the branches of the external carotid
artery supplying the dura and skull base. The Japanese term
moyamoya was coined by Suzuki and Takaku in 1969 and
means something hazy, like a puff of cigarette smoke drifting in the air, referring to the characteristic angiographic
appearance of the abnormally dilated collateral network of
vessels that forms at the base of the brain in response to the
carotid artery stenosis.
This cerebrovascular syndrome is now more widely
recognized and easily diagnosed because of the advent
of magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA), and moyamoya has been reported throughout the Western Hemisphere with increased
frequency, usually in association with a wide spectrum of
disorders including prior cranial radiation therapy, neurobromatosis, optic glioma, and genetic disorders such
as Down syndrome. The precise etiology of this vasculopathy remains unclear, but it has become apparent that the
angiographic phenomenon is associated with a variety of
congenital syndromes and diseases that are associated with
acute or chronic structural changes or injury to cerebral
arterial walls. For this reason, we use the term moyamoya
syndrome when referring to the condition; however, about
one-half of the patients in the large recently published
Childrens Hospital, Boston series have no known cause or
The incidence of moyamoya syndrome is said to have
two age peaks: one in children <10 years of age, and the
other in adults 30 to 40 years of age. In adults, the disease
usually presents with acute subarachnoid, intraventricular, or intracerebral hemorrhage, although the majority of
the adult patients in the senior authors (RMS) series have
ischemic symptoms such as stroke or transient ischemic
attack (TIA). Most children present with recurrent TIAs,
strokes, seizures, or headaches; ~3% of pediatric patients
in The Childrens Hospital series had an intracerebral hemorrhage as their first symptom. The natural history of this
disease is unpredictable.

There is no known effective medical therapy that will halt

the progression of the arteriopathic process. Antiplatelet
agents such as aspirin help prevent the formation of thrombus at areas of vessel narrowing and are an important mainstay of treatment in these patients; our patients remain on
aspirin permanently. Calcium channel blockers afford certain
patients relief from severe headaches and may reduce TIAs.
Several surgical revascularization procedures have been
devised to increase blood ow to the chronically ischemic
hemispheres, including supercial temporal artery-middle
cerebral artery (STA-MCA) anastomosis, encephalomyosynangiosis (the placement of muscle on the exposed cerebral
cortex), omental transposition and transplantation, various
combinations of these techniques, and the procedure we
utilize most frequently in children, pial synangiosis, a modication of Matsushimas encephaloduroarteriosynangiosis
(EDAS) procedure.
The advantages of pial synangiosis are several-fold. The
operative procedure is relatively short~3 to 4 hours for
one hemisphere, and it avoids the technical challenges of
carrying out and maintaining the patency of a very narrowcaliber vessel anastomosis in a child, the diameter of whose
supercial temporal artery is frequently <0.5 mm. Donor
arteries are almost always available. Both hemispheres can
be operated upon under the same anesthesia, avoiding the
risks of repeated anesthetic induction in this population.
The procedure can revascularize an entire hemisphere, not
just a specic arterial territory. The clinical and angiographic
follow-up results have been excellent.

Patient Selection
Any child with a diagnosis of moyamoya syndrome should
be considered for pial synangiosis. The condition is invariably progressive, both clinically and radiographically, and
permanent decits can occur while patients are being observed to verify the syndrome's progressive nature.

Preoperative Preparation
The patients require no special preoperative medications,
unless dictated otherwise by preexisting conditions such as
prior radiation treatment for craniopharyngioma or hypoth-


14535_C26.indd 171

4/11/08 11:23:11 AM

172 Pediatric Neurosurgery

Figure 261 Patient positioned for bilateral synangiosis, left
side rst. The head is turned away from the anesthesiologist
and secured in a Mayeld head-holder (Schaerer Mayeld,
Cincinnati, OH). The course of the parietal branch of the supercial temporal artery has been marked out on the scalp using a
Doppler pencil-type probe. A roll has been placed under the left
shoulder to avoid compression of the cervical vessels.

the initial side is operated on, surgery on the contralateral

hemisphere is postponed.
There is no evidence that one anesthetic agent or technique is superior in patients with moyamoya syndrome.
Muscular blockade is established by a nondepolarizing muscle relaxant prior to intubation. Any uid decits are partially replaced by intravenous crystalloid without glucose
(10 mL/kg) over 15 minutes after induction. Anesthesia is
maintained with low-dose isourane and a balanced nitrous
oxide/oxygen mixture with fentanyl. The rationale for the
use of these agents is that isourane is a cerebral vasodilator
and may even provide a protective effect against ischemia.
However, any anesthetic technique that will maximize the
balance between cerebral blood ow and oxygen consumption is probably reasonable. End-tidal CO2 is maintained between 36 and 42 mm Hg. We avoid the use of hyperventilation or any anesthetic technique that would cause cerebral
vasoconstriction because hyperventilation in a child with
compromised cerebral circulation could precipitate further
ischemic sequelae. Normotension, appropriate for age, is
maintained. Diuretics such as mannitol and Lasix (furosemide) are unnecessary and possibly risky in this patient population because of the possibility of dehydration leading to

alamic glioma. Standard preoperative tests include routine

laboratory tests such as complete blood count and coagulation studies, particularly if the patient has been taking antiplatelet agents preoperatively. Unless contraindicated by
study results or other clinical indications, we now continue
aspirin right up to the time of surgery in hopes of reducing intraoperative and postoperative thrombus formation
in stenotic proximal vessels. All patients undergo selective
internal and external carotid angiography prior to surgery.
This study assists the surgeon in the identication of vascular anatomy and ow patterns, and demonstrates any preexisting collateral anastomoses from the external circulation that should be preserved during the craniotomy. Single
photon emission computed tomography with and without
Diamox (acetazolamide) administration have been used in
the past on a routine basis to detect any cerebrovascular instability, but it is currently utilized only in unilateral disease
or in patients with atypical arteriographic ndingswhen
there is a question as to whether cerebral blood ow in
unstable enough to merit surgery. We routinely admit patients the night before surgery for intravenous hydration.
If the patient is taking anticonvulsants, the morning dose
is administered prior to surgery. Patients receive one dose
of prophylactic antibiotics, tailored for the specic hospital
environment, 1 hour prior to the skin incision. Steroids, cerebral dehydrating agents such as mannitol, and anticonvulsants are not administered on a routine basis.

Operative Procedure



Premedication adequate to avoid hyperventilation and crying in children is essential, and patients undergo typical
craniotomy monitoring, including arterial lines and bladder catheterization. Intraoperative electroencephalography
(EEG) monitoring, with a full array of scalp electrodes (except directly at the operative site) is performed on all patients in whom bilateral surgery during the same anesthetic
is contemplated. If any signicant changes occur on EEG as

The patient is placed in the supine position with a roll under

the shoulder opposite the side being operated. Care must
be taken to ensure that the cervical vessels are not compressed by extreme head rotation. The head is placed in
the lateral position parallel to the oor in skeletal xation,
with a single anterior pin placed in the high midline forehead to allow rotation of the head without new pin placement if the opposite side is to be operated on under the

14535_C26.indd 172

4/11/08 11:23:11 AM


Moyamoya Syndrome in Children with Pial Synangiosis 173

same anesthesia (Fig. 261). In a child under the age of 1

year, the head is positioned on a doughnut or cerebellar
headrest if the skull does not appear sturdy enough for pin
xation. After completion of the rst procedure, the head
is turned (maintaining skeletal xation) and the shoulder
roll is placed under the opposite shoulder. If a circular or
horseshoe headrest is used in the very young child, care
should be taken to ensure that the new craniotomy site is
not compressed by the device.

ous tissue is made directly over the vessel at its most distal
marked point with a no. 15 scalpel. The artery is identied
by scalp retraction with toothed forceps and dissected using a delicate curved pediatric hemostat. A linear incision
following the course of the artery is then performed, using
the hemostat to dissect and then protect the STA as the assistant incises the skin overlying it. The skin edges rarely
require coagulation, and most scalp-edge bleeding will stop

Skin Incision

Synangiosis Procedure

The posterior (parietal) branch of the supercial temporal

artery (STA) is identied by a pencil Doppler probe, and the
artery is traced from its base above the zygomatic arch to
the parietal convexity. Its course is accurately marked on the
skin with an 18 or a 21 gauge needle. We attempt to mark
out at least 10 cm of vessel in most patients, but in small
children, often only 6 to 7 cm can be identied. Although
the posterior branch of the STA is most frequently utilized
because the skin incision can be kept behind the hairline
and the majority of the MCA circulation lies beneath that
branch of the artery, the frontal branch can be used if absolutely necessary. Standard skin prep and draping can then
be performed.
We utilize the microscope right from the beginning of the
artery dissection, nding it particularly helpful in very small
children because of the fragility and small size of STA, and in
young adults because of the frequent tortuosity of the vessel
and its branches. A small skin incision down to subcutane-

After the artery is exposed, we use a needle tip cautery

(Colorado needle) at a low setting to separate the artery
with its subjacent galea strip from the galea on either side.
The artery pedicle is then encircled with a vessel loop distally
and elevated and separated from the underlying periosteum
and temporalis fascia using standard monopolar cautery, attempting to preserve as much adventitia and loose areolar
tissue beneath the vessel as possible (Fig. 262).
Anterior and posterior scalp aps are then developed
with electrocautery dissection to minimize bleeding, and
disposable shhook-type retractors are used to maintain
scalp retraction. The artery pedicle is retracted out of the
eld as needed using the vessel loop, and the temporalis
muscle is incised with the electrocautery into four equal
quadrants, which are retracted widely using the previously
placed skin hooks. Generous bur holes are drilled inferiorly
and superiorly in the exposure, and as large a craniotomy
as possible is performed using power equipment. The dura

Figure 262 The supercial temporal artery dissection has
been completed under the microscope with the aid of a needle
electrocautery, and the vessel has been encircled with a vessel loop to aid in its displacement during the remainder of the

14535_C26.indd 173

4/11/08 11:23:13 AM

174 Pediatric Neurosurgery

Figure 263 The craniotomy has been performed, and the
scalp edges and temporalis muscle retracted with shhook
retractors (Lone Star retractor system; Lone Star Medical Products Inc., Stafford, TX). The dura is opened in multiple aps
secured with sutures.

is then opened vertically along the exposure and cut into

six separate aps that are retracted on sutures (Fig. 263);
care is taken to preserve any signicant middle meningeal
collateral observed on the preoperative arteriogram and the
dural opened around or between such vessels. Under high
power of the microscope, the arachnoid is incised linearly
over the exposed cortex using a disposable arachnoid knife
and jewelers forceps, beginning inferiorly (temporally) in
a sulcus and then opening the arachnoid linearly along the
sulcus and when possible laterally toward the crown of adjacent gyri (Fig. 264). Vannas ophthalmic scissors are helpful in making long continuous arachnoid openings over MCA

branches in certain patients. The pial vasculature is profuse

and tortuous in patients with advanced disease, and these
areas of the pia should be avoided when the arachnoid is
opened. Bleeding that occurs from the pial surface or from
small vessels within the sulci usually stops after a few moments of irrigation with a microirrigator or the application
of a minuscule pledget of Gelfoam soaked in thrombin solution. After completion of the arachnoid opening through
as much of the length of the exposure as possible, the STA
with its galea investiture is brought down onto the surface of the brain, placing the vessel over areas of opened
arachnoid. Using jewelers forceps and a Castroviejo needle-

Figure 264 The arachnoid is being opened over a sulcus adjacent to a cortical vein using jewelers forceps and a disposable

14535_C26.indd 174

4/11/08 11:23:14 AM


Moyamoya Syndrome in Children with Pial Synangiosis 175

(Rapide) to approximate the skin edges. A sterile Telfaover-Xeroform gauze dressing is applied, but the head is
not wrapped tightly to avoid compressing the scalp arteries.
Patients are extubated and sent to the intensive care unit
following the procedure.

Postoperative Management Including

Possible Complications

Figure 265 The arachnoid has been opened in a roughly linear trajectory from temporal to frontal across the sylvian ssure using a disposable arachnoid knife, jewelers forceps, and ophthalmic (Vannas)
scissors. The synangiosis has been accomplished with four interrupted
sutures of 10-0 nylon passed through pia and supercial temporal artery adventitia, with the donor vessel kept in apposition to areas of
opened arachnoid. The arrows point to parallel arachnoid openings
made over middle cerebral artery branches anterior to the major
arachnoid opening. A small focus of subpial hemorrhage can be seen
to the left of the donor vessel, an unavoidable complication in many
moyamoya patients with hypervascular cortical surfaces.

holder, the vessel is xed to the cortical surface by placing

approximately four interrupted 100 nylon sutures through
the vessel pedicle soft tissue and the outermost layer of the
pial-cortical surface. This tight pial approximation leads to
a more satisfactory postoperative result then simply placing
the vessel on the brain or suturing the vessel into the dura
(Fig. 265).
The dura is then loosely laid on the brain surface and not
sutured becasue collateralization of the underlying brain
will also occur from the cut edges of the dura. A large piece
of Gelfoam soaked in saline, or DuraGen (Integra LifeSciences, Plainsboro, NJ, trimmed to the size of the bony opening, is placed over the entire exposure, and the bone ap is
repositioned using three to four miniplates. The bur hole
openings at either end of the ap should be enlarged if necessary prior to its repositioning to ensure that the artery is
not compressed when the ap is replaced. The temporalis
muscle is approximated only from inferior to superior so
that the supercial temporal artery is not compressed. The
scalp is closed in two layers, using 30 or 40 Vicryl suture
on the galea followed by 40 or 50 monolament Vicryl

14535_C26.indd 175

There are very few special postoperative orders. The patient's head is slightly elevated to assist in venous return
and avoid cerebrospinal uid accumulations under the skin
aps. The patient is given sufcient pain medication so that
there is a minimum of crying and hyperventilation. Antibiotics are given postoperatively for 24 hours. Patients are
mobilized as tolerated and kept well hydrated. Blood loss
with the procedure should be less than 100 cc, and transfusions should be avoided if at all possible because of the
rheologic problems that can be created by a high hematocrit. Postoperative seizure medication is continued on an
individualized basis, and aspirin is begun 24 hours after
surgery. Postoperative hypertension is rarely treated unless
signicant elevations occur, and care must be taken if treatment is required to avoid rebound hypotension.
The most signicant postoperative complication in our
series has been stroke, which in a consecutive series of 143
patients occurred at ~4% per operated hemisphere. Patients
at greatest risk appear to be those with neurological instability around the time of surgery or those who have suffered a stroke within 2 months of the operationalthough
angiographic risk factors, such as involvement of the posterior circulation, must also play a role. There have been
two perioperative deaths related to ischemic stroke: one
in a 5-year-old child operated on in the midst of a urry of
strokes who developed additional strokes in the immediate
postoperative period and died of brain edema and herniation, and one in a 15-year-old boy with progressing disease
and a preoperative dominant hemisphere stroke and basilar
artery disease whose internal carotid arterythe sole supply of his posterior circulationthrombosed several hours
following a unilateral operation. There have been several
late subdural hematomas in this surgical series that have
required evacuation, a complication probably related to preexisting brain atrophy and chronic antiplatelet therapy. A
death 6 years following surgery occurred from the rupture
of an aneurysm on a deep moyamoya collateral vessel.

All patients in this series in the past had undergone arteriography 1 year following the procedure to document the efcacy of the procedure and to provide a new baseline, along
with MRI and MRA studies, for future evaluation (Fig. 266).
Because of the rapid improvements that are being made in
MRA imaging, we have begun following patients with MRI
studies alone because the formal arteriograms have conrmed the surgerys effectiveness. Late angiographic studies
have conrmed the durability of the collateral induced by

4/11/08 11:23:16 AM

176 Pediatric Neurosurgery

pial synangiosis, with one patients arteriogram repeated
out 9 years after synangiosis because of the development of
a brain tumor demonstrating even more luxuriant collateral
than the initial 1-year follow-up study.
This chapter is a revision of the chapter, Treatment of
Moyamoya Syndrome in Children with Pial Synangiosis
by R. Michael Scott, M.D., and Richard G. Ellenbogen, M.D.
The chapter appeared in the Neurosurgical Operative Atlas,
Volume # 4, edited by Setti S. Rengachary and Robert H.
Wilkins. The Neurosurgical Operative Atlas was published by
the American Association of Neurological Surgeons (AANS)
from 1991 to 2000.
We would like to acknowledge and thank Richard G. Ellenbogen, M.D., for his help and efforts on the original chapter
published in the rst edition of this work.

Figure 266 One-year postoperative magnetic resonance angiogram

in a patient who has undergone bilateral pial synangioses. The arrowheads denote new collateral to the hemispheres via the supercial
temporal and middle meningeal arteries. The intracranial internal carotid arteries are now totally occluded.

14535_C26.indd 176

4/11/08 11:23:17 AM

Selective Dorsal Rhizotomy for Spastic
Cerebral Palsy
Tae Sung Park and James M. Johnston
Cerebral palsy (CP) is a major neurological problem in children; it occurs in 1 of 500 live births and inicts lifelong disabilities on those with the disorder. Among several factors
contributing to the disabilities of patients with CP is spasticity. Spasticity affects nearly 80% of patients with CP, and once
fully developed, it never resolves spontaneously. It hinders
motor activities in daily living and also causes muscle contractures and orthopaedic deformities in growing children.
Selective dorsal rhizotomy (SDR) reduces spasticity in CP;
this reduction facilitates patients motor performance and
alleviates orthopaedic deformities. Several surgical techniques for SDR are currently used. The standard technique
requires an L1-S1 laminectomy or laminoplasty for visualization of all dorsal nerve roots exiting at their respective
foramina. The following is a description of an operation that
includes an L-1 laminectomy, ultrasonographic localization
of the conus medullaris, and partial deafferentation of L1S2 roots with electromyographic (EMG) testing under an
operating microscope. Advantages of this technique include
decreased operative time, reduced postoperative pain, as
well as minimal risk of progressive lumbar instability. This
procedure is both effective and well tolerated in appropriately selected children and adults.
We have performed this SDR in more than 1500 children
and young adults since 1991.

Patient Selection
The primary beneciaries of SDR are children with spastic diplegia, the most prevalent subtype of CP in which the
lower extremities are affected with minimal or no involvement of the upper extremities (Table 271). Children with
spastic quadriplegic CP, in whom all extremities and the
trunk are involved, also benet from SDR. In spastic hemiplegic CP, spasticity is not a predominant cause of motor
impairments, and reduction of spasticity does not greatly
improve motor functions. Some adults younger than 40
years of age who have relatively mild spastic diplegia and
can walk independently are also able to benet from SDR.
When evaluating a patient, one should rst be certain
that a patients motor impairment dates back to infancy and
has taken a course of steady improvement rather than progressive deterioration during the preschool years. A careful
review of the patients perinatal history and medical his-

Table 271 Indications for Selective Dorsal Rhizotomy
for Spastic Cerebral Palsy
Children Younger than 18 Years
At least 2 years of age
Diagnosis of spastic diplegia or spastic quadriplegia
Some form of independent mobility (e.g., crawling or
walking) with or without an assistive device
History of premature birth; if born at term, child must have
typical signs of spastic diplegia
Patients exhibit potential for improvement in functional skills
after dorsal rhizotomy
Adults between Ages 19 and 40 Years
Diagnosis of spastic diplegia
History of premature birth
Currently ambulates independently without assistive device
Relatively mild xed orthopaedic deformities
Patients exhibit potential for functional gains after dorsal
Patients exhibit motivation to perform home exercise

tory provided by physicians and therapists is invaluable.

A history of premature birth is considered a positive factor in the selection of candidates for SDR. Findings from a
neurological examination determine whether spasticity is
a major cause or the only cause of muscle hypertonia and
signicantly hinders motor activities (e.g., sitting, crawling,
standing, and walking). In addition, the severity of orthopaedic deformities and their effects on a patients motor
performance should be assessed in detail. A thorough understanding of orthopaedic problems by the neurosurgeons
performing SDR is desirable, although orthopaedists and
therapists can assist in evaluation. Radiological evaluation
includes lumbosacral spine and hip radiographs. The spine
radiographs may show the presence of lumbar hyperlordosis, scoliosis, spondylolisthesis, and congenital anomalies.
Hip radiographs may reveal hip subluxation and dislocation,
deformities that inuence the timing of surgical interventions. In children who were delivered at term, head magnetic resonance imaging (MRI) is obtained because diverse
abnormalities underlie the diagnosis of spastic diplegia or
quadriplegia. In adults with back pain, spine MRI is obtained
to rule out herniated disk and other intraspinal disorders. A
gait analysis is not routinely obtained, but it helps to con-


14535_C27.indd 177

4/11/08 11:34:15 AM

178 Pediatric Neurosurgery

rm the presence of spasticity before SDR and also to assess
changes in motor performance after SDR.
SDR is not considered for children younger than 2 years
because CP cannot be diagnosed with certainty in young
children. Dystonia that is concomitant with spasticity is not
a contraindication for rhizotomy. Dystonia is exceedingly
rare in spastic diplegia but is common in nonambulatory
children with whole body involvement. Dystonia becomes
clinically evident by age 5 years. SDR in patients with dystonia can improve motor function, and this improvement
does not worsen the dystonia. Rigidity is also rare in spastic
diplegia, but rigidity cannot be ascertained in the presence
of concomitant spasticity. Severe damage to the basal ganglia, as revealed with MRI, is considered a contraindication
because of the possibility of concomitant rigidity. A history
of multiple orthopaedic operations is generally a contraindication for rhizotomy, mainly because of severe muscle
weakness and xed deformities. Patients with severe xed
joint deformities are excluded from SDR because the deformities limit gains in motor function after rhizotomy. The
best example is severe crouch knees due to overstretched
or lengthened heel cords after heel cord release procedures.
Severe scoliosis is only a relative contraindication for SDR
through a single-level laminectomy. Children with increased
muscle tone due to severe hydrocephalus, intrauterine and
neonatal infections, and head trauma are not candidates for
SDR. In general, neuronal migration disorders are contraindications for rhizotomy, but children with typical clinical
features of spastic diplegia due to schizencephaly can benet from rhizotomy.

Preoperative Preparation
Oral midazolam is administered, if deemed necessary. Intubation is performed while the patient is under deep serourane anesthesia; sometimes intubation is facilitated by
short-lasting muscle relaxants (e.g., atracurium or vecuronium). Anesthesia is induced with serourane, and nitrous
oxide and is maintained with fentanyl (10 g/kg), 2% serourane, and 70% nitrous oxide. Propofol is avoided because it
alters EMG activities. The patient receives a dose of antibiotic before a skin incision. A bladder catheter is inserted.

Operative Procedure
The patient is placed in a prone position on the operating
table so that cerebrospinal uid (CSF) is pooled rostrally and
CSF loss from the intracranial compartment is minimized
(Fig. 271). Needle electrodes are placed bilaterally in the
adductor longus, vastus lateralis, anterior tibialis, medial
hamstring, and medial gastrocnemius muscles in preparation for intraoperative EMG examinations.

Localization of the Conus Medullaris and a SingleLevel Laminectomy

Normally, the conus medullaris terminates between the
T-12 and L-3 spinal levels. For a laminectomy to be limited
to a single level, the conus medullaris must be localized

Figure 271 After electromyographic electrodes are inserted,
the patient is placed prone in the Trendelenburgs position to
minimize cerebrospinal uid loss during the operation.

14535_C27.indd 178

4/11/08 11:34:15 AM

27 Selective Dorsal Rhizotomy for Spastic Cerebral Palsy 179

in several steps with ultrasound before a laminectomy is
In children younger than 10 years, the conus medullaris
and cauda equina are localized with axial views of the spinal
cord ultrasound through the skin and paraspinal muscles.
An ultrasound probe is placed lateral to the spinous process
to get axial views of intradural structures at a few levels
(Fig. 272A). On axial views, the conus appears hypoechogenic and circular. The ventral and dorsal roots are attached
to the lateral aspects of the conus. Pulsatile movements of
the conus are always present. The cauda equina appears
as a hyperechogenic mass. It also has pulsatile movements
(Fig. 272B).
In children younger than 10 years, the skin incision is made
over the interlaminar space where the conus has been localized with ultrasound. The paraspinal muscles are injected
bilaterally with saline solution that contains epinephrine
in a concentration of 1:400,000. The interlaminar space is
exposed, and the interspinous ligament and ligamentum
avum are removed. The conus is localized with axial and
sagittal views of ultrasound. If, with ultrasound, the spinal
cord is rostral to the conus or the cauda equina is caudal
to the conus, then the skin incision is extended rostrally
or caudally. The next interlaminar space is exposed. If the
conus is clearly localized with axial and sagittal views of ultrasound, the next caudal interlaminar space is exposed. The
cauda equina is localized with ultrasound (Fig. 272B).
In older children (older than 10 years) and adults, ultrasound localization of the conus is impossible. Thus the
spinous process of the L-1 vertebra is localized with a lateral
radiograph of the lumbosacral spine and marked by ~0.2
ml of indigo carmine. First, the T12-L1 interlaminar space
is exposed. The ligamentum avum is removed to expose

the extradural fat tissue. Ultrasound examination of intradural structure is done through the interlaminar space (Fig.
272A). If the interlaminar space is tight for the ultrasound
examination, it is enlarged with a Kerrison punch. Two levels of interlaminar space are examined to localize the conus
and cauda equina.
If the conus and cauda equina are localized, a single-level
laminectomy is performed with a Midas Rex craniotome
with a B5 attachment (Medtronic Powered Surgical Solutions, Fort Worth, Texas) (Fig. 273A). After the lamina is
removed, ultrasound examination of the laminectomy site is
obtained again to conrm that the conus and cauda equina
are at the laminectomy site. On the ultrasound examination,
the conus is distinguished from the cauda equina as follows
(Fig. 272B): a sagittal examination reveals the conus as a
hypodense triangle tapering caudally. The ventral and dorsal
spinal roots appear hyperdense. When only a caudal end of
the conus is in the laminectomy area, sagittal examination
does not clearly delineate the conus, and an axial examination is required. A hypoechogenic circular structure on
axial view is sought at the center of the dural tube; it is
most reliable in localizing the conus. Also, on axial view,
one can notice a small cleft between the dorsal and ventral
spinal roots on the lateral aspect of the conus. This cleft is an
important anatomical landmark because it guides the surgeon in separating the dorsal roots from the ventral roots.
Sometimes the patent central canal appears hyperechogenic
within the conus.
Even in 2-year-old children, only a single-level laminectomy is needed for SDR. The laminectomy should cover at
least 5 mm of the conus so that the dorsal roots are safely
separated from the ventral roots at a later stage of operation. If it is needed for the adequate exposure of the conus,

Figure 272 (A) Through an L1-2 interlaminar space or a keyhole
laminotomy, the conus medullaris is identied with ultrasound. (B) The
conus appears hypoechogenic (arrowhead) in contrast to the hyperecho-

genic cauda equina. If the conus is not identiable through the interlaminar space, then an ultrasound examination is repeated through the
L1-2 interlaminar space.


14535_C27.indd 179

4/11/08 11:34:19 AM

180 Pediatric Neurosurgery

Figure 273 (A) After the conus is clearly identied, a single laminectomy is done entirely with a Midas Rex craniotome. At least 5 mm of
the caudal conus should be exposed. The laminectomy extends laterally
close to the facet joint. (B) After the dural incision, an operating microscope is brought into the eld. The L-1 and L-2 spinal roots are identied
at the corresponding intervertebral foramina, and the lum terminale in
the midline is found. (C) The L-2 dorsal root and the dorsal roots medial
to the L-2 root are retracted medially to separate the L2-S2 dorsal roots

from the ventral roots. The thin S3-5 spinal roots exiting from the conus
are identied. A cotton patty is placed over the ventral roots and lower
sacral roots. (D) A 5 mm Silastic sheet is placed under the L2-S2 dorsal
roots, after which the surgeon again inspects the L-2 dorsal root at the
foraminal exit, the lateral surface of the conus between the dorsal and
ventral roots, and the lower sacral roots near the lum terminale. The
inspection ensures placement of only the L2-S2 dorsal roots on top of
the Silastic sheet.


one-third of the lamina that is immediately rostral to the

laminectomy is removed. A wide laminectomy is necessary
to expose the spinal nerve roots for EMG testing.
Other neurosurgeons perform SDR through L2-S1 or L1-S1
laminectomy or laminoplasty. The multilevel laminectomy
makes SDR easy. Nevertheless, the multilevel laminectomy

14535_C27.indd 180

or laminoplasty, when performed in children with spastic

diplegia or quadriplegia, can lead to late spine deformities,
including spondylolisthesis and increased lumbar lordosis
or scoliosis. The multilevel laminectomy may carry a higher
risk of spine deformities after SDR in older children and
adults than in young children.

4/11/08 11:34:30 AM

27 Selective Dorsal Rhizotomy for Spastic Cerebral Palsy 181

Separation of Dorsal Roots from Ventral Roots
After bleeding from the epidural veins and bone is controlled, a dural incision is made. Saline irrigation is not used
after the dura is opened because it alters EMG responses. An
operating microscope is then brought into the eld and used
during EMG testing and sectioning of dorsal root fascicles.
The operating table may be slightly rotated away from the
surgeon as the contralateral spinal roots are dissected. The
arachnoid is removed, and the conus and lum terminale
are identied. At this point, EMG activities are continuously
monitored to determine if any movement of the nerve roots
evokes EMG activities. Stretching and pressure on the ventral roots but not on the dorsal roots evoke EMG activities
and often movement of the patients lower extremity.
Next, the L-2 spinal roots are identied at the neural foramen, and the L-2 dorsal root is separated from the ventral
root (Fig. 273B). The L-2 ventral and dorsal roots are traced
back to the conus until the cleft between the ventral and
dorsal roots is identied. Then the L-2 and adjacent dorsal
roots are gently retracted medially, and a cotton patty is
placed over the ventral roots (Fig. 273C). The L-1 root is
left untouched at this point. Next, the conus and the lum
terminale are examined, and the S2-5 sacral roots that exit
the conus are identied. The S-2 dorsal root can be bulky,
especially in patients with the postxed lumbosacral plexus,
but there is always an abrupt and marked decrease in size
of the S-2 root. The individual S3-5 spinal roots appear as
thin threads. The dorsal and ventral roots at this level are
close together without intervening space between them, so
all of the S3-5 spinal roots are left intact. The lower sacral
roots can best be identied with a gentle lift at the dorsal
roots from the entry zone on the dorsal aspect of the conus.
Whenever the surgeon is unsure of the exact identication
of the S3-5 spinal roots, then sparing the S-2 dorsal root
would be prudent.
Once the L2-S2 dorsal roots are identied, a 5 mm wide
blue Silastic sheet (Dow Corning, Midland, Michigan) is
placed around all of the dorsal roots and distant from the
conus (Fig. 273D); the Silastic sheet keeps the L2-S2 dorsal
roots safely separate from the ventral and lower sacral roots
during the rest of operation. Before starting EMG testing,
the surgeon reexamines three structures to ensure that no
ventral root or lower sacral root is over the Silastic material:
the L-2 foraminal exit, the cleft lateral to the conus between
the ventral and dorsal roots, and the S3-5 roots.

Identication of Individual Dorsal Roots

A shortcoming of this technique, as compared with alternative techniques, is difculty in the identication of individual dorsal roots with certainty. Precise identication of
the roots, however, is not critical for SDR because all major
lower extremity muscles of children with spastic CP receive
motor innervation from several segments. As is shown in
animal experimental studies, signicant somatotopic organization and sprouting may occur in the spinal cord and
brain after deafferentation.
The L-2 dorsal root is readily identied at the neural foramen. The L3-S2 dorsal roots below the conus are close

14535_C27.indd 181

together without a natural separation, so unequivocal identication of the individual dorsal root is difcult. Nevertheless, dorsal root bers of individual segments are roughly
identied as follows. First, the dorsal roots are spread on top
of the Silastic sheet. The L-3 and L-4 dorsal roots, which are
located medial to the L-2 root, are identied; each of the roots
consists of two and three naturally separated rootlets. The
L-5 and S-1 roots are medial to the L-4 root and largest of all
the lumbosacral roots. The L-5 and S-1 dorsal roots consist of
three or four rootlets with natural separation. The S-2 root
has a single fascicle. Second, an innervation pattern of each
root is examined with EMG testing. An individual dorsal root
is placed over two hooks of the Peacock rhizotomy probes
(Aesculap Instrument Co., Burlingame, California) (Fig. 274A),
and responses to electrical stimulation with a threshold voltage are recorded from the lower extremity muscles. The entire dorsal root is tested at each level immediately before
subdividing the dorsal root into rootlets.

EMG Examination and Sectioning of Dorsal Roots

After the innervation of a dorsal root is determined, the root is
sharply subdivided into three to ve smaller rootlet fascicles of
equal size with a Scheer needle (Storz Instruments, St. Louis,
MO) (Fig. 274B). The rootlet fascicles are suspended over two
hooks of the rhizotomy probes (Fig. 274C). Single constant
square wave pulses of 0.1 milliseconds duration are applied to
the rootlet at a rate of 0.5 Hz. The stimulus intensity is increased
stepwise until a reex response appears from the ipsilateral
muscles. After the reex threshold is determined, a 50 Hz train
of tetanic stimulation is applied to the rootlet for 1 second. The
reex response is then graded according to the criteria detailed
in Table 272. Our experience has been that most rootlets produce 1+ to 4+ responses. Thus we base our decision to section
a given rootlet on the number of rootlets producing sustained
responses at that level and the intensity of the responses. The
rootlets that produce a response of 0 are left intact. The rootlets
producing 3+ and 4+ responses are cut, and those producing
1+ and 2+ responses are sometimes spared. The dorsal rootlets
spared from sectioning are placed behind the Silastic sheet and
kept separated from rootlets yet to be tested (Fig. 274D). If only
1+ and 2+ responses are detected, then rootlets with the most
active responses are cut (Fig. 274E). At least one rootlet is left
irrespective of EMG responses to avoid postoperative sensory
loss. The procedure is performed in sequence on the remaining
L3-S2 dorsal roots (Fig. 274F).
Using the criteria given in Table 272, we section 60 to 65%
of the rootlets examined. Finally, the L-1 dorsal root is identied
at the neural foramen, and half of the dorsal root is cut without
EMG testing. In our experience EMG testing of the L-1 root is unreliable. The sectioning of the L-1 dorsal root is necessary to further reduce spasticity in hip exors, especially in patients with a
large L-1 root associated with prexed lumbosacral plexus. The
rhizotomy is repeated on the contralateral side.
The intradural space is irrigated with saline solution.
Bipolar cautery is seldom required for control of bleeding
from the cut ends of fascicles. The dura is closed in a running
fashion with 40 monolament nylon. Clonidine (2 g/kg
up to 7 years of age and 1 g/kg over 8 years of age) mixed
with morphine at 15 g/kg of body weight is injected intra-

4/11/08 11:34:50 AM

182 Pediatric Neurosurgery


Figure 274 (A) The L-2 dorsal root is easily identied. In an attempt to
identify the L3-S2 dorsal roots, all the dorsal roots are spread over the
Silastic sheet and grouped into presumed individual dorsal roots. Then
the innervation pattern of each dorsal root is examined with electromyographic (EMG) responses to electrical stimulation with a threshold voltage. (B) With a Scheer needle, each dorsal root is subdivided into three
to ve rootlet fascicles, which are subjected to EMG testing. (C) Stimula-

tion of an L-2 rootlet fascicle elicits an unsustained discharge to a train of

tetanic stimuli. (D) The rootlet is thus spared from sectioning and placed
behind the Silastic sheet. (E) Stimulation of a rootlet fascicle elicits sustained discharges from multiple muscles. The rootlet is thus sectioned. (F)
The rootlets spared from sectioning are under the Silastic sheet, and the
roots to be tested are on top of the Silastic sheet. Note that EMG testing
and sectioning of the dorsal roots are performed caudal to the conus.


14535_C27.indd 182

4/11/08 11:34:50 AM

27 Selective Dorsal Rhizotomy for Spastic Cerebral Palsy 183

Table 272 Criteria for Grading Electromyographic
Responses in Selective Dorsal Rhizotomy for Spastic
Cerebral Palsy

Electromyographic Response


Unsustained or single discharge to a train of stimuli

Sustained discharges from muscles innervated
through the segment stimulated in the ipsilateral
lower extremity
Sustained discharges from muscles innervated
through the segment stimulated and immediately
adjacent segments
Sustained discharges from segmentally innervated
muscles and muscles innervated through segments
distant to the segment stimulated
Sustained discharges from contralateral muscles
with or without sustained discharges from the
ipsilateral muscles




durally. The Trendelenburgs position is reversed. A strip of
Gelfoam is left over the laminectomy defect, and the wound
is closed in layers.

Postoperative Management Including

Possible Complications

Patients are transferred to the ward the next day, and the
fentanyl drip is continued for another 24 to 48 hours. On
the third postoperative day, patients are allowed to sit, and
physical therapy is started. The patients are discharged to
home on the fth postoperative day and receive outpatient
physical therapy from local therapists.

Postoperative Course and Complications

With spastic diplegia the operation invariably reduces spasticity, whereas with spastic quadriplegia there is a small
chance of recurrent spasticity. Most patients who were
independent walkers preoperatively resume independent
walking within 2 weeks after undergoing the SDR technique
previously described. Patients who walked with assistance
preoperatively take a slower postoperative course. Within
2 months, however, all patients show motor performance
exceeding preoperative levels.
The major complications of SDR include paraplegia, sensory loss, bladder and bowel incontinence, CSF leak, and
infection. There has been one case of CSF leak requiring
operative repair in more than 1500 children and adults who
have undergone single-laminectomy SDR at our institution,
clear evidence of the safety of the procedures. Also, no patients had late spine deformities that required medical or
surgical intervention. Many patients, however, did experience hyperesthesia in the legs for several months.

Postoperative Care
Patients stay overnight in the intensive care unit where they
receive an intravenous infusion of fentanyl, at a dose of 1 to
3 g/hour per kg of body weight, and diazepam, as needed.

14535_C27.indd 183

4/11/08 11:35:30 AM

Treatment of Lambdoidal Synostosis with
Calvarial Reconstructive Techniques
David F. Jimenez, Constance M. Barone, and James Tait Goodrich
Although premature closure of the lambdoid suture is the
least common of all craniosynostosis, it produces a marked
posterior deformity characterized by ipsilateral flattening
of the occiput, skull base changes leading to cranial scoliosis, and inferior displacement of the temporal-mastoid
region (Fig. 281, Fig. 282, and Fig. 283). These changes
may be mild or severe depending on the time of onset
of the synostosis, (in utero vs. early or late infancy) and
extent of suture closure. Care must be taken not to confuse true lambdoidal craniosynostosis with deformational
plagiocephaly, which may be due to torticollis and/or positional sleeping patterns. True lambdoidal synostosis will

give the cranium a trapezoidal appearance, when viewed

from above, versus a parallelogram shape in the case of
deformational plagiocephaly.

Figure 281 Three-month-old male with left lambdoidal synostosis.

Prominent ndings include ridging of the affected suture, inferolateral
displacement of the ipsilateral posterior fossa, asymmetric skull base,
and inferior displacement of temporomandibular complex.

Figure 282 Minimal asymmetry of the forehead is seen in lambdoid

craniosynostosis when compared with deformational plagiocephaly.
Sagittal calvarial scoliosis is evident with deviation of the parietal area
toward the involved suture can be seen with stenoidal involvement.



14535_C28.indd 184


Patient Selection
Patients presenting with severe deformation of the occiput,
skull base, and calvaria who are older than 9 months of
age are ideal candidates for extensive calvarial reconstructive techniques. Although every attempt should be made
to diagnose and treat these infants at an earlier time, often referring pediatricians and primary care providers will
watch and follow these patients expectantly hoping for selfcorrection. Parents should be advised that these are risky

4/11/08 11:28:27 AM

28 Treatment of Lambdoidal Synostosis with Calvarial Reconstructive Techniques 185

Preoperative Preparation

Figure 283 Patient with left lambdoid synostosis shows no evidence

of vertical dystopia (vs. contralateral coronal synostosis); inferior displacement of ipsilateral mastoid complex and contralateral parietal

procedures, and the surgical and anesthesia teams must be

ready to deal with unexpected complications such as injury
to the conuence of sinuses, or transverse or sigmoid sinus
tears. Whenever possible, family should be encouraged to
have donor-directed blood available for possible intra- and
postoperative transfusions.

As these children are generally healthy, no specic preoperative preparations are needed besides baseline hematological studies. Generally, the anesthesia team requires
only normal electrolytes and hematocrit levels. The patient
should be given a single dose of anti-staphylococcal antibiotic (oxacillin 50 mg/kg intravenously) 1 hour prior to
surgical incision. A Foley catheter is inserted to follow the
intravascular volume status during the operative procedure.
Many anesthesia teams prefer to place intra-arterial and
central venous lines to closely monitor the patients during
the surgery.
To allow access for both plastic surgery and neurosurgical teams, the patient is placed prone, in the center of the
room on an adequately padded horseshoe headholder. The
patients head is placed at 180 degrees from the anesthesia
team, and the scrub nursing team can be located to the patients left. A Mayo stand is placed over the patients back
and holds the active surgical instruments. A second larger
table is kept behind the main tray set-up of both surgical
teams. A third sterile table, equipped with a complete drill
and cranial xation systems, is provided for the plastic surgery team. All cautery cords, suction tubes, and other lines
are placed toward the feet of the patient so that there is no
clutter on the oor within the operating circle. The patient
is placed in the prone position with the forehead and face
resting on the horseshoe-shaped headrest covered with a
viscoelastic polymer pad. The eyes must be checked to ensure that no pressure is being applied by the headrest. The
head is placed in a position level to the heart to reduce the
risk of air embolism. Shaving is performed only in an area
around the proposed incision line for a width of 2 cm (Fig.
284). In children with thin hair, no shaving is required. The
entire occipitoparietal area, including the ears, is scrubbed
with Betadine (providone-iodine) soap followed by Betadine solution and paint.

Figure 284 Three-month-old female placed
prone in padded horseshoe. A small area is
shaved extending biparietally. The scalp ap is
retracted posteriorly to expose the occiput for
bilateral reconstruction.

14535_C28.indd 185

4/11/08 11:28:28 AM

186 Pediatric Neurosurgery

Operative Procedure
Skin Incision
If desired, scalp hemostasis is enhanced prior to skin incision by inltrating the dermis with 0.25% lidocaine and
epinephrine (1:400,000). If not, a relatively bloodless scalp
opening can be made using the monopolar electrocautery
unit with a ne needle tip set at 15 W with a blend of 80%
cut and 20% coagulation. The incision is made to extend from
ear to ear and is carried just behind the ear helix. The scalp
ap is elevated in the subgaleal plane and carried posteriorly to allow full exposure of the occiput. The pericranium
is elevated as a separate layer and taken down to the nuchal
musculature, which is elevated with the pericranium using
the monopolar needle tip electrocautery. Once elevated, the
pericranium and muscles are covered with a moist sponge
to keep them from dehydrating and shrinking.

Correction of a severe lambdoidal malformation requires
complete disassembly of the parieto-occipital region. Our
plastic surgery team marks out the occipital site with methylene blue. First, a new bandeau is marked out (Fig. 285)
that will provide an anchoring unit upon which reconstruction is based. A new backhead has to be identied, for

which we use a Marchac template to nd an area of calvaria

that is symmetrically close to normal. Depending on how far
forward the symmetric reconstruction is carried, we mark
out an osteotomy that is taken from each asterion up to just
behind the anterior fontanelle. A step-off is marked out at
the asterion; this will provide an anchoring point to which
the new bandeau is placed in a tongue-and-groove fashion
(Fig. 286). Another osteotomy is then carried around as
long as possible under the inion pointing to the base of the
These osteotomies result in three units of harvested bone:
(1) the new bandeau; (2) the new backhead; and (3) the
remaining occipital unit. This third piece is then cut up into
units of bone to be replaced in a mosaic fashion. Another
technique is to cut the occipital bone into strips, which
then can be radially placed to give a sunrise appearance
to the new occipital complex (Figs. 287 and 288). The
bur holes are made with a round drill bit (MedtronicMidas Rex M33; Medtronic Inc., Ft. Worth, Texas) and a 0.5 cm
bur hole is placed on either side just behind the asterion.
Once the dura is cleared, a footplate attachment (Midas
Rex B5) is used and a bone flap elevated. Elevation of the
bone flap may be accomplished in several ways. The footplate attachment is used to create an osteotomy from the
asterion to midline. If difficulties are encountered near the
sagittal suture, then a second osteotomy is begun at the contralateral bur hole and united with the previous osteotomy.
This can be done in sequential fashion until three to five

Figure 285 Artistic reconstruction of the bandeau and
backhead and sites from which these are harvested.

14535_C28.indd 186

4/11/08 11:28:31 AM

28 Treatment of Lambdoidal Synostosis with Calvarial Reconstructive Techniques 187

Figure 286 Artists rendition of removal of posterior occipital bone plate. The reconstructed tiara, xed with miniplates, is seen in the lower

segments are cut. The area where the backhead was elevated is used as an anatomical opening to dissect various
sutures, whereas typically the dura is very adherent. If the
dura is extremely adherent or the team feels the sinus may
be injured, an additional osteotomy is made just lateral to
the sagittal suture and carried parallel to and down over
the occipital bone to the lower osteotomy. This cut allows
easy exposure of the sagittal sinus along its length. Extra
care must be taken with all of the major dural sinuses and
their points of confluency. This is especially the case at the
asterion. If the sinuses are interrupted here, the bleeding
can be profuse and life-threatening. The asterion suture
is often involved in lambdoidal synostosis and must be
released, which can be done easily with careful dissection
using a no. 1 Penfield dissector. This osseous area also has
to be released, as a tongue-and-groove bandeau will be
anchored here.
The bone ap is now handed off to the plastic surgery
team to complete the reconstruction at a separate table.

14535_C28.indd 187

The tiara (backhead and bandeau unit) is made using the

three anchor miniplates, and this is the foundation on which
the reconstruction will be based. The tiara unit is placed
tongue-and-groove into the asterion region. A critical unit
is the strut piece that is placed from the region of the anterior fontanelle to the most superior point of the tiara (Fig.
288). The rest of the calvaria is then reconstructed with the
remaining bone pieces, taking care to maintain symmetry.
We use the combination of two Nurolon or three Nurolon
vicarial and titanium miniplates and screws to anchor the
bone pieces. Once all bone aps are reattached, the cylinder
drill bit (Midas Rex M12) can be used to smooth all edges to
give the new construct a rounded and symmetric contour.
The pericranial ap is then reected over the bone aps and
secured in place to the adjacent pericranium with absorbable sutures. The galea and skin are closed in a standard
fashion. A 7 mm drain is left in the subgaleal space and
tunneled to a separate exit. We do not routinely suture the
drain in place.

4/11/08 11:28:34 AM

188 Pediatric Neurosurgery

Figure 287 Artistic reconstruction showing the tiara
placed tongue and groove into the asterion. A strut piece
is placed from the anterior fontanelle to the top of the
tiara. The rest of the calvaria is reconstructed in a mosaic

Figure 288 (A) Lateral intraoperative photographs depicting the reconstructed occipital unit. (B) The tiara forms the new occiput along with
radially placed strips of bone, resembling a sunrise.

14535_C28.indd 188

4/11/08 11:28:35 AM

28 Treatment of Lambdoidal Synostosis with Calvarial Reconstructive Techniques 189

Postoperative Management Including

Possible Complications
All patients are placed in pediatric intensive care unit for at
least 24 to 48 hours postoperatively. The drain is removed 24
hours postoperatively unless there is still signicant drainage. Prophylactic antibiotics are not routinely used following surgery. The majority of the patients will have elevated
temperatures (3839C) during the rst 4 to 5 days postoperatively. We no longer culture patients on a routine basis
unless there is evidence of an infection or the fever persists
for more than 4 days. The procedure is generally well tolerated, but the patients usually require a strong analgesic regiment. Patients are kept prone or rotated from side to side
to prevent reoccurring occipital asymmetry due to pressure
from continuously lying on one side. Patients are usually
discharged on the sixth to seventh postoperative day.

The most serious complication is injury to the asterion region
venous sinuses during elevation of one of the bone aps. In
most cases, an Avitene pledget (C.R. Bard, Inc., Murray Hill,
New Jersey) and direct pressure to the bleeding area with
a combination of Gelfoam, cottonoids, and, most recently,
Surgio (Ethicon, Inc.; Somerville, New Jersey), will stop the
majority of bleeding. Occasionally the sinus tear will need to
be closed primarily with a small suture. Meticulous atten-

14535_C28.indd 189

tion has to be paid to any dural tears, which must be directly

repaired. Any subsequent cerebrospinal uid leak will lead
to persistent subgaleal uid collection and problems with
proper bone healing. Prior to elevation of the skin ap, all of
the bone, ap gutters, and surrounding areas are copiously
irrigated. The removal of the remaining debris and collected
bone dust markedly reduces postoperative infection. Subdural
hematomas are a possible complication but are extremely
rare. If they do occur, they most likely are due to improper repair of a venous sinus injury. Persistent subgaleal uids in the
postoperative period need to be tapped and tested for bacteria, if the child has a persistent fever. Results indicate that
excellent long-term outcomes in older patients treated with
these complex reconstructed techniques can be achieved if
these surgical principles are closely followed.
This chapter is a revision of the chapter, Lambdoidal Synostosis by David F. Jimenez, M.D., Constance M. Barone,
M.D., Ravelo Argamaso, M.D., and James T. Goodrich, M.D.
The chapter appeared in the Neurosurgical Operative Atlas,
Volume # 4, edited by Setti S. Rengachary and Robert H.
Wilkins. The Neurosurgical Operative Atlas was published by
the American Association of Neurological Surgeons (AANS)
from 1991 to 2000.
We would like to acknowledge and thank Ravelo Argamaso, M.D., for his help and efforts on the original chapter
published in the rst edition of this work.

4/11/08 11:28:38 AM

Early Treatment of Lambdoid Synostosis
with Endoscopic Assisted Craniectomy
David F. Jimenez and Constance M. Barone

Patient Selection
When a patient presents with cranial vault deformational
changes secondary to lambdoid synostosis at <6 months
of age, our preferred method of treatment is to release the
synostosed suture by performing a suturectomy and then
placing the patient in a cranial orthosis postoperatively. The
concept is based on utilization of the rapid brain growth
that takes place during the rst 18 months of life to correct the deformational skull changes associated with the
synostosis. Invariably, the younger the patient at the time
of the release, the better the results will be. Very young
infants tolerate this surgery well, with minimal blood loss
and without the need for blood transfusions.

Preoperative Preparation
No special preparations are needed if the patient is found
to have no signicant medical problems or past medical
history. A spun hematocrit is drawn following anesthetic
induction, and no other serum laboratories are obtained. A
single dose of anti-staphylococcal antibiotic (oxacillin 50
mg/kg) is given within 1 hour prior to incision. To help with
postoperative pain management, an acetaminophen rectal
suppository (10 mg/kg) is given prior to the start of the
case. The scalp is minimally shaved and then prepped with
Betadine (providone-iodine) paint, which is allowed to dry
for maximal effectiveness.

Figure 291 Patient lies on a padded horseshoe headholder with the

head parallel to the oor. This position allows access to the lambda
and asterion regions of the affected suture.

Operative Procedure
The patient is induced under general anesthesia and placed
supine with the head turned and parallel to the oor (Fig.
291). A shoulder roll is used to elevate the ipsilateral
shoulder and to minimize cervical rotation, and one or two
peripheral intravenous lines are inserted. There is no need
for placement of arterial or central venous lines or urinary
catheter. It is of paramount importance to radiographically
identify the ipsilateral lambda and the asterion. This can be
done using plain x-rays and metallic markers (Fig. 292).
The extent of synostosis is correlated between the preoperative computed tomography (CT) scan and the intraoperative
x-rays. The incisions are marked to include the full extent of
the synostosed suture (Fig. 293).

Figure 292 Markers (Xspot) are used intraoperatively to localize

the medial and lateral ends of the stenosed lambdoid suture. A plain
x-ray allows for proper placement of the incisions.



14535_C29.indd 190

4/11/08 11:22:50 AM


Early Treatment of Lambdoid Synostosis with Endoscopic Assisted Craniectomy 191

Figure 293 Diagram showing the location of the medial and lateral
incisions. These must be made directly over the stenosed suture.

Operative Technique
The incisions are placed lateral to the lambda along the stenosed lambdoid suture and medial to the ipsilateral asterion.
Each incision is ~2 cm in length. The incision is made with
the needle tip monopolar electrocautery set at 15 W, blend
1. Dissection is taken down below the galea but does not
include the pericranium. Insulated malleable retractors are
used to elevate the scalp and a zero degree rigid endoscope
is used to perform a subgaleal dissection, with the monopo-

Figure 294 Dissection is undertaken under the galea and directed

laterally toward the asterion. A needle tip electrocautery is used to
develop a bloodless plane above the pericranium ~3 cm in width.

Figure 295 A 4 mm Kerrison rongeur is used to enlarge the bur hole.

The endoscope and sectional dissector are inserted under the bone to
separate the dura from the galea.

lar laterally toward the asterion. Care is taken not to elevate

the pericranium, as this maneuver will create unnecessary
bleeding. A second incision is made medial to the asterion,
and the subgaleal dissection is extended medially to join the
previously described exposure (Fig. 294).
A 7 mm (pediatric) craniotome is used to create a bur hole
at each incision. The bur holes are enlarged longitudinally
using a 4 mm Kerrison rongeur (Fig. 295). This maneuver
allows the insertion of the rigid endoscope under the calvarial bone (Fig. 296). A no. 6 or 7 French malleable suction
tube is inserted under the bone and advanced in front of
the endoscope to visualize the osseous-dural interface. The
endoscope and suction tips are advanced in tandem, visualizing the stenosed suture and the underlying dura. This dissection may be carried medially from either incision. Once
the entire stenosed suture has been fully isolated from the
overlying scalp and underlying dura, preparations are made
for resection of the stenosed suture.
The suturectomy is begun at the lambdoid incision and
extended toward the asterion by cutting a small (23 cm)
wedge of bone along the affected suture (Fig. 297). Further
wedge sections can be removed until the asterion incision is
reached. Care must be taken not to tear the dura, particularly
near the asterion, as it can lead to an injury to the transverse
or sigmoid sinus. Once the suturectomy is nished, the edges
are smoothed with either the Mayo scissors or a small Leksell
rongeur. Diploic homeostasis is obtained using a suction electrocautery unit (Valley Laboratory, Valley Forge, PA) set at 50
to 60 W. Coated malleable retractors must be used to protect
the dura and scalp from being thermally injured.
The wound is irrigated, further hemostasis is obtained as
needed, and then the scalp incisions are closed using 40
Vicryl (Ethicon; Somerville, NJ) galeal sutures. Dermabond


14535_C29.indd 191

4/11/08 11:22:54 AM

192 Pediatric Neurosurgery

Figure 296 Artists diagram shows a 30 rigid endoscope being advanced under the stenosed (and ridged) suture and above the dura. A small
malleable suction tip is used to dissect and keep the endoscopic eld bloodless.

(Ethicon; Somerville, NJ) is used for nal dermal closure.

The scalp is inltrated with 0.25% marcaine (without epinephrine) at 1 ml/kg to help with postoperative pain. Following extubation, the patient is admitted overnight for observation and discharged the following morning. Any residual
postoperative swelling is allowed to resolve over the next 2
to 3 days, and then the patient is scanned and properly tted
for a helmet, which is to be worn to help correct the associated cranial deformational changes (Fig. 298).

Postoperative Management Including

Possible Complications
All patients are hospitalized overnight and discharged the following morning. Pain management is obtained by alternating
use of acetaminophen (15 mg/kg every 3 hours) and ibuprofen (10 mg/kg weight every 3 hours). If stronger analgesic is

Figure 297 Diagram showing extent of the craniectomy, which

should be 0.7 mm 0.2 mm in width. Care must be taken that the
craniectomy extends from lambda to asterion.

14535_C29.indd 192

needed, Nubain or morphine sulfate is used. The patients

irritability is most evident during the rst 8 hours following surgery and then fully recedes by the rst postoperative
morning. The patient is allowed to nurse as soon as possible
following the surgery. Two days after surgery, the patient is
scanned and molded for a cranial orthosis, which is worn
during the rst 10 to 12 months (Figs. 299 and 2910).

Possible complications include dural tears, dural sinus injury,
embolism, and any of the complications associated with the
infant cranial surgery. We have had no complications and
overall are very satised with the results. None of the patients
have required blood transfusions or further reoperations.

Figure 298 A custom-made Surlyn helmet (Dupont, Wilmington,

Delaware) is used postoperative to direct brain growth and achieve
calvarial correction.

4/11/08 11:22:55 AM


Early Treatment of Lambdoid Synostosis with Endoscopic Assisted Craniectomy 193

Figure 299 Preoperative view of a 5-month-old with lambdoidal synostosis and marked skull base asymmetry and cranial scoliosis.

14535_C29.indd 193

Figure 2910 Postoperative view at 2 years of age showing signicant

correction. Estimated blood loss was 15 cc, surgery time was 40 minutes, and helmet was worn for 7 months.

4/11/08 11:22:59 AM

Posterior Plagiocephaly
Richard G. Ellenbogen, Sudesh J. Ebenezer, and Richard Hopper

Patient Selection
Diagnostic Criteria
Posterior plagiocephaly is an abnormality of the posterior aspect of the skull where the ipsilateral occipital and
parietal bones are flattened. Posterior plagiocephaly may
result from fusion of a lambdoid suture. It can also occur
in the absence of lambdoid suture synostosis. A child with
posterior plagiocephaly, regardless of the etiology, will
have flattening of the affected occipital-parietal region.
In general, posterior plagiocephaly resulting from lambdoid suture synostosis can require surgical management.
Deformational or nonsynostotic posterior plagiocephaly
secondary to external compressive forces whether in utero,
at delivery, or after birth generally does not require surgical treatment. Thus it is of paramount importance to distinguish the cause of the posterior plagiocephaly, as the
treatment paradigms are different. Table 301 describes
the morphology of unilambdoid synostosis versus nonsynostotic posterior plagiocephaly.
Patients are seen in multiple visits in our craniofacial
clinic. A thorough history and physical examination is important at each visit. The features in Table 301 are assessed at each visit, to determine whether the patient has
positional posterior plagiocephaly or synostotic posterior

plagiocephaly. Fig. 301, Fig. 302, and Fig. 303 show the
differences in morphology.
Photographs are taken at each visit for comparison. Mild
deformities can be observed. A child with a moderate to
severe deformity undergoes a computed tomography (CT)
of the brain along with three-dimensional (3-D) reconstructions. Bony windows are obtained and also reconstructed in
the coronal plain. The underlying brain, skull base, and the
regions undergoing compensatory change should be evaluated closely. Fig. 304 and Fig. 305 show pre- and postoperative 3-D CT scans.

It must be emphasized that isolated true lambdoid synostosis is extremely rare. In a series of 519 children with
craniosynostosis at the Childrens Hospital, Boston reported
by Shillito and Matson in 1968, the incidence of lambdoid
synostosis was 2.3%. In 130 patients with craniosynostosis
reported on by Huang et al, the incidence was 3.1%. A critical review of the literature in 1998 by Rekate revealed the
incidence of lambdoid craniosynostosis ranges from 3 to
20%. Differences in diagnostic criteria accounted for the
variability. We believe that the true incidence is closer to
that reported by Shillito and Matson, 2 to 3%. In contrast
to posterior plagiocephaly caused by lambdoid suture syn-

Table 301 Features and Diagnosis of Posterior

Posterior Plagiocephaly without Synostosis

Unilambdoid Synostosis

View from vertex

Skull base

1. Open
2. No ridging
Perpendicular to suture

Ipsilateral inferior tilt
1. Fusion
2. External ridging
1. Parallel to suture
2. Restricted perpendicular to suture
Inferiorly or posteriorly
1. Ipsilateral occipital and mastoid
2. Contralateral occipital and parietal
Ipsilateral constriction

Skull growth
Displacement of ipsilateral ear
Volume of posterior fossa
Ipsilateral occipital and parietal bones

1. Contralateral occipital
2. Ipsilateral frontal


14535_C30.indd 194

4/11/08 11:28:03 AM

30 Posterior Plagiocephaly 195

Figure 301 (A) Normal head shape from posterior view. Growth occurs perpendicular to suture. Skull base is horizontal. (B) Normal head shape
from vertex view.

Figure 302 Posterior view ([A] diagram and [B] photograph) of synostotic posterior plagiocephaly involving left lambdoid suture. Compensatory growth is parallel to fused lambdoid suture. Ipsilateral occipital

mastoid and contralateral parietal bossing results. Inferior displacement

of ipsilateral ear and ipsilateral inferior tilt of skull base occurs.


14535_C30.indd 195

4/11/08 11:28:03 AM

196 Pediatric Neurosurgery

Figure 303 Differences in head shape from vertex view. (A) Positional molding showing parallelogram-shaped head. (B) Right lambdoid synostosis
showing trapezoid-shaped head.

Figure 304 (A) Preoperative three-dimensional (3-D) computed tomography (CT) of right lambdoid synostosis. Note fused right lambdoid
suture, ipsilateral skull base tilt, ipsilateral occipital-mastoid bossing, and

contralateral parietal bossing. (B) Immediate postoperative 3-D CT of
right lambdoid synostosis. Note radial barrel-staven-like osteotomies.


14535_C30.indd 196

4/11/08 11:28:06 AM

30 Posterior Plagiocephaly 197

Figure 305 (A) Preoperative three-dimensional (3-D) computed tomography (CT) of left lambdoid synostosis. Note fused left lambdoid
suture, ipsilateral skull base tilt, ipsilateral occipital-mastoid bossing,

and contralateral parietal bossing. (B) Two years postoperative 3-D CT
of left lambdoid synostosis. Note symmetric contour of parietal-occipital area.


ostosis, nonsynostotic plagiocephaly can occur commonly,

and is seen in up to 48% of healthy neonates. Thus it is
imperative that surgeons distinguish the etiology of the
posterior plagiocephaly before contemplating surgical

Surgical Indications
The most common cause of posterior plagiocephaly is a nonsynostotic deformation. This rarely requires surgery. The
other major cause is premature closure of one or both of
the lambdoid sutures. Many infants with lambdoid synostosis have a signicant posterior cosmetic deformity that
progresses. If the progressive nature is conrmed, and is
severe, surgical management is offered. Progression should
be documented over the 6 to 12 weeks that the patient is
followed in the craniofacial clinic prior to surgery. Surgery is
timed for when the child is 3 to 12 months old. Because this
operation represents an expansion of the posterior cranial
surface area, we inform parents to expect gaps in bone coverage, which will heal secondarily with time. As a result of
these bone gaps, surgery should occur within the rst year
of life and ideally at ~6 months of age when bone growth is
still very active from the immature dura. The parents may
believe that the deformity will adversely affect the infants
psychological development. They should understand that
the deformity cannot be covered by hair. Parents must understand that surgery will not cause intellectual impairment
or developmental delay.
If a nonsynostotic deformity (deformational) progresses
even with efforts to change the infants position, the patient
should be reassessed for synostosis with repeat clinical examination and imaging.

14535_C30.indd 197

Contraindications to Surgery
Posterior plagiocephaly caused by a nonsynostotic deformation very rarely requires surgery. Causes of nonsynostotic
deformation include positional molding and torticollis. Positional molding is the most common type of nonsynostotic deformation. This can occur when an infant lies at, or
when one has intrauterine compression of the skull. Positional molding usually causes a mild deformity that generally does not progress. It generally improves as the infant
grows, learns to roll, crawl, and walk. Posterior plagiocephaly caused by torticollis can often be corrected by regular
stretching exercises.

Alternative Considerations to Surgery

Most infants with posterior plagiocephaly will be nonsynostotic (deformational). Thus the majority will not require
surgery. The parents must try to change the infants position
while lying down. There is the risk of sudden infant death
syndrome while in a prone position. The child can be in a
lateral position, or supine with the head toward the unaffected occiput.
Helmet and band therapy are other alternatives. A potential difculty is that it is labor intensive. A dedicated
orthotics team is required. Parents may not be compliant
with helmets.

Preoperative Preparation
Infants have routine laboratory studies. This includes a complete blood count (CBC) and coagulation studies. Parents can

4/11/08 11:28:10 AM

198 Pediatric Neurosurgery

over the proposed incision, which makes for a more efcient
Lacri-Lube (Allergan, Irvine, CA) is applied to the eyes,
which are then taped shut. The endotracheal tube is secured to the mandible with wire or suture for added stability. Three layers of foam are placed over bony prominences
of the cheeks and forehead to avoid pressure on the globes
(Fig. 306). The patient is then turned prone. The head is
placed in a neutral position, face down, in a padded horseshoe headrest. Special care is taken to conrm there is no
ocular compression. The chest and hips are supported by
padded gel cushions. There should be no pressure on the
abdomen. The lower legs are exed 15 degrees using a pillow, to enhance venous return (Fig. 307).

Draping and Skin Incision

Figure 306 Foam placed over bony prominences to avoid pressure

on the globes.

choose parentologous blood donation. Two pediatric units

of blood should be available in the operating suite. Some
centers use a pediatric cell saver.
The anesthesiologist administers one intravenous dose of cefazolin 30 minutes prior to skin incision. No specic anesthetic
technique is used. All patients have a Foley catheter, a radial
arterial line, pulse oximetry, and two intravenous lines. End
tidal partial pressure of carbon dioxide PCO2 is monitored.

A biparietal zig-zag incision is drawn with a marking pen

(Fig. 308). This is done in a coronal plane behind the external auditory meatus, ~2 to 3 cm posterior to the coronal
sutures. This provides maximum exposure down to the insertion of the neck muscles into the posterior fossa, and also
the affected parietal-temporal regions. The eld is prepped
in the standard fashion with alcohol and Betadine (providone-iodine). Draping is completed in a standard manner. We drape from the muscles of the neck to the coronal
suture. This way the back of the head can be completely
visualized. Five to 10 minutes prior to incision, a solution
of 0.25% Marcaine (bupivacaine) with 1:200 000 dilution
of epinephrine is injected into the subcutaneous tissue to
decrease blood loss.

Operative Technique

Operative Procedure
Patient Positioning
We do not shave any hair unless the infant has a signicant
amount of hair. In this case a narrow strip of hair is shaved

The zig-zag incision is made in the subgaleal plane. Bipolar

electrocautery is used to obtain hemostasis. Raney skin clips
are not used to decrease trauma to the tissues and hair follicles. Subperiosteal dissection is done to a level below the
torcula but above the foramen magnum. The neck muscles
inserting into the occipital and posterior fossa regions are

Figure 307 The patient is positioned prone.


14535_C30.indd 198

4/11/08 11:28:11 AM

30 Posterior Plagiocephaly 199

Figure 308 (A) Biparietal zig-zag incision is marked. (B) Biparietal zig-zag incision.

Figure 309 Skin ap retracted.


14535_C30.indd 199

elevated using a sharp periosteal dissector. Inferiorly, the

skin ap should be exposed to the occipitomastoid suture
(Fig. 309).
Bur holes are placed adjacent to the transverse sinus and
on either side of the superior sagittal sinus (Fig. 3010A).
We use a Stryker Universal Drill (Stryker, Kalamazoo, MI)
with a 6 mm round bur. It is set at 60,000 rotations per
minute. A no. 3 Peneld dissector is used to bluntly dissect
the dura from the bone ap. A biparietal-occipital craniotomy is performed. We use a Stryker 1.8 mm craniotome
for this. The craniotomy is taken in two pieces (Fig. 3010B).
If the sagittal suture is excessively adherent to the sinus,
we leave a strip of bone over the sinus, which is later used
during the operation to reattach the two remolded cranial
plates. The craniotomy should extend to the level just above
the transverse sinus and far enough anterior to incorporate the entire attened parietal, occipital, and temporal
The two parietal-occipital bones are reshaped to create a
symmetrical contour using radial osteotomies and a Tessier
bone bender (Lorenz Surgical, Jacksonville, FL). The contralateral bossed parietal bone is often rotated and used
to create the cranial expansion on the affected side. With
this technique, the two bone aps are essentially switched.
While we are reshaping, we often place the bone aps back
over the occipital region and reect the scalp ap to its original position to ensure the reshaped bones are satisfactory.

4/11/08 11:28:14 AM

200 Pediatric Neurosurgery

Figure 3010 Left lambdoid synostosis. (A) Bur holes are placed adjacent to the transverse sinus and on either side of the superior sagittal
sinus. (B) Biparietal-occipital craniotomy taken in two pieces. (C) Bar-

rel-staven-like osteotomies. (D) Bone plates are reattached after being



14535_C30.indd 200

4/11/08 11:28:15 AM

30 Posterior Plagiocephaly 201

Some centers plicate the contralateral dura with either suture or bipolar electrocautery to decrease the compensatory
bossing on the normal occiput. The efcacy of plication is
We then work on the infratentorial occipital and posterior
fossa bones. The dura is stripped from the bone. Vertically
directed barrel-stavenlike osteotomies are then created
along the length of the craniotomy. The cuts are made from
the level of the transverse sinus down to above the level of
the foramen magnum. The osteotomies extend the length
of the craniotomy aps (Fig. 3010C). While the osteotomies are made, the dura and underlying brain is protected
with a malleable brain retractor. The barrel-stavened bone
is remolded and bent posterior using a Tessier bone bender
and mosquito snaps.
The temporoparietal region on the attened side may require osteotomies and remolding to create a symmetrical
contour. The bone plates are reattached using resorbable
microplates, and microscrews (Fig. 3010D). This may require pulling down the scalp ap several times and viewing the remodeled bone work prior to actual xation. Fig.
3011A and Fig. 3011B are intraoperative photographs of
left lambdoid synostosis, shown from posterior.
When there are only mild to moderate anterior fossa compensatory changes, we do not suggest surgery of the frontal region. We reassure the parents that with time, these
changes become less noticeable as the hair grows and the
posterior contour improves. However, in the rare patient
who suffers extremely severe compensatory ipsilateral frontal bossing, we will alter our surgical approach. We have
seen only two children with such severe cosmetic frontal
changes that reshaping of their anterior fossa during the
same anesthetic was required. Both were older children at
the time of diagnosis; one was 11 months old and the other
22 months old. In these children, a more aggressive ap-

proach was required. Discussion of this is beyond the scope

of this chapter.
Throughout the case, bleeding is controlled with bipolar
electrocautery along with injectible cellulose and thrombin
(FloSeal; Baxter International Inc., Deereld, IL). Half a liter
of room temperature saline with bacitracin is used for nal
irrigation. A Jackson Pratt (JP) drain is placed in the subgaleal space and tunneled out through a separate small skin
incision. The scalp ap is closed in two layers with all tension on the galeal closure and none on the skin. Using piercing towel clips on the skin edges can facilitate closure over a
large expanded skull volume. Absorbable suture such as 30
or 40 Vicryl (Ethicon, Somerville, NJ) is used for the galea.
The skin is approximated with interrupted absorbable suture that will not require removal, such as 50 Chromic Gut
(CP Medical, Portland, OR). The incision is dressed with a
light layer of bacitracin ointment. Five years ago we stopped
using any external head dressing and found a decreased
infection rate on internal quality review and no increase in
hematoma rate.

Postoperative Management Including

Possible Complications
The patient is extubated in the operating room. A CT scan of
the brain is obtained to rule out an intracranial hemorrhage.
The patient is then admitted to the pediatric intensive care
unit (PICU). While the patient is in the PICU, an arterial line
is kept in place, along with good intravenous access. A CBC
and coagulation studies (prothrombin time [PTT] and international normalized ratio [INR]) are obtained immediately
when the patient arrives in the PICU. Heart rate, blood pressure, urine output, and JP drain output are monitored closely.

Figure 3011 Intraoperative posterior occipital view. (A) Markings are
made for a biparietal-occipital craniotomy. A strip of bone along the
sagittal sinus is left behind. (B) The bone plates have been switched

and rotated. Barrel-staven-like osteotomies have been made. Plates are

reattached using absorbable microplates and microscrews.


14535_C30.indd 201

4/11/08 11:28:18 AM

202 Pediatric Neurosurgery

If the patients hematocrit is <22, blood will be transfused.
Fresh frozen plasma will be given to maintain the INR <1.4.
If the JP drain puts out >100 cc of blood over 2 consecutive
hours, repeat CBC, PTT, and INR are obtained. The child usually stays in the PICU for 24 hours, after which the patient is
transferred to the surgical ward and monitored in a routine
manner. Immediately after surgery, the child may be active
as tolerated, with the head at 30 degrees. Diet may also be
as tolerated. While in the PICU pain is controlled with an
intravenous (IV) morphine drip starting at 5 g/kg/hour. As
needed oral Tylenol (acetaminophen) along with oral oxycodone are also used. Odansetron and metaclopromide control
postoperative nausea and emesis. After day 1 when the patient is on the ward, the morphine drip is rarely needed. The
child receives 24 hours of postoperative IV cefazolin for staphylococcus prophylaxis. The JP drain is withdrawn by postoperative day 2 or 3. By this time it is usually draining <5 cc
over 24 hours. The child is discharged by postoperative days
3 to 5, as long as the child is able to open his or her eyes.

Complication Avoidance
The endotracheal tube is rst secured by anesthesia. We
then secure the tube to the mandible for added protection.
Points on the forehead and face that will contact the gelcovered horseshoe headrest are well padded with foam.
These points include the lateral supraorbital ridge, the malar eminences, and the zygomatic arch. This will prevent
ischemia to the skin of the forehead and face. To avoid ocular ischemia, the orbital rims are padded with foam that
has areas that are cut out for the eyeballs. Upon adequate
foam protection as described, the infant is then carefully
placed prone in a gel-covered horseshoe headrest. Finally,
all pressure points in the extremities are padded with foam.
A small pillow is placed under the feet to enhance venous
return to the heart.
The scalp opening should be done with minimum blood
loss. Raney clips are not used to decrease injury to hair follicles and skin edges. While drilling the bone, saline is used
to cool the heat that is generated. This will theoretically
decrease the amount of bone resorption.
Fluid and blood volume should be replaced timely. Waiting to transfuse as blood is slowly being lost can end in

14535_C30.indd 202

a sudden catastrophic intraoperative event. We transfuse

blood when the bone is being cut. This is the period in which
the infant loses the most volume.
Injury to venous sinuses and dura must be avoided. Bur
holes are placed above the transverse sinus, and on either side of the superior sagittal sinus. The dura is then
carefully stripped from overlying bone. The occipital bone
is then removed in two pieces. The dura and underlying
brain must be protected with a malleable brain retractor when bone cuts are made. Dural tears are sutured
watertight with 40 Nurolon (Ethicon, Somerville, NJ).
Thrombin-soaked Gelfoam (Pfizer, New York, NY) and
DuraSeal (Confluent Surgical, Waltham, MA) are then applied over the tear.
The risk of venous embolism is minimized by maintaining
the patient in a neutral position, maintaining meticulous hemostasis, and copiously irrigating during any osteotomies.
The scalp must be closed without skin tension, which can
be challenging given a newly expanded cranial vault. This
can be facilitated with the use of piercing towel clips to
achieve stress relaxation of the scalp within 10 minutes. All
tension is then transferred to the galeal closure. Interrupted
suture technique is less damaging to the skin than a running
suture, which can strangulate the skin edges.
Finally, microplates and microscrews can present as scalp
irritation years after an operation due to screw pull-out. We
currently use the minimum number of resorbable plates
and screws to achieve stability of the expansion.
This chapter is a revision of the chapter, Surgical Management of Posterior Plagiocephaly by Richard G. Ellenbogen,
M.D., and Michael H. Mayer, M.D. The chapter appeared in
the Neurosurgical Operative Atlas, Volume # 5, edited by Setti
S. Rengachary and Robert H. Wilkins. The Neurosurgical Operative Atlas was published by the American Association of
Neurological Surgeons (AANS) from 1991 to 2000.
We would like to acknowledge and thank Michael H.
Mayer, M.D., for his help and efforts on the original chapter
published in the rst edition of this work.
All illustrations in this chapter were drawn by Michaele
Miller. Amy Cheney provided editorial assistance, and
graphic design was done by Carrie Wachob.

4/11/08 11:28:19 AM

Sagittal Synostosis
Larry A. Sargent and Timothy A. Strait

Sagittal synostosis is an abnormal condition of the skull

involving premature closure of the sagittal suture. A suture represents the zone between two adjacent skull plates
that has the capacity to form new bone in response to the
underlying expansion of the brain. The growth of the brain
is responsible for the enlargement of the skull and its subsequent shape. As the brain expands, new bone is formed
perpendicular to the cranial sutures. However, if a suture
prematurely closes, it loses the capacity to form new bone.
Therefore, as the brain expands, the skull will not grow
perpendicular to the suture. Premature closure of the sagittal suture will reduce the width of the skull, producing
an abnormally elongated skull called scaphocephaly. The
appearance of the boat-shaped skull is characterized by
anteroposterior elongation and transverse narrowing. With
the compensatory anteroposterior growth, there usually is
increased bossing of the frontal and occipital regions. The
abnormality of the forehead is exaggerated further by the
narrowing that is present at the bitemporal level and by the
loss of the squamous temporal convexity.
Synostosis of the sagittal suture is the most frequent type
of single-suture craniosynostosis. It accounts for 50 to 60%
of all patients within North America with craniosynostosis.
With sagittal synostosis, males are affected three to four
times as frequently as females. Although an inherited tendency has been reported in the literature, most cases of
isolated sagittal synostosis are sporadic in occurrence, with
no known specic etiology. Because the normal lambdoid
and coronal sutures permit compensatory growth in the
anteroposterior axis, infants with sagittal synostosis rarely
have restriction in brain growth, and neurological development is usually normal.
The degree of the abnormality of the skull varies in sagittal synostosis. Complete fusion along the entire sagittal
suture is the most severe form and usually produces a characteristic boat-shaped skull present at birth. The less severe
forms range from fusion of the anterior portion of the sagittal suture with isolated frontal bossing, to fusion of the
posterior portion of the sagittal suture with isolated occipital bulging. Despite the extent or magnitude of premature
suture fusion, the goal of surgery is to release the fused area
and normalize skull shape. It has been our policy to perform
a total vault remodeling procedure with immediate correction of the scaphocephalic deformity.

Patient Selection
All infants born with abnormal skull shapes are suspected
of having craniosynostosis. However, abnormal skull
shapes at birth may occur from either birth trauma or fetal
head position. Generally, this positional type of deformity
will correct itself in a few months. The typical child with
complete sagittal synostosis has an elongated head with
a palpable bony ridge along the sagittal suture, biparietal
narrowing, occipital bulging, and frontal bossing. The diagnosis of sagittal synostosis can usually be made on clinical grounds and confirmed by skull radiographs. We routinely obtain computed tomography (CT) scans to confirm
the diagnosis, to rule out other intracranial abnormalities, and to assess the full extent of the deformity. Threedimensional reconstruction imaging can also be helpful
in delineating the abnormalities. In general, the decision
to recommend surgical intervention is guided by the risk
of potential psychological harm from having an abnormal
and unsightly appearing head. Untreated sagittal synostosis is fully capable of producing increased intracranial
pressure and subsequent brain damage in a small percentage of cases. Release of the fused suture with total skull
remodeling not only virtually eliminates the risk of brain
damage due to growth restriction, but it also restores the
skull shape, creating the potential for normal growth and
In our experience, the best results from reconstruction
occurred when infants underwent surgical repair between
8 and 10 months of age. This time period has several advantages. First, their blood volumes are greater than in newborns. Second, remodeling is simpler (less operative time)
because the bone is far more malleable and easier to contour. Third, bone defects or gaps left after skull remodeling
are not a problem due to rapid bone healing. Finally, the
rapid brain growth that occurs during the rst 3 years of life
benets additional bone remodeling.
Older patients will require modication of this plan due to
the presence of more rigid bone and slower cranial growth
patterns. In the patient >3 years of age, more extensive surgical maneuvers are required with denitive bone work,
better xation, and no bony gaps. Furthermore, older patients may require reconstruction of secondary facial deformities.


204 Pediatric Neurosurgery

Preoperative Preparation
Patients with craniosynostosis should be evaluated and
treated at a craniofacial center with a team approach. Our
center has a multidisciplinary team to ensure that the
patients and their families are evaluated and treated in a
coordinated manner. The team combines the expertise of
each specialist to provide a level of comprehensive care
unmatched by a single physician. The pediatricians play a
key role in early diagnosis and referral so that successful
treatment is accomplished. A craniofacial surgeon, neurosurgeon, and pediatric anesthesiologist compose the operative team. Each of these specialists sees the patient in a
preoperative evaluation. A pediatric intensivist is consulted
when associated medical problems are present. The planned
surgical procedure is fully explained to the parents by the
craniofacial surgeon and the neurosurgeon in their ofce
A CT scan is routinely performed on each patient who is
clinically diagnosed as having craniosynostosis. This helps
to fully assess the deformity as well as plan the operative
Because blood transfusion is generally necessary, the parents are asked to provide designated-donor blood. Aspirin
and other medications capable of prolonging the blood clotting time are avoided.

Nasotracheal general anesthesia is preferred, with the tube
sutured to the caudal septum with 30 silk suture. An arterial line is placed in the radial artery to monitor blood pressure, with one to two peripheral intravenous lines inserted
in the extremities. A central line silicone catheter is inserted
into the femoral vein and sutured in place. A Foley catheter
is inserted to monitor urine output. Plastic sterile drapes are

Figure 311 (A) The patient is positioned prone with the neck slightly
extended. This provides excellent exposure of the entire skull. A bean


placed anteriorly and posteriorly from the shoulders down

to minimize heat loss and to keep the infant dry. Broadspectrum intravenous antibiotics are administered at the
induction of anesthesia.

Operative Procedure
Proper positioning of the patient is very important in total
vault remodeling (Fig. 311). Following administration of
the anesthetic agent, the infant is placed in the modied
prone position with the neck only slightly extended. The
chin and neck are placed in a stable position by a molded,
suctioned bean bag. Padded foam is placed beneath the chin
to avoid pressure ulceration of the skin. All other joints are
appropriately padded with foam.

The rst stage of the procedure involves mobilization of
the scalp aps. To gain access to the entire calvaria, a zigzag bicoronal skin incision is designed from ear to ear and
centered approximately between the anterior and posterior
fontanelles. This type of incision yields a superior cosmetic
result compared with the straight bicoronal incision. Only a
small amount of scalp hair is cut, creating a 0.5 cm margin
on either side of the incision. Prior to incising the scalp, 0.5%
lidocaine with 1:200,000 epinephrine is injected intradermally to diminish bleeding. The scalp edges are compressed
while the skin is incised with a scalpel. Hemostatic clips are
applied to the scalp margins. The scalp aps are elevated in
the supraperiosteal plane using a Colorado microcautery tip
to separate the scalp and its underlying areolar tissue from
the periosteum. Scalp aps are mobilized anteriorly to the
supraorbital rims and posteriorly to slightly below the level
of the external occipital protuberance. After the aps have

bag padded with foam is used to support the chin and jaw in this position. (B) The zig-zag bicoronal incision is outlined.

been mobilized, the temporalis muscle, with a 2 cm cuff
of periosteum, is elevated off the squamous portion of the
temporal bone with a periosteal elevator through a large
semicircular incision. The temporalis muscle is taken down
to the level of the oor of the middle fossa. With this exposure, total vault remodeling can be performed (Fig. 312).
The neurosurgical portion of the procedure involves
dismantling the skull in several pieces. Skull plates are removed with care so as not to tear the dural covering of the
brain and the dural venous sinuses (Fig. 313). The vault is
removed in four separate pieces, leaving intact bone overlying the sagittal suture. It is helpful to utilize the M-8 attachment of the Midas Rex drill (Medtronic Inc., Ft. Worth, TX)
to make multiple bur holes for elevation of the individual
bone plates. The bur holes are connected with the B-5 attachment of the Midas Rex drill, and the bone edges are

Sagittal Synostosis 205

waxed. Initially, bilateral parietal craniotomies are elevated.

Multiple bur holes are made ~2 cm off the midline from
the sagittal suture. Additional bur holes are made along the
coronal suture anteriorly, slightly below the parietosquamosal suture inferiorly and at the parieto-occipital junction
posteriorly. Through each individual bur hole, a ball-hook
dissector is utilized to separate the underlying dura from
the bone plate. Once the bone plate has been elevated, a
single layer of Surgicel (Ethicon, Somerville, NJ) is placed
overlying the dura and moistened with a liquid thrombin
spray. Next, the biparieto-occipital craniotomy is elevated.
Multiple bur holes again are placed along a line at the level
of the external occipital protuberance. To facilitate elevation
of this posterior bone plate, it is helpful to place bur holes on
each side of the lambdoid sutures and connect them using
the M-8 attachment. A ball-hook dissector is then used to

Figure 312 Sagittal synostosis. (A) The design of the craniotomies
is shown leaving a sagittal bone strut. The two parietal bone aps are
removed rst, followed by the occiput and the frontal bone aps. (B)
Anteroposterior shortening is accomplished with posterior inclination
of the frontal bone and anterior inclination of the occipital bone. This
effectively results in biparietal widening. (C) Radial osteotomies are used

to recontour the bone aps with the help of bone-contouring forceps.

Barrel-stave-like osteotomies are placed in the temporal bone and fractured laterally. The frontal and occipital bone plates are secured at their
base with 28 gauge wires. Both of these bones are secured to the sagittal bone strut with 28 gauge wires.

206 Pediatric Neurosurgery

Figure 313 (A) Lateral intraoperative photograph of the scaphocephalic

skull with the frontal bone on the left. The bone aps to be removed
are marked. The temporalis muscle with a cuff of periosteum has been

separate the dense adhesion between the cranial suture and

the overlying bone. This technique is repeated in the midline
at the level of the torcula. Bur holes are made on each side of
the sagittal sinus and connected using the M-8 attachment.
Frequently, a keel of bone is present that requires careful
separation of the dura while drilling away the small bridge
of bone. Emissary veins from the sagittal sinus are readily
controlled with bipolar coagulation. Finally, elevation of a
single bifrontal bone plate is performed with the inferior
margin placed 1 cm above the supraorbital rim. Again, it is
useful to place bur holes close to the midline along the inferior saw cut. The two bur holes are then connected using the
M-8 attachment to divide the small keel of bone that corresponds to the superior aspect of the crista galli. This maneuver facilitates elevation of the bifrontal bone plate. After the
bone plates have been removed, the craniofacial surgeon
recontours the plates at a side table. Simultaneously, the
neurosurgeon makes vertically oriented barrel-stave-like
osteotomies in the squamous portion of the temporal bone,
which are outfractured at their base to increase the lateral
projection or width of the skull base.
For a child with marked bilateral frontal bossing, the bossing is lessened by plicating the dura in the frontal region
from superolaterally to inferomedially with 40 Vicryl suture. Frequently, a similar effect is achieved with bipolar
coagulation to the dura itself.
Dural tears are repaired with 40 Vicryl suture. In the
event of a dural tear or laceration overlying a venous sinus,
compression with Gelfoam soaked in thrombin is maintained until a dural graft is sutured in place overlying the
dural rent. Interruption of the venous connections between
the bone plates and underlying dural sinuses accounts for
the major source of bleeding. The cumulative effect of removing the entire skull produces a relatively substantial
blood loss and mandates the need for routine blood transfusions. Generally, blood transfusions are initiated as the
bone plates are being removed. The rate of transfusion varies according to the amount of bleeding encountered as

elevated off the skull (dashed line). (B) View from the top showing the
four bone aps to be removed and the sagittal bone strut that is left
in place.

each bone plate is removed. Communication between the

surgeons and the anesthesiologist is imperative during this
portion of the procedure.
The four bone plates are taken to a side table where the
craniofacial surgeon begins recontouring (Fig. 314). The
occipital bone ap is generally too convex, with a narrow
transverse diameter. To create a more normal convex curvature of this bone, radial osteotomies are performed with
bone-cutting scissors. The bone is attened using Tessier
(Lorenz Surgical, Jacksonville, Florida) bone-contouring forceps, creating a more gentle curvature with a wider transverse diameter. The frontal bone plate is remodeled next in a
fashion similar to the occiput. The frontal bone is too narrow
in the bitemporal plane, in addition to the frontal bossing.
To correct this abnormal contour, radial osteotomies are performed with recontouring using the Tessier bone-contouring
forceps to atten and widen the forehead like the occipital
bone. The two parietal bone aps are abnormally at, lacking the convexity needed to increase the transverse width.
Radial osteotomies are performed on these two bone aps,
with inward bending to create a more normal curvature. The
occipital bone ap is returned rst. Prior to replacement, a
strip of bone ~6 cm 1 cm at the base of the skull is cut
with a reciprocating saw, and it is infractured and secured
with 28 gauge wire. The posterior bone plate is then anteriorly inclined with xation to the base of the skull in the
midline using two 28 gauge wires. A portion of the posterior
end of the sagittal strut is resected to allow shortening of
the anteroposterior length. The occipital bone is secured to
the sagittal strip of bone with two 28 gauge wires that are
slowly twisted down. The frontal bone ap is returned next.
Two 28 gauge wires are used to secure the frontal bone to
the supraorbital rims. The frontal bone ap is inclined posteriorly, resetting a small portion of the anterior end of the
sagittal strut where it is secured to the frontal bone. Approximately 1 to 2 cm of anteroposterior length shortening of
the skull is accomplished once the wires securing the frontal
and occipital bones to the sagittal strut are secured down.


Figure 314 (A) The four bone plates, frontal (bottom of photograph),
biparietal, and occipital are taken to a side table. The abnormal elongation of the skull is shown in this intraoperative photograph. (B) Radial
osteotomies are used to contour the bone aps with the help of bone10.1055/978-1-60406-039-3c031_f004

Sagittal Synostosis 207

contouring forceps. This intraoperative photograph demonstrates how

the bone aps are replaced. Anteroposterior shortening is accomplished
with biparietal widening.

Figure 315 Preoperative photographs (AC) of infant with sagittal synostosis and postoperative results (DF) several months after total vault

208 Pediatric Neurosurgery

These wires are cinched down slowly and gradually to allow
the brain and dura to compensate with lateral expansion.
To facilitate the shortening of the anteroposterior length
without creating undue intracranial pressure, a small dural
incision is made in the temporal region to vent cerebrospinal uid. This shortening due to the posterior inclination
of the frontal bone and anterior inclination of the occiputs
results in a transverse widening of the brain and thus, the
parietal bones (Fig. 312). Due to the shortening of the skull,
the anterior and posterior edges of the parietal bones will
need to be trimmed using bone scissors. The parietal bones
are next secured only to the sagittal strip of bone with 28
gauage wires. The lack of xation on the remaining three
sides of the parietal bones allows for lateral brain expansion and skull growth in this area. Once all four bone plates
have been replaced as described, a dramatic improvement
in overall skull shape can be seen. The temporalis muscles
are replaced and secured to drill holes in the skull in multiple places with 40 Dexon.
A subgaleal drain is left beneath the posterior scalp ap.
The scalp is closed in layers using 40 Dexon to approximate
the galea. A 50 plain catgut is used to close the coronal skin
incision so as to avoid sutures that require removal. A bulky
head dressing of Kerlix sponges and rolls is applied.

Postoperative Management Including

Possible Complications
At the completion of the operative procedure, infants are
taken directly to the pediatric intensive care unit. They are

positioned in an infant car seat with the head elevated. Elbow restraints are used to avoid pulling on the subgaleal
drain or the head dressing. In general, most infants remain
intubated overnight and are then extubated on the rst
postoperative day. Both suction drain and bulky pressure
head dressing are removed after 48 hours. Intravenous prophylactic antibiotics are continued for an additional 3 days.
Transfer out of the intensive care unit to the oor is usually
possible on the second or third postoperative day, with the
total hospital stay usually 5 days.

Operative risks include anesthetic risks, blood loss, infection, air embolus, dural/sinus tears, and cortical injuries. A
total vault reconstruction requires longer operative time,
with the potential for greater blood loss. However, this approach provides immediate normalization of the skull and
virtually eliminates the chance of either recurrent synostosis or persistent skull deformities associated with strip
craniotomies. Any potentially increased risks associated
with total vault remodeling are lessened by the surgical
team approach and meticulous attention to detail. The signicant improvements in skull shape seem to justify this
more aggressive and comprehensive approach to sagittal
synostosis (Fig. 315).

The Separation of Craniopagus Twins
Sami Khoshyomn and James T. Rutka

Patient Selection
Craniopagus twins are the rarest form of conjoined twins
in the human organism. They account for 2 to 6% of all conjoined twins, occurring in only 1 in 2.5 million live births.
Conjoined twins are always genetically identical and share
the same gender. Females are affected more commonly, and
no association with race, age, parity, maternal age, heredity,
or environmental factors has been found. Recently Spencer
suggested that craniopagus twins are formed due to secondary fusion of two primitive neural folds of two dorsally
oriented embryonic disks prior to the fourth week of gestation. OConnell et al developed an extensive classication of
craniopagus phenotypes, and Bucholz et al proposed four
subclassications for total craniopagus twins: frontal, parietal, temporoparietal, and occipital. Bucholz et al dened
frontal craniopagus twins as facing each other with the axis
of bodies forming an acute angle. Temporoparietal twins
are joined above the external auditory meatus and occipital
craniopagi are joined at or above the occipital protuberance. The parietal craniopagi, also referred to as vertical
craniopagi by OConnell, are fused at the vertex with varying degrees of rotations of one head as compared with the
other. The classication of the morphology of fusion is important, as it directly relates to underlying abnormalities of
the cortex and shared arterial or venous anatomy. A shared
venous system, especially a common superior sagittal sinus,
is believed to be the single most important factor affecting survival and long-term morbidity when separation of
craniopagus twins is attempted. The main questions to be
answered prior to considering separation are:

tated at 30 (Fig. 321). The twins were investigated using

computer tomography (CT), magnetic resonance imaging
(MRI), and angiography to determine whether they shared
cerebral structures. Three-dimensional reconstructed CT
scan showed the relationship of the cranial vaults of the
twins (Fig. 322). MRI scan of the brain showed a common
dural shelf separating the brains of the twins along half
of the plane of attachment and interdigitation of the two
brains across an undulating layer of leptomeninges for the
other half (Fig. 323). They did not appear to share any brain
tissue. Magnetic resonance venography (MRV) revealed as
anticipated a common or shared superior sagittal sinus. Cerebral angiography as the next imaging study demonstrated,
in addition to a common superior sagittal sinus, the passage
of middle cerebral artery (MCA) branches from one twin to
the other (Figs. 324 and 325). Utilizing interventional
neuroradiological techniques, we embolized the distal
branches of the MCA from twin A bridging to twin B. This
achieved a partial separation of their cerebral arterial circulations. A preoperative endovascular cannulation and occlusion of the shared superior sagittal sinus was attempted but
was unsuccessful. Other authors have advocated an open
staged ligation of shared venous structures prior to separation, hoping that over time this would promote deep venous
collaterals and drainage in the twin without a complete superior sagittal sinus.

1. Is there any shared brain tissue?

2. Are there any shared major arterial or venous channels?
3. Can the resulting skin and bony skull defect be closed?
In this chapter, using our experience with separation of
total craniopagus twins at the Hospital for Sick Children in
Toronto, we provide an overview of the surgical considerations and steps undertaken for separation.

Preoperative Preparation
Craniopagus twins referred to our neurosurgical service
were 3 years old and were a total vertex craniopagus ro-

Figure 321 Preoperative photograph of craniopagus twins A (right)

and B (left) playing.



210 Pediatric Neurosurgery

In addition to the key issues of shared neuronal tissue or
arterial and venous anatomy, one important perioperative
consideration is the complex problem of bony and soft tissue defect and reconstruction after separation. In this regard, we were aided by our craniofacial plastic surgeons,
who inserted tissue expanders into the subcutaneous tissues of the scalp at least 2 months prior to the attempted
separation so that the scalp wounds could be closed without
using skin grafts. On the morning of the separation, the skin
incisions were outlined by the plastic surgeons (Fig. 326).
The craniofacial team also helped us with the harvesting of
autogenous split-thickness bone grafts for rigid repair and
xation of the bony skull defects after the separation was

Operative Procedure
Figure 322 Computed tomography scan with three-dimensional reconstruction demonstrating the relationship of the cranial vaults of
the twins.

On the morning of surgery each craniopagus twin was

placed on her own operating room table and own horseshoe
headrest. Anesthesia was induced by two separate teams by
techniques previously described. Arterial and venous lines
and Foley catheters were placed. The scalps were shaved entirely to allow the craniofacial plastic surgeons to mark the
skin aps. The twins were prepped and draped separately
so they could be rotated together and remain sterile for the
closure of the wounds following the separation.

Figure 323 T1-weighted magnetic resonance imaging (MRI) scans;
sagittal (left) and coronal (right) views. A dural shelf separating the anterior half of the plane of attachment is appreciated. Interdigitation of

the parietal lobes is also seen. MRI suggests a shared sagittal sinus but
no shared cerebral tissue.

32 The Separation of Craniopagus Twins 211

Figure 324 Cerebral angiogram showing the sagittal sinus of one

twin joining the sagittal sinus of the other twin.
Figure 326 Artists illustration of the scalp incision markings and the
tissue expanders.

Figure 325 Cerebral angiogram of the twins in early venous phase (left) and arterial phase (right). The images show branches of the middle
cerebral artery of twin A traversing and supplying brain tissue of twin B.

212 Pediatric Neurosurgery

Figure 327 The scalp is opened along the incision markings, and a
craniotomy is performed.

Figure 328 The sagittal sinus of one twin is divided just before it
joins the sagittal sinus of the other twin.

Next, the scalp incision was made and the tissue expanders were removed. Using intraoperative neuronavigation,
we selected the site of the rst craniotomy and dural opening so that the location of the shared sagittal sinus could
be exposed rst (Fig. 327). Following removal of a bone
segment, we observed the superior sagittal sinus of one
twin entering the sagittal sinus of the other twin. It was
necessary to ligate the sinus of one twin and divide it just
before it joined the normal sagittal sinus of the other twin
(Fig. 328). The twins were then rotated together circumferentially under the supervision of the neuroanesthesiologists, exposing more skull so further craniotomies could be
performed and the dura could be opened circumferentially.
In this fashion, the underlying cerebral hemispheres of the
twins were exposed along the plane of attachment. The preoperative embolization of the MCA branches bridging from
twin A to B greatly facilitated the separation of the interdigitation between the twins where the dura was decient. An
ultrasonic aspirator (Cavitron; Dentsply International Inc.,
York, PA) was used to expedite the separation of two cerebral hemispheres in a subpial plane in this region.
Division of one twins falx below the level of the other
twins superior sagittal sinus represented the last step of the
separation. Upon complete separation of the twins brains,
the operating room tables were pulled apart allowing two
neurosurgical teams to simultaneously repair the dural defects. Large dural grafts were used to repair the dural defects
(Fig. 329). The harvested bone fragments from the craniotomies were split by the craniofacial plastic surgeons and
utilized to repair the cranium of each twin using titanium
plates and screws (Fig. 3210). The nal step in this complex
procedure was the scalp closure over subcutaneous drains
(Fig. 3211).

Figure 329 A large dural graft is used to repair the dural defect.

32 The Separation of Craniopagus Twins 213

Postoperative Management Including

Possible Complications
The management of craniopagus twins in the postoperative
period requires close attention to several potential issues
that may arise. Raised intracranial pressure (ICP) can develop postoperatively in one or both of the separated twins
from venous congestion if one or more of the shared venous
sinuses are sacriced. Venous congestion may be complicated by an intracerebral hematoma that may require neurosurgical evacuation. Raised ICP may also arise from hydrocephalus, which can develop in the postoperative period.
Symptomatic hydrocephalus may require the insertion of a
ventriculoperitoneal (VP) shunt. Finally, care must be taken
to ensure that the rotational scalp aps and split-thickness
cranioplasties heal without signs of infection, bone resorption, or vascular compromise.


Figure 3210 The skull is reconstructed by the plastic surgeon using

harvested bone fragments from the craniotomies utilizing titanium
plates and screws.

In summary, prior to undertaking the separation of craniopagus twins, a detailed neuroradiological map consisting
of CT, MRI, and cerebral angiography must be obtained. In
addition, the use of intaroperative neuronavigation can be
invaluable in localizing shared anatomical structures, particularly a superior sagittal sinus. Lastly, endovascular techniques may be used as an alternative to staged surgery to
detach the cerebral circulations joining the twins.

Figure 3211 The scalp is closed, and a drain is placed.


214 Pediatric Neurosurgery

The separation of craniopagus twins must be considered
one of the most challenging and difcult procedures in the
repertoire of neurosurgical cases. Morbidity and mortality
are high. In the case of the twins described here, both twins
survived the separation. However, the twin whose superior sagittal sinus outow was taken during the separation
suffered a deep intracerebral hematoma that was managed
conservatively, and developed hydrocephalus requiring a
VP shunt. This twin died from a pulmonary complication ~1
month after surgery. The other twin whose superior sagittal
sinus was maintained is alive and well, and leading a normal
life in her home country of Pakistan (Fig. 3212).
This chapter is a revision of the chapter, The Separation of
Craniopagus Twins by Harold J. Hoffman, M.D., and James
T. Rutka, M.D. The chapter appeared in the Neurosurgical
Operative Atlas, Volume # 6, edited by Setti S. Rengachary
and Robert H. Wilkins. The Neurosurgical Operative Atlas
was published by the American Association of Neurological
Surgeons (AANS) from 1991 to 2000.
The intellectual contributions and neurosurgical skills of
Dr. Harold Hoffman who led the Toronto Hospital for Sick
Children team in the separation are acknowledged here. Dr.
Hoffman died November 14, 2004.
The authors would also like to acknowledge the photographic art contributions by Mr. Robert Teteruck, senior
photographer at the Hospital for Sick Children.

Figure 3212 Postoperative photograph of the surviving twin (twin

A) 1 year after separation.

Endoscopic Approaches to the Ventricular
David F. Jimenez
Utilization of endoscopic techniques to treat ventricular lesions has markedly increased in recent years, as evidenced by
the large number of published articles with neuroendoscopic
topics. Already commonplace in pediatric neurosurgery, increased training of graduating neurosurgical residents and
many postresidency training courses have led to continued
popularization of these techniques. With increased experience has come improved outcomes, less morbidity, and
higher patient satisfaction. Corporate interest in the eld by
prominent endoscopic companies has also led to the development of several endoscopic systems dedicated to neurosurgery. Miniaturization of forceps, graspers, balloons, scissors,
bipolar and monopolar units, lasers, and other instruments
now contribute to an ever expanding array of endoscopic
instrumentation. Because the use of such scopes and instruments creates a working environment that is inherently different from that for which we have classically trained, careful planning, preparation, and execution are paramount to
avoid complications and poor outcomes. Described herein
are practical pointers that relate to a variety of approaches to
the ventricular system and the pathology that may affect it.

Prepoperative Preparation
Careful and thoughtful preoperative planning is essential for
a successful outcome when performing neuroendoscopic
procedures. As such, magnetic resonance imaging (MRI) with
adequate visualization of the ventricular system should be
obtained for proper planning and execution. The exquisite
detail given by the multiplanar MR images greatly enhances
preoperative planning, preparation, and familiarization with
pertinent surgical anatomy. Based on anatomical corridors,
appropriate measurements can be made on the scans to plan
bur hole placement and scope angle trajectories. In cases of
multiloculated hydrocephalus, axial computed tomography
(CT) scans with intraventricular contrast medium are needed
to ascertain the level of compartmentalization. Based on
these images, the best angle of approach and number of
fenestrations can be planned and executed.

Surgical Suite Organization

Appropriate setup and ergonomic organization of the operating room are of utmost importance to minimize surgeon

fatigue. The operating table should be reversed or placed in

such a manner as to place its base away from the patients
head. This maneuver allows the surgeon and assistant to sit
comfortably without obstruction of the feet or legs by the
tables base. Accessories such as intravenous lines, electrical cords, laser bers, and irrigation lines should exit the
operative eld toward the patients legs, which will diminish tangling of lines at the operative site. The anesthesia
team can be placed toward the patients legs with the use
of extended tubing, which allows full access to the patients
head. Equipment such as video monitor, camera, and light
source should be located on the side opposite and across
from the surgeon (Fig. 331). Newer visualization systems
such as the OR1 (Karl Storz, Germany) allow for strategic
placement of multiple monitors so that the entire operating room team can see the procedure continuously. This
setup is extremely important, as it allows the surgeon to sit
comfortably and have direct and neutral visual access to the
monitor during the procedure. This design will minimize
cervical and upper back strain, as is commonly seen when
the monitor is located lateral or behind the surgical team.
Sitting also allows the surgeon to use both feet to control
separate instruments (e.g., laser and irrigation) in addition
to bilateral hand control. A craniotomy tray should be available (not opened) during all endoscopic procedures. Should
a complication arise and should it suddenly become necessary to perform a craniotomy, little time will be spent by the
staff looking for the appropriate instruments.

Ventricular Irrigation
Paramount to excellent visualization is a clear uid ventricular environment and as such, adequate irrigation is of
utmost importance. Lactated Ringers solution is the irrigant of choice for neuroendoscopic procedures, as its pH
and electrolyte composition are closest to cerebrospinal
uid (CSF). Normal saline should be avoided due to its low
pH (5.1) and associated irritating effect. An irrigating system can easily be set up either using motorized units or
with instrumentation available in most operating rooms.
The components of such a system include a pressure bag, a
uid warmer and the Malis bipolar irrigator unit (Codman
& Shurtleff, Inc., Raynham, MA). A 1 liter bag is placed in the
pressure bag at 300 mm Hg. The uid is then passed through
a standard uid warmer. The warmed uid is then advanced


14535_C33.indd 215

4/23/08 3:13:41 PM

216 Pediatric Neurosurgery

Figure 331 Surgical setup. The surgeon should sit on a chair with armrests. Depending on the side of entrance, a monitor should be placed
directly across for direct viewing. Anesthesia is placed at the foot of
the table along with accessories such as suction bipolar and monopolar

through the Malis irrigating bipolar unit and delivered to

the endoscopes irrigating port. The rate of delivery of irrigating uid into the ventricular system can be adjusted
and controlled with a foot pedal. Regardless of which unit
or system is used, extreme care must be taken to ensure
that the uid inow equals the outow. Unattended uid
delivery without proper egress can lead to acute intracranial
pressure elevations and tentorial herniation.

Operative Procedure
General Principles
General endotracheal anesthesia should be used for the
majority of endoscopic procedures. Occasionally, an adult
patient who has a high surgical risk can be operated upon
under a conscious sedation protocol without problems. Two
peripheral intravenous lines are sufcient, and no central
venous access is necessary. A urinary catheter is needed
for complex procedures but not for simple fenestrations. A
single dose of an antibiotic of choice (nafcillin, oxacillin, or

14535_C33.indd 216

units. The laser (when used) is placed near the patients head due to
the length of the laser ber. Newer systems (OR1; Karl Storz, Germany)
allow for the strategic placement of multiple monitors.

rst-generation cephalosporin) is given ~30 minutes prior

to surgery. Only a very thin strip of hair needs to be shaved
over the proposed incision site. The head is prepped with
povidone-iodine and then draped wide enough for a possible emergency craniotomy. An irrigating uid collecting
pouch should then be placed in the operative eld.
Prior to skin incision the endoscopic system should be
assembled and appropriately tested to verify a fully functioning unit. When using a exible endoscope, a self-retaining retractor should be used to hold the body of the scope.
A simple Greenberg or Leyla self-retaining retractor can
be adequately set up to hold the scope so that the exible
component can be easily manipulated. Newer pneumatic
systems (Mitaka, Japan) offer greater exibility and stability and can be used with either exible or rigid endoscopes
(Fig. 332). Following the skin incision, care is taken to preserve a pericranial ap that may be used at the end of the
procedure to cover large dural defects from the cannulating units. After bur hole placement, a small dural opening is made and a standard ventricular catheter is inserted
in the corresponding ventricle or cyst and the intracranial
pressure is measured manometrically; CSF should be sent
to the laboratory for the appropriate and indicated stud-

4/23/08 3:13:41 PM

33 Endoscopic Approaches to the Ventricular System 217

monitor, the same orientation is seen and the movements of
the scope are correlated with visual images. If this important step is not done, disorientation and reverse sensing will
occur, making the procedure very difcult to perform.

Foramen of Monro Access

Figure 332 A left parietal approach to an intraventricular meningioma. The pneumatic Mitaka arm is attached to the left side of the
bed and positioned at an appropriate height to accommodate the

ies. Following removal of the ventricular catheter, a no. 14

French peel-away introducer is inserted into the ventricle.
Some systems have rigid cannulas that may be used instead
of the peel-away disposable introducer. The scope is then
inserted into the cannula or introducer and the intraventricular anatomy is ascertained.
Prior to insertion of the scope into the ventricular system,
it is very important to align the cameras view with the
actual spatial orientation of the ventricle. An aerial view of
the patients head and operative eld will easily permit the
surgeon to ascertain the correct orientation. The camera
can be rotated in such a manner so that when looking at the

Figure 333 (A) Intraoperative view of a large suprachiasmatic cyst extending into the third ventricle. A thick capsule is seen expanding the
foramen of Monro. The junction of the choroid plexus, thalamostriate
vein (inferiorly), and the septal vein (superiorly) is seen en-face. (B) Following yttrium-aluminum-garnet (YAG) laser fenestration, the deated

Access to the foramen of Monro for lesions such as colloid

cysts, suprachiasmatic cysts, or even septum pellucidotomies is best gained by placing the bur hole far more lateral
than standard ventriculostomy access (Figs. 333). A coronal MRI should be used to measure the distance from the
midline, and it is usually 6 to 7 cm lateral on the side with
the greatest frontal horn dilation (Fig. 334). This maneuver
will allow en-face visualization of the cyst and easier entrance into the third ventricle. Depending on anteroposterior
extension of the lesion, the bur hole may be located more or
less anteriorly to the foramen of Monro (Fig. 335).

Anterior Third Ventricular Access

The patient should be placed supine with the head neutral
to the midline and in slight extension (5 to 15 degrees). A
bur hole is placed at or slightly behind the coronal suture
and 3 cm lateral to the midline. Perpendicular access to the
ventricle will place the scope slightly behind the foramen of
Monro (Fig. 336). The choroid plexus, the thalamostriate
vein, and the septal vein should be easily and clearly visualized. The scope is advanced through the foramen of Monro,
being careful to avoid injury to the veins or the columns of
the fornix. Once in the third ventricle, the mammillary bodies, tuber cinereum, and infundibular and suprachiasmatic
recesses will be identied and seen (Fig. 337). A 30 degree
angled scope will permit the visualization of the lamina terminalis, anterior commissure, and anterior third ventricular

cyst wall can be seen inside the third ventricle. The top of the cyst can
be seen attached to the undersurface of the anterior commissure. The
ipsilateral fornix is seen superiorly (12 oclock) and contralateral foramen
of Monro at 9 oclock.


14535_C33.indd 217

4/23/08 3:13:42 PM

218 Pediatric Neurosurgery

Figure 334 Access to a colloid cyst is best obtained by placing the bur hole much more laterally than standard Kochers
point. Otherwise, attempts at removing the lesion may injure the fornix secondary to too much medial tension.

roof (Fig. 338). A third ventriculocisternostomy can now be

easily performed, choosing the site of fenestration anterior
to the mammillary bodies and posterior to the infundibular
recess. Care should be taken to avoid injury to the basilar
artery. Lesions, masses, or cysts of the third ventricle can
easily be reached with this approach. Exquisite visualization
of the suprasellar cistern may be obtained when fenestrating suprachiasmatic cysts (Fig. 339).

Posterior Third Ventricular Access

Figure 335 Axial diagram shows entrance into the lateral tip of the
frontal horn for better access to lesions at the foramen of Monro that
extend into the posterior third ventricle. A second bur hole may be
used for improved access and instrument manipulation.

The patient is placed supine, head neutral but with moderate extension of the neck (30 to 45 degrees). An incision is
made in the forehead, 2 to 3 cm above the eyebrow, at or
slightly lateral to the midpupillary line (Fig. 3310). Care
should be taken to avoid the supraorbital nerve. The incision
should be placed in one of the forehead creases to conceal
postoperative scars. A bur hole is made directly under the
incision, and the dura is opened. A ventricular catheter can
be inserted into the frontal horn and proper trajectory ascertained. Next, a peel-away introducer or scope cannula is
then inserted into the tip of the frontal horn, using the same
trajectory. Directly in front, the foramen of Monro and body
of the lateral ventricle and occipital horn can be visualized
(Fig. 3311). Entrance into the third ventricle is gained by
passing the scope medially through the foramen of Monro.
At this point, the posterior third ventricle is now easily
visualized. The massa intermedia, posterior commissure,
pineal recess, pineal gland region, aqueduct, tela choroidea,
and posterior third ventricular wall can also be easily seen
(Fig. 3312). This approach is ideal for rigid scopes. Careful review of sagittal MR images will help in choosing the
proper location of the bur hole and the trajectory of the
endoscope during these frontal approaches (Figs. 3313 and


14535_C33.indd 218

4/23/08 3:13:44 PM

33 Endoscopic Approaches to the Ventricular System 219

Figure 336 For access to the anterior third ventricle, the access bur hole should be placed behind the coronal suture along the midpupillary line.
Placement anterior to the coronal suture will make visualization of the oor difcult and can potentially injure the fornix.

Figure 337 Endoscopic view of the oor of the third ventricle demonstrates the paired mamillary bodies and tuber cinereum. The basilar
artery bifurcation is seen along with both proximal posterior cerebral
arteries between the mammillary bodies. The pons is visualized behind the basilar artery.

Figure 338 Endoscopic visualization of the undersurface of the

anterior commissure using a 30 degree angled endoscope. The two
columns of the fornix are seen extending posteriorly and forming the
anterosuperior margins of the foramina of Monro.



14535_C33.indd 219

4/23/08 3:13:45 PM

220 Pediatric Neurosurgery

Figure 339 After fenestration of a suprachiasmatic cyst, the suprasellar contents are visualized. Notice a stretched infundibulum and the
pituitary gland lobes located in the sella turcica. The carotid bifurcation is seen with the rst portion of middle cerebral artery (left) and
rst portion of anterior cerebral artery (right) segments along with
an elongated posterior communicating artery and its hypothalamic

3314). With this approach there is no injury to the foramen

of Monro structures because no torquing is necessary, as
is commonly done with a coronal approach. Resection of
lesions, biopsies, and fenestrations can now be easily performed. Once the procedure is nished, a single bur hole
titanium cover is placed over the bur hole and the galea is
closed. A subcutaneous absorbable suture and Steri-Strips
are used to close the skin. This approach not only gives excellent cosmetic results but also a superb, direct view to the
posterior third ventricle.

Figure 3310 For lesions of the posterior third ventricle, a direct

frontal approach can be made through a forehead crease and slightly
lateral to the supraorbital neurovascular bundle. The frontal horn is
accessed through its frontal tip.

Figure 3311 (A) The location and angle of the endoscope can be varied depending on the superior or inferior location of the lesion in the
posterior third ventricle. An inappropriate trajectory can make the tar-

get area difcult to visualize. (B) The scope enters the tip of the frontal
horn. Medial rotation allows passage into the third ventricle and access
to lesions in the posterior wall of the third ventricle.


14535_C33.indd 220

4/23/08 3:13:46 PM

33 Endoscopic Approaches to the Ventricular System 221

Figure 3312 (A) Endoscopic view of the posterior third ventricle. A
dilated aqueduct of Sylvius with a membrane is seen inferiorly. The
posterior commissure is seen immediately superior to the aqueduct.
The pineal gland is visible between the posterior commissure and the

massa intermedia. (B) The roof of the third ventricle composed of
the tela choroidea and choroid plexus is seen between the two thalami;
the pineal gland and habenular commissure are seen above the massa


Figure 3313 Preoperative axial T1-weighted magnetic resonance

image for an 18-year-old patient with a history of papilledema and
severe headaches. Note the large septum pellucidum cyst.

Figure 3314 Postoperative T1-weighted magnetic resonance image

of the same patient following endoscopic laser fenestration using a
laterally placed coronal bur hole. Symptoms are fully resolved.



14535_C33.indd 221

4/23/08 3:13:50 PM

222 Pediatric Neurosurgery

Figure 3315 Supine position for access to the frontal horn
and anterior third ventricles.

Frontal Horn Access

As with approaches to the anterior third ventricle, the patient should be supine with slight neck exion (Fig. 3315).
The bur hole should be at or 1 cm in front of the coronal
suture. However, its lateral location may vary between 3
and 7 cm from the midline or medially located lesions (Fig.
3316). An extreme lateral placement is of particular importance for lesions of the septum pellucidum (Fig. 3317) or
medial roof of the horn. Cysts, lipomas, or metastatic lesions

can occur in this area. The classic view of the foramen of

Monro will be encountered (Fig. 3318). The rst structure
to be identied is the choroid plexus, which, when followed,
will lead to the foramen. In addition, the thalamostriate and
septal veins should be easily identiable. In cases of longstanding hydrocephalus, the septum pellucidum is not existent and the contralateral ventricle will be visualized with
all of its neural structures (Fig. 3319).
Although not all colloid cysts are resectable using endoscopic techniques, many can easily and safely be removed
with minimal injury to the patient. The authors method
of choice is the two-portal approach to the right frontal
horn. Using the MR image as a guide, two ipsilateral bur
holes are placed further laterally than normal to allow direct

Figure 3316 To gain adequate access to the frontal horn, a standard

coronal bur hole is used. By placing it progressively more lateral, medial pathology will be visualized and approached.

Figure 3317 Diagrammatic fenestration of a lesion located in the

medial wall of the frontal horn. A more laterally located bur hole will
give better exposure to the inferior aspect of the lesion.



14535_C33.indd 222

4/23/08 3:13:51 PM

33 Endoscopic Approaches to the Ventricular System 223

Temporal Horn Access
For access to this area, the patient should be placed supine
with an ipsilateral shoulder roll and extreme lateral rotation of the head. A 1 cm incision is made slightly above the
top of the pinna. Entrance into the temporal horn is gained
through a bur hole located over the middle temporal gyrus.
The exact location of the bur hole is ascertained with MR
images. In the majority of cases, the affected horn will be
dilated and the pathology readily visualized (Fig. 3320). By
varying the angle of approach, access can be gained to the
roof of the temporal horn, the oor, or the lateral wall (Fig.
3321). The choroid plexus will be seen coursing along the
medial superior aspect of the horn.
Figure 3318 The landmarks of the foramen of Monro and lateral
ventricle as seen endoscopically. The fornix makes up the anterior and
superior margins of foramen and the thalamus composes the inferior
posterior margins. The choroid plexus is located posteriorly and the
rostrum of the corpus collusum is seen anteriorly along with the head
of the caudate.

medial access to the cyst. One bur hole should be placed

directly lateral to the foramen of Monro. Through this opening, an endoscope and its working channels can be inserted.
The second bur hole, placed anteriorly, is used to insert a
large bore slotted sucker through a no. 14 French peel-away
sheath into the anterior horn. The sucker should be placed
at high suction. Under direct visualization, the capsule may
be opened using an Nd:YAG (neodymium:yttrium-aluminum-garnet) or KTP (potassium-titanyl-phosphate) laser
ber or a radiofrequency probe. As soon as the gelatinous
cyst contents are visualized, the large sucker is placed next
to the opening and the contents are rapidly aspirated. This
maneuver will prevent the cyst contents from entering the
ventricular system. The capsule can then be removed using
other available endoscopic instruments.

Figure 3319 The head and body of the caudate are visualized resting on top the thalamus. Above the caudate, the crossing bers of the
corpus collusum can be seen.

14535_C33.indd 223

Occipital Horn Access

The occipital horn may be approached with the patient positioned prone or lateral (Figs. 3322 and 3323). The exact
position is the surgeons preference, although lateral positioning seems to provide less surgeon fatigue and makes
ventricular ushing and irrigation easier. Similarly, there
is less chance for debris to settle into the frontal horn and
third ventricle with the aid of gravity. A bur hole is located
in the classic position for occipital access: 3 cm lateral to
the inion and 6 to 7 cm superiorly (Fig. 3324). Again, the
exact location and trajectory should be guided by preoperative MR images. This approach can be superior access to the
anterior tip of the temporal horn, the occipital horn, the
ventricular atrium, and even the frontal horn (Fig. 3325).
Familiarity with the anatomy in this area is crucial. Struc-

Figure 3320 The circles demonstrate the locations of lesions easily accessible using an approach over the middle temporal gyrus. The
roof of the temporal horn will be identied by choroid plexus running
along the ssure.

4/23/08 3:13:52 PM

224 Pediatric Neurosurgery

Figure 3322 Prone position for access to the posterior fossa lesions.
Figure 3321 Entrance into the temporal horn is achieved by placing
a bur hole slightly superior to the pinna of the ear. The cortical incision
is made on the middle temporal gyrus.

tures such as the calcar avis, collateral trigone, and choroid

plexus should be easily identiable. Sometimes the occipital
horn is not fully dilated, which makes a direct occipital approach much more difcult and risky. When the rest of the
ipsilateral ventricular system is dilated and a small occipital
horn is present, it is preferable to approach the lesion from
a frontal approach as seen in Fig. 3316.

Posterior Fossa Access


heavily padded horseshoe (Fig. 3322). The neck should be

placed in maximal allowable exion to aid in the exposure
of this region. Generally, a paramedian incision is made and
a bur hole is placed 2 to 3 cm from the midline (Fig. 3326).
This is particularly important in infants who may have large
midline venous sinuses. The appropriate angle is chosen
for a particular lesion, and access is easily obtained (Fig.
3327). When placing a catheter in a trapped fourth ventricle, only a very small bur hole is needed and a 1.2 mm 10K
scope can be inserted within the ventricular catheter and
pushed forward through a previously made slit at the distal
end of a catheter. In this way, difcult to reach and small

Several pathological conditions occur in the posterior fossa

that may be accessible with endoscopic techniques. Such
lesions include trapped fourth ventricles, arachnoid cysts,
cystic tumors, and cerebral pontine angle lesions. For lesions at or near the midline, the patient should be placed in
the prone position in either three-point rigid xation or on a

Figure 3323 Supine position with 90 degree cranial rotation may be

used for entrance into the temporal horn or occipital horns.

14535_C33.indd 224

Figure 3324 To gain access to the posterior compartment of the

lateral ventricles, the standard occipital bur hole is used.

4/23/08 3:13:54 PM

33 Endoscopic Approaches to the Ventricular System 225

Figure 3325 The circles demonstrate the locations of lesions that

can be reached using an occipital bur hole approach.

midline cysts can be safely and accurately approached with

endoscopic techniques. For lesions of the cerebellopontine
angle, the patient should be placed in a lateral position with
the affected side up. A spinal drain can be inserted into the
lumbar cistern, and when sufcient spinal uid is removed,
the cerebellar hemisphere is allowed to fall away from the

Figure 3326 The endoscope is inserted frontally to access a trigonal

or occipital lesion when the involved occipital horn is not dilated. Although further away, it is easier to maneuver in the frontal and lateral
ventricle that is dilated.

lateral aspect of the skull, thereby exposing the cerebellopontine angle in the superb fashion. In these cases, the
incision is made more lateral and the bur hole is placed
immediately medial to the mastoid tip. Closure of the dura
should be done with a piece of pericranial ap, and if the
mastoid air cells are entered, generous amounts of bone
should be used to prevent postoperative CSF leaks.

Postoperative Management Including

Possible Complications

Figure 3327 Endoscopic localization and biopsy/removal of difcult

to reach midline infratentorial lesions can be safely and easily done
with paramedian bur hole placement.

Although relatively infrequent, complications can and do

occur while approaching the ventricular system with endoscopic methods. Given the steep learning curve associated
with these procedures, mishaps are more likely to occur
early during training. Bleeding is probably the most common and bothersome complication. Even small amounts
of blood within the CSF can rapidly obscure visualization
within a few seconds. The risk of bleeding can be decreased
by using the laser, which will obliterate small bleeders. By
holding the laser ber several millimeters away from the
vessels, it will diffuse the energy and fully obliterate it. Most
bleeding will be venous in nature, and time and irrigation
will lead to clot formation and bleeding cessation. Even if
the entire ventricular system is casted with clotted blood,
the clot can then be safely removed with some diligence.
Grasping forceps or suction can be used to accomplish clot
evacuation. Arterial bleeding on the other hand, if uncontrollable, may require an open craniotomy. Fortunately, this
is not a common occurrence.


14535_C33.indd 225

4/23/08 3:13:55 PM

226 Pediatric Neurosurgery

Acute increase of intracranial pressure can occur if the
outow of irrigant is not carefully matched with the inow. This is particularly true when working inside the third
ventricle. Sudden changes in vital signs should alert the
surgeon to this possibility. Conversely, if too much uid is
removed from the ventricles, especially in a patient with
thin cortical mantle, bilateral subdural hematomas can develop. Theoretically, the infection rate increases with the
increased number of instruments used and with manipulation. However, this appears not to be the case because the
infection rate in the surgical procedures is ~0.8%. Seizures
can be problematic in some patients. However, prophylaxis
has not been done routinely even for two portal procedures.
Due to the large size of the dural opening, large subgaleal
uid collections can occasionally be found if dural closure
is not properly done. A small piece of pericranium can be
used to close the dura and prevent these complications. Endocrine disturbances have been described in third ventriculostomies by some authors who perform the opening of the
oor of the ventricle with the tip of the endoscope. Perhaps
this is due to mechanical injury to the hypothalamus. We
have not seen any endocrine or hypothalamic dysfunction

14535_C33.indd 226

following third ventriculostomies. A dangerous complication that has been seen with increased frequency is injury to
the tip of the basilar artery. Besides the acute intraoperative
hemorrhage, a pseudoaneurysm commonly forms that must
be immediately and adequately diagnosed and treated with
endovascular techniques.

Neuroendoscopic techniques provide the neurosurgeon
with an increased array of techniques and alternatives for
treating simple or complex intraventricular lesions or for restoring adequate CSF ow dynamics. The minimally invasive
nature of these procedures leads to less neural tissue destruction/injury, less patient discomfort, decreased length
of stays, and, ultimately, to decreased hospitalization costs.
However, to achieve excellence results, adequate training,
careful planning and execution, as well as proper patient
selection are not only necessary but mandatory.

4/23/08 3:13:57 PM

Intraventricular Endoscopy
Jonathan P. Miller and Alan R. Cohen

Minimally invasive neurosurgery sounds like an oxymoron

because neurosurgical procedures are, by their very nature,
highly invasive. However, located deep within the substance
of the brain and lled with clear uid, the cerebral ventricles are ideally suited for endoscopic navigation. Recent
improvements in optics and miniaturization now make it
possible to perform certain intracranial procedures through
small exposures with minimal trauma to neural structures.
Endoscopic techniques have been used to facilitate insertion
of ventricular shunt catheters, to fenestrate loculated ventricular cysts, and to inspect, biopsy, and sometimes remove
certain intraventricular cysts and tumors. In some cases of
noncommunicating hydrocephalus, endoscopic fenestration
of the oor of the third ventricle can eliminate the need for
ventricular shunting altogether. The endoscope can also be
introduced through small cranial exposures at the skull base
to allow visualization of structures otherwise hidden from
the operative microscope. In this chapter, the history of neuroendoscopy is briey reviewed, followed by a discussion
of the neurosurgical applications of endoscopy focusing on
operations performed within the cerebral ventricles.

History of Neuroendoscopy
The application of endoscopic techniques to neurosurgical
conditions began with an effort to diagnose and treat hydrocephalus at the turn of the 20th century. The rst recorded
endoscopic neurosurgical procedure took place in 1910
when Victor Darwin Lespinasse (18781946), a urologist in
Chicago, used a cystoscope to fulgurate the choroid plexus
in two hydrocephalic infants. Eight years later, Walter Edward Dandy (18861946) used the instrument for the same
purpose, and in 1922 was the rst to use the term ventriculoscope. Both Lespinasse and Dandy reported discouraging results, and ventriculoscopy was all but abandoned at
that time because of high surgical morbidity rates and the
development of better indirect means of imaging intracranial contents. Ultimately, valved shunts were developed to
effectively treat hydrocephalus, and ventricular endoscopy
almost became a historical curiosity.
As experience with ventricular shunts increased, however,
it became apparent that ventricular shunts were not as safe
and effective as initially thought. Ventricular shunts remain
fraught with problems related to malfunction and infection,
and shunt surgery is today associated with a higher rate of

complications than any other commonly performed neurosurgical procedure. The current resurrection of interest in
neuroendoscopy was fueled initially by efforts to simplify or
eliminate ventricular shunt systems. Recent technical breakthroughs in optical design have made it possible to approach
even deep-seated ventricular tumors via the endoscope. A
variety of neurosurgical endoscopic procedures have been
developed, such as endoscopic third ventriculostomy, cyst
fenestration, retrieval of adherent ventricular catheters, and
biopsy or removal of ventricular tumors.

Several of the innovative contributions to the eld of neuroendoscopy were pioneered by Harold H. Hopkins, a British
physicist working at the University of Reading. In 1960, he
improved on the conventional endoscope system initially developed in Vienna by Nitze in 1887. Instead of a train of biconvex lenses spaced out in an air-containing metal tube, which
is optically inefcient, Hopkins used a solid rod lens system
that results in a signicantly smaller diameter instrument
with a wider viewing angle, improved light transmission,
improved image resolution, and reduced chromatic aberration. Hopkins also developed the coherent ber-optic bundle
system that serves as the basis for the exible endoscope by
wrapping a small quartz ber many times around a drum and
then making a single cut across all the bers, thereby leaving
them identical in length and location at each end. By ensuring
that each ber at one end was identical in position to its cut
counterpart at the other end, Hopkins was able to create a