Advances in Spinal Stabilization PDF

dramroo
Advances in Spinal Stabilization
dramroo
Progress in Neurological
Surgery
Vol. 16
Series Editor
L. Dade Lunsford Pittsburgh, Pa.
dramroo
Advances in Spinal
Stabilization
Volume Editors
Regis W. Haid, Jr. Atlanta, Ga.

Brian R. Subach Atlanta, Ga.
Gerald E. Rodts, Jr. Atlanta, Ga.
145 figures, 2 in color, and 17 tables, 2003
Basel Freiburg Paris London New York

Bangalore Bangkok Singapore Tokyo Sydney
dramroo
Regis W. Haid, Jr., MD
Emory University School of Medicine
Department of Neurosurgery
The Emory Clinic
Atlanta, GA 30322 (USA)
Library of Congress Cataloging-in-Publication Data
Advances in spinal stabilization / volume editors, Regis W. Haid, Jr., Brian R. Subach,
Gerald E, Rodts, Jr.
p. ; cm. (Progress in neurological surgery, ISSN 00796492 ; vol. 16)
Includes bibliographical references and index.
ISBN 3805574312 (hard cover : alk. paper)
1. SpineSurgery. 2. Spinal implants. I. Haid, Regis W. II. Subach, Brian R. III. Rodts,
Gerald E. IV. Series.
[DNLM: 1. Spinesurgery. 2. Orthopedic Proceduresmethods. 3. Spinal
Fusionmethods. WE 725 A2447 2003]
RD533 A285 2003
617.56059dc21 2003047736
Bibliographic Indices. This publication is listed in bibliographic services, including Current Contents and
Index Medicus.
Drug Dosage. The authors and the publisher have exerted every effort to ensure that drug selection and
dosage set forth in this text are in accord with current recommendations and practice at the time of publication.
However, in view of ongoing research, changes in government regulations, and the constant flow of information
relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for
any change in indications and dosage and for added warnings and precautions. This is particularly important
when the recommended agent is a new and/or infrequently employed drug.
All rights reserved. No part of this publication may be translated into other languages, reproduced or
utilized in any form or by any means electronic or mechanical, including photocopying, recording, microcopying,
or by any information storage and retrieval system, without permission in writing from the publisher.
Copyright 2003 by S. Karger AG, P.O. Box, CH4009 Basel (Switzerland)

www.karger.com
Printed in Switzerland on acid-free paper by Reinhardt Druck, Basel
ISSN 00796492
ISBN 3805574312
dramroo
Contents
VIII Series Editors Note

Lunsford, L.D. (Pittsburgh, Pa.)
IX Foreword
Sonntag, V.K.H. (Phoenix, Ariz.)
Biological Advances
1 Bone Morphogenetic Protein (rhBMP-2): Experimental Review

and Clinical Update
Subach, B.R.; Haid, R.W., Jr.; Rodts, G.E., Jr.;
Petraglia, C.A. (Atlanta, Ga.)
14 Cellular and Genetic Approaches for Spinal Fusions
Helm, G.A.; Anderson, D.G. (Charlottesville, Va.)
29 Bone Substitutes
Swartz, K.R.; Trost, G.R. (Madison, Wisc.)
39 Resorbable Technology for Spinal Stabilization
Resnick, D.K. (Madison, Wisc.); Alexander, J.T. (Winston-Salem, N.C.);
Welch, W.C. (Pittsburgh, Pa.)
55 Lumbar Interbody Fusion Using Bone Morphogenetic Protein:
Results and Fusion Assessment
Burkus, J.K. (Columbus, Ga.); Foley, K.T. (Memphis, Tenn.);
Haid, R.W., Jr. (Atlanta, Ga.)
dramroo
Neuronavigational Advances
71 Overview of Spinal Navigation

McLaughlin, M.R.; Bartolomei, J. (Lawrenceville, N.J.)
84 Virtual Fluoroscopy: Overview and Future Implications
Rampersaud, Y.R. (Toronto); Foley, K.T. (Memphis, Tenn.)
96 Placement of Thoracic Pedicle Screws
Polly, D.W., Jr.; Kuklo, T.R. (Washington, D.C.)
107 Cervical Techniques with Image-Guided Spinal Navigation
Kalfas, I.H. (Cleveland, Ohio)
Instrumentation and Technique Update
128 C1 Lateral Mass Fixation

Fiore, A.J. (Atlanta, Ga.); Birch, B.D. (Scottsdale, Ariz.);
Haid, R.W., Jr. (Atlanta, Ga.)
142 Cervical Laminoplasty
Casey, A.T.H. (London/Stanmore); Crockard, H.A. (London)
154 Cervical Pedicle Screws: Advances in Spinal Stabilization
Chun, J.Y. (San Francisco, Calif.); Mummaneni, P.V. (Atlanta, Ga.);
Birch, B.D. (Scottsdale, Ariz.); Sasso, R.C. (Indianapolis, Ind.)
165 Cervical Lateral Mass Advances
Mummaneni, P.V. (Atlanta, Ga.); Traynelis, V.C. (Iowa City, Iowa);
Sasso, R.C. (Indianapolis, Ind.)
176 Interbody Carbon Fiber
Bertagnoli, R. (Straubing)
188 Thoracic Pedicle Screw Placement
Lenke, L.G. (St. Louis, Mo.)
204 Percutaneous Lumbar Pedicle Screws: Indications,
Technique, Results
Rodts, G.E., Jr. (Atlanta, Ga.)
213 Thoracolumbar Deformity Advances.
1. Nonoperative Treatment of Thoracolumbar Deformity
Mummaneni, P.V. (Atlanta, Ga.); Ondra, S.L. (Chicago, Ill.);
225 Thoracolumbar Deformity Advances.
2. Operative Treatment of Thoracolumbar Deformity
Mummaneni, P.V. (Atlanta, Ga.); Ondra, S.L. (Chicago, Ill.);
Contents VI
dramroo
240 Vertebral Augmentation for Osteoporotic and Osteolytic
Vertebral Compression Fractures: Vertebroplasty
and Kyphoplasty
Lieberman, I. (Cleveland, Ohio)
Minimally Invasive Update
251 Endoscopic Posterior Cervical Foraminotomy

and Microdiscectomy
Perez-Cruet, M.J.; Fessler, R.G. (Chicago, Ill.)
266 An Update on Endoscopic Thoracic Spinal Surgery:
Thoracic Microendoscopic Discectomy
Perez-Cruet, M.J.; Kim, B.-S.; Sandhu, F.; Fessler, R.G. (Chicago, Ill.)
277 Laparoscopic versus Mini-Open Anterior Lumbar
Interbody Fusion
Kaiser, M.G.; Haid, R.W., Jr.; Subach, B.R.; Rodts, G.E., Jr. (Atlanta, Ga.)
Socio-Economic Evolution
290 Epidemiology and Variations in Care of Spine Disease

Burke, J.P.; Gerszten, P.C. (Pittsburgh, Pa.)
302 The Changing Economics of Spine Surgery
Przybylski, G.J. (Orange, N.J.)
314 Author Index
315 Subject Index
Contents VII
dramroo
Series Editors Note
Three eminent spine surgeons have collaborated as editors for this volume
of Progress in Neurological Surgery. There is no question that new spinal sta
bilization techniques, minimally invasive approaches to the spine, and new con
cepts of biomechanics have revolutionized our understanding of disorders of
spinal stability. In this comprehensive volume, the authors describe various
techniques and clinically relevant procedures designed to improve the outcomes
of spine surgery. Neuronavigational techniques, new bone fusion concepts, and
both open and percutaneous spinal stabilization techniques are described.
Progress in Neurological Surgery is dedicated to providing timely updates of
important neurosurgical paradigm shifts. The volume of spinal surgery across
the world has dramatically increased. The chapters in volume 16 help to eluci
date the role and rationale of these new kinds of spinal stabilization. I am grate
ful to the editorial efforts of Dr. Haid, Dr. Subach and Dr. Rodts and to all the
authors.
L. Dade Lunsford, MD
VIII
dramroo
Foreword
Spinal stabilization has changed dramatically over the last 10 years and the
pace of change continues to accelerate. This volume is an excellent mirror of
the evolution of spinal stabilization from the nuts, rods, and bolts first used
years ago to the current applications of bilateral fusion using bone morpho
genetic protein and gene manipulation or sophisticated bone extenders. Further
possibilities for stabilizing the spine, as reflected in this book, include
absorbable stabilization devices.
Through the combination of technical advances in minimally invasive
spine surgery and neuronavigation, spinal surgery will become less destructive,
less painful, and more successful. Most importantly, patient outcomes will
improve. The section on instrumentation and technique represents the most
up-to-date advances in surgical technique and management of spinal disorders.
The authors have provided a valuable service by compiling the most recent
advances in spinal stabilization. This volume is a welcome addition to the per
sonal libraries of spinal surgeons and to institutional libraries.
V.K.H. Sonntag
IX
dramroo
This page intentionally left blank
dramroo
Biological Advances
Haid RW Jr, Subach BR, Rodts GE Jr (eds): Advances in Spinal Stabilization.
Prog Neurol Surg. Basel, Karger, 2003, vol 16, pp 113
Bone Morphogenetic Protein
(rhBMP-2): Experimental Review
and Clinical Update
Brian R. Subach, Regis W. Haid, Jr., Gerald E. Rodts, Jr.,

Carmen A. Petraglia
The NeuroSpine Institute, Department of Neurological Surgery,
Emory University, Atlanta, Ga., USA
Over 30 years ago, Urist et al. [67, 68] identified a group of protein
extracts, derived from the ground substance of mature bovine bone, capable of
inducing both cartilage and bone formation when implanted into the soft tissues
of study animals. Aptly named, bone morphogenetic proteins (BMPs) by Urist,
these glycoproteins comprise a subset of the transforming growth factor-
family of related growth and differentiation factors. Of the more than 20 BMPs
isolated to date, 6 appear to be structurally related to each other and capable of
initiating the process of endochondral bone formation. The presence of such
factors within the matrix of mature bone indicates a likely role in the regenera
tion and remodeling of bony structures after injury or repetitive stresses [4, 9,
18, 24, 43, 44].
The Basics of BMP Biology
Each of the six known osteoinductive BMPs shares significant similarities

on a molecular level. Synthesized within the cell in precursor form, each mol
ecule has a hydrophobic leader or secretory sequence with the mature portion
of the protein at the carboxy terminus marked by a highly conserved, seven-
cysteine repeat. Each mature BMP begins as 2 monomers of 120 amino acids each,
which undergo disulfide linkage dimerization to form either a homologous or
a heterologous protein chain. In the case of BMP-2 and BMP-7, the specific
structure was first identified by isolating the bovine protein from bone extracts.
dramroo
Oligonucleotide probes were used to obtain the human complementary DNA
(cDNA) sequence. The cDNA clones were then spliced into a viral expression
vector and transfected into a carrier cell in a process called recombination. In
the case of BMP-2, the cells used were Chinese hamster ovary cells. Such cells
produce the pure recombinant differentiation factor rhBMP-2 in large quanti
ties in a process similar to fermentation. This process avoids potential compli
cations related to the transmission of infectious materials from human donor
bone tissue and eliminates the possibility of xenograft interactions with human
recipients of BMP derived from bovine sources.
An Osteoinductive Role for BMP
In order to function as a suitable graft for bridging bone defects or fusing

fracture lines and unstable motion segments, the prospective material would
ideally possess three characteristics. The material would provide a source of
primitive osteoprogenitor cells, which, under the appropriate influence, would
form osteoblasts and osteocytes (osteopromotive). Such precursor cells are
unfortunately relatively scarce. Bone marrow, for example, contains a ratio of
only one osteprogenitor cell to approximately 50,000 nucleated cells in a young
adult. This ratio may dip to 1:200,000 cells in an elderly individual afflicted
by degenerative spinal disease [21]. Despite techniques to concentrate marrow
extracts, successful efforts have only resulted in a maximum of 5-fold improve
ment of the unfavorable cellular ratio. Second, the graft material would produce
local growth factors to stimulate bone growth and vascularity in the area
(osteoinductive). There are numerous reports in the literature detailing the com
plex interaction of various autocrine and paracrine growth factors released from
fibroblasts, platelets, and even the local hematoma that forms at the site of
injury [1, 7, 16, 37, 46]. Finally, the third property of the graft material ideally
would be its ability to act as a scaffold for bony ingrowth. Such ability is known
as osteoconduction.
There are a number of possible reasons, which may account for the osteo
inductive role of BMP. BMP acts as a chemotactic agent, a growth factor, and a
differentiation factor. BMP acts as a chemotactic factor and initiates the recruit
ment of progenitor and stem cells toward the area of bone injury. In vivo stud
ies of the local effects of BMP indicate an initial migration of mesenchymal
stem cells to the area of implantation far in excess of that supplied by bone mar
row grafting [12]. BMP acts as a growth factor and stimulates both angiogene
sis and the proliferation of stem cells from surrounding mesenchymal tissues.
BMP also acts as a differentiation factor by promoting the maturation of stem
cells into chondrocytes, osteoblasts, and osteocytes.
Subach/Haid/Rodts/Petraglia 2
dramroo
Some cells respond to the growth factor aspect of BMP by altering their
rates of proliferation. Yamaguchi et al. [71] demonstrated this in vitro by quan
tifying cellular proliferation of the rat C26 calvarial osteoprogenitor cells after
treatment with BMP-2. This BMP-2 effect, however, appears to maintain speci
ficity for certain cell types. For example, although BMP-7 has been shown to
be mitogenic for a human osteosarcoma cell line (TE-85), treatment with BMP-2
showed no measurable effect on proliferation. In contrast, treatment of an
osteoblast cell line (MC3T3-E1) with BMP-4 results in an inhibition of growth
and a globally diminished proliferative index. By the effects on both mature and
immature cell types, it seems reasonable that BMPs must be involved in the
regulation of bone growth and maintenance of bone structure.
BMPs may also initiate the differentiation of stem cells into a specific
phenotype. For example, the rat calvarial stem cell line (C26) is considered
multipotential, in that such cells may be precursors for adipocytes, muscle cells,
or osteoblasts. When BMP-2 is added to the culture medium, such cells become
mature osteoblasts with increased surface expression of receptors for parathy
roid hormone, alkaline phosphatase, and calcitonin [71]. This effect may also
be observed in bone marrow cells. For example, the mouse line of marrow cells
(W-20-17) may differentiate into either adipocytes or osteoblasts, depending
upon the specific hormonal influence. The BMP-2 treatment of such cells
results in both the differentiation of the cells into osteoblasts and the surface
expression of receptors normally seen on mature cells.
Sources of BMP
At present there are three ways to obtain bone growth and differentiation fac
tors: extraction of the factors from animal or human bone matrix, production of a
single factor by cellular hosts using recombinant technology, and direct delivery of
the DNA encoding for the factor to cells at the site of desired bone formation.
The first of these was initially employed by Urist et al. [67, 68]. From
massive quantities of bovine bone, the group was able to extract a mixture of
proteins found to stimulate bone growth in vivo. Under clinical evaluation in
Europe as NeOsteo (Sulzer Spinetech, Wheat Ridge, Colo., USA), this mixture
of BMPs and other associated proteins is derived through a well-engineered
isolation process. The precise combination of factors comprising this substance
has not yet been fully characterized, but appears to be reproducible through the
manufacturers process. This substance has shown experimental promise in
bridging both segment skeletal defects in dogs and in bringing about spinal
fusion in animal models of posterolateral arthrodesis. [12, 14, 17, 26, 27]. Like
other growth and differentiation factors, a carrier substance is necessary to
BMP Update 3
dramroo
maintain adequate concentrations at the site of fusion. Substances such as
natural coral (hydroxyapatite), collagen, and calcium sulfate have each been
investigated [26, 27].
The second method of obtaining bone growth and differentiation factors
has previously been discussed (see Molecular Biology). The process of obtain
ing recombinant human BMPs such as rhBMP-2 (Medtronic Sofamor Danek,
Memphis, Tenn., USA and Genetics Institute, Cambridge, Mass., USA) and
rhBMP-7 (Stryker Biotech, Hopkinton, Mass., USA) has been described. Such
proteins differ from mixtures of extracted substances, mainly in terms of purity
of product. Original studies of these substances focused upon animal models of
segmental bone defects in the appendicular skeleton of rats, sheep, and dogs
[23, 24, 35]. Cole et al. [23] compared rhBMP in a carrier to autologous graft
ing in a skeletal defect model with considerable success. Gerhart et al. [35]
showed the utility of rhBMP in healing segmental femoral defects in sheep.
Shortly thereafter, recombinant BMP was applied to animal models of spinal
fusion and later, humans.
The third strategy for engineering bone formation involves gene therapy, or
the delivery of the appropriate gene, or cDNA encoding for BMP to the local
cells, rather than the actual factor. There are two obvious benefits to the strategy
as compared to recombinant technology. First, the cost of genetic manipulation
is significantly less than that required to both produce and market the purified
rhBMP. Second, the potential for prolonged local production of the factor is
greater with gene therapy when compared to the relatively short-lived effect of
the rhBMP/carrier complex. Attempts to introduce BMP-2 cDNA into animal
models are preliminary, but have met with limited success [1, 17, 29, 70].
From the Laboratory
Early work in the isolation of proteins with osteoinductive activity suggested

that BMP-2 and BMP-7 were primarily responsible for the effects observed
in vivo [3, 22, 23, 53, 59]. As a result, rhBMP-2, produced in a Chinese hamster
ovary cell line, was the first of these molecules studied in detail. Implantation of
the recombinant factor in a rat model resulted in ectopic bone formation with a
dose-effect relationship temporally identical to that of bone-derived extracts;
however, the amount of pure rhBMP-2 required to induce formation of a given
amount of bone was approximately 10-fold less than that required of the bone
extract [5, 12, 13, 51]. Subsequent studies in nonhuman primates show no differ
ence in the dose required to effect consistent posterolateral spinal fusions.
The ability to form bone at ectopic sites, however, had little application to
current spinal fusion techniques. Realizing the limitations inherent in
dramroo
a b
Fig. 1. Photomicrograph of intertransverse spinal fusion using carrier matrix alone (a)
as a control and rhBMP-2/carrier (b). Histologic section demonstrates transverse processes
at inferior-lateral corners with bridging collagen scar tissue and minimal bone formation (a)
and abundant new membranous bone formation (b).
autogenous and allogeneic bone grafting, investigators began applying BMP

technology to animal models of spinal fusion procedures [32, 36, 38, 62]. Multiple
studies, involving various concentrations of BMP in a variety of carrier sub
strates, have shown remarkable results. Early work by Boden et al. [1215] and
Holliger et al. [45] compared rhBMP-2 to autologous bone graft in a rabbit pos
terolateral lumbar fusion model. Remarkably, all BMP-treated animals (100%)
attained solid, bony fusions across the operated level, which were biomechani
cally stiffer and stronger than the autograft-only fusions observed in 42% of the
control group. Similar studies in a canine model also confirmed the efficacy of
rhBMP-2 in producing mature fusion masses [28]. The canine study by David
et al. [28] demonstrated a dose dependence to the BMP effect, with greater
concentrations producing greater effects; however, this contradicts a study by
Sandhu et al. [5860] in a similar canine model, which shows BMP to be more
effective than autologous bone graft, but in a dose-independent manner [63].
Most investigators developed the opinion that bone induction was relatively
simple in lower species, but was only indirectly applicable to human models.
As a result, research focused on developing spinal fusion models in primates.
As a developmentally higher species, primates provide a more realistic test
environment for evaluating the effectiveness of BMP [40, 41]. Boden et al. [10]
applied this belief to a nonhuman primate model of intertransverse spinal
fusion and demonstrated effective fusion rates using rhBMP-2 on a ceramic
carrier delivered by a minimally invasive approach (fig. 1). Sandhu et al. [60]
went a step further to demonstrate clinically, mechanically, and radiographically
equivalent spinal arthrodesis using rhBMP-2 without decortication of the
prospective fusion bed. Boden et al. [7, 16] and Martin et al. [51], by adding
rhBMP-2 to autograft, were able to demonstrate subsequent induction of BMP-6,
osteocalcin, and collagen within the graft itself.
BMP Update 5
dramroo
Control 0.75 mg/ml 1.50 mg/ml
Fig. 2. Threaded titanium interbody cage implanted into the lumbar spine of a primate.
Control group is cage alone. Experimental groups are rhBMP-2-impregnated sponges at con
centrations of 0.75 and 1.50 mg/ml. Actual spines at 6 months from surgery. Control demon
strates a pseudarthrosis with fibrous scar tissue within the cage. BMP groups both show
mature bone bridging the interspace within the cage.
Posterolateral fusion models attempted to replace autograft with a BMP/

carrier complex, but still required internal fixation. Attention was then focused
upon interbody spinal fusion techniques, which could possibly obviate the
need for both autograft and fixation [19, 20, 34, 42, 47, 48, 55, 56, 64, 65].
Zdeblick et al. [73] performed three level anterior cervical fusions in goats
using a BAK cage filled with either local autograft or a collagen impregnated
rhBMP-2 sponge [49]. Eleven of 21 animals (52%) in the autograft group had
histologic evidence of pseudarthrosis, while only 1 (5%) of the BMP group
failed to form a solid bony fusion. The biomechanical stiffness of the BMP
construct was equal to that of an autograft/cervical-plated level. Boden et al. [8]
performed the same procedure in the lumbar spine of primates. Using rhBMP-2
on a collagen carrier in both titanium-threaded interbody cages and
threaded bone dowels, both were delivered laparoscopically with a docu
mented improvement of fusion rates over empty cages and autograft-filled
cages [8, 57] (fig. 2, 3).
Preliminary results of human clinical studies have been encouraging. In a
recent report of a randomized, prospective controlled clinical pilot study, Boden
et al. [8] demonstrated solid bony fusions by both clinical and radiographic
dramroo
a b
c d
Control rhBMP-2
Fig. 3. Bone dowel interbody device implanted into the lumbar spine of a primate.
Control group is dowel with autograft. Experimental group is a bone dowel with an rhBMP-2
impregnated sponge inside. Actual spines at 6 months from surgery. Control group demon
strates a solid fusion (a) and reabsorption of the bone dowel/autograft construct (c). BMP
animals (b, d) both show mature bone bridging the interspace with resorption of the dowel.
criteria in 11 patients undergoing anterior lumbar interbody fusion procedures

with a tapered, threaded titanium cage filled with an rhBMP-2-impregnated
collagen sponge (fig. 4). Pain scores, as documented by the Oswestry Disability
and the Short Form-36 questionnaires, improved concomitantly as fusion
progressed [8, 30, 39, 56, 66, 69].
Based upon the findings in both the primate models and human trials, the
United States Food and Drug Administration (FDA) approved rhBMP-2 for lim
ited use in patients with spinal disorders. Specifically, the FDA approved BMP
for use as an adjunct to spinal fusion in patients with degenerative disc disease
undergoing anterior lumbar interbody fusion using titanium cages (LT cages).
BMP Update 7
dramroo
Patient 11 rhBMP-2
c
6 months 12 months 24 months
Fig. 4. Computed tomographic reconstructions of a human patient receiving a

threaded, titanium cage (Lordotec cage, Medtronic Sofamor Danek, Memphis, Tenn., USA)
with rhBMP-2 on a collagen sponge carrier. a Sagittal view of the left cage at 6, 12 and 24
months. b Right cage at the same time intervals. c Coronal views at the same time intervals.
All images show fusion with increasing bone density over time.
The cages attempt to both restore and maintain intervertebral height and protect
the BMP from exposure to diluting substances such as blood and irrigation fluid.
Although approved, BMP has yet to reach the market. Concerns raised by
experts include the improper use of the BMP such that vascular and neural ele
ments may come in contact with the protein causing injury, lower fusion rates
due to improper or unapproved implantation techniques, and stimulation of
infectious or neoplastic processes due to use in patients with such diseases.
Discussion
Research over the past decade has shown the utility of using the growth
and differentiation factor, BMP-2, to promote bone formation at the site of bone
loss or injury. The in vivo role of BMP-2 and its complex interaction with other
dramroo
growth and differentiation factors remains to be clarified. The use of BMP-2 as
a means of replacing harvested autograft and obviating the need for internal fix
ation, each with its attendant morbidity, appears likely as a result of outcomes
from both animal and human studies. Although dose-effect relationships and
carrier substrates may provide continued investigational challenges, the use of
recombinant technology and gene therapy in the field of bone fusion have been
firmly established. In the past 30 years since Marshall Urist first coined the
term bone morphogenetic protein, one doubts that he could have envisioned the
monumental strides and clinical progress, which researchers in the field have
achieved to this point.
Conclusions
The widespread use of spinal fusion procedures in the management of

spinal disorders has led investigators to explore the use of growth and differen
tiation factors in such procedures. Either as an adjunct to allograft bone or as a
replacement for harvested autograft, BMPs appear to improve fusion rates after
spinal arthrodesis in both animal and human models, while reducing the donor
site morbidity previously associated with such procedures [2, 6, 11, 14, 23, 25,
50, 54, 72]. The use of recombinant genetic technology in the production of
BMP has improved the efficiency, cost-effectiveness, and safety of producing
and using such materials. rhBMP-2, as one of the first factors identified in the
process of endochondral bone formation, has been extensively researched over
the past decade. The efficacy and dose profile of this differentiation factor in
the context of various carrier substrates has been investigated [15, 26, 27, 33,
51, 58, 61]. Based upon the encouraging results of preliminary studies, the
future role of rhBMP-2 may lie in the elimination of autologous bone grafting,
the reduction of the need for instrumented fixation, and augmentation of
accepted fusion rates.
Acknowledgments
The authors wish to thank Medtronic Sofamor Danek for their continued support and
assistance in the preparation of the figures.
References
1 Aronow MA, Gerstenfeld LC, Owen TA, Tassinari MS, Stein GS, Lian JB: Factors that promote
progressive development of the osteoblast phenotype in cultured rat calvarial cells. J Cell Physiol
1990;143:213221.
BMP Update 9
dramroo
2 Arrington ED, Smith WJ, Chambers HG, Bucknell AL, Davino NA: Complications of iliac crest
bone graft harvesting. Clin Orthop 1996;329:300309.
3 Asahina I, Sampath TK, Nishimura I, Hauschka PV: Human osteogenic protein-1 induces both
chondroblastic and osteoblastic differentiation of osteoprogenitor cells derived from newborn rat
calvaria. J Cell Biol 1993;123:921933.
4 Aspenberg P, Jeppsson C, Wang JS, Bostrom M: Transforming growth factor beta and bone mor
phogenetic protein 2 for bone ingrowth: A comparison using bone chambers in rats. Bone 1996;
19:499503.
5 Aspenberg P, Turek T: BMP-2 for intramuscular bone induction: Effect in squirrel monkeys is
dependent on implantation site. Acta Orthop Scand 1996;67:36.
6 Banwart JC, Asher MA, Hassanein RS: Iliac crest bone graft harvest donor site morbidity: A sta
tistical evaluation. Spine 1995;20:10551060.
7 Boden SD, Hair G, Titus L, Racine M, McCuaig K, Wozney JM, Nanes MS: Glucocorticoid-
induced differentiation of fetal rat calvarial osteoblasts is mediated by BMP-6. Endocrinology
1997;138:29202928.
8 Boden SD, Martin GJ, Horton WC, Truss TL, Sandhu HS: Laparoscopic anterior spinal arthrod
esis with rhBMP-2 in a titanium interbody threaded cage. J Spinal Disord 1998;11:95101.
9 Boden SD, McCuaig K, Hair G, Racine M, Titus L, Wozney JM, Nanes MS: Differential effects
and glucocorticoid potentiation of bone morphogenetic protein action during rat osteoblast differ
entiation in vitro. Endocrinology 1996;137:34013407.
10 Boden SD, Moskovitz PA, Morone MA, Toribitake Y: Video-assisted lateral intertransverse
process arthrodesis: Validation of a new minimally invasive lumbar spinal fusion technique in the
rabbit and nonhuman primate (rhesus) models. Spine 1996;21:26892697.
11 Boden SD, Schimandle JH: Biologic enhancement of spinal fusion. Spine 1995;20(suppl):
S113S123.
12 Boden SD, Schimandle JH, Hutton WC: 1995 Volvo Award in basic sciences. The use of an
osteoinductive growth factor for lumbar spinal fusion. II. Study of dose, carrier, and species. Spine
1995;20:26332644.
13 Boden SD, Schimandle JH, Hutton WC: Lumbar intertransverse-process spine arthrodesis using a
bovine-derived osteoinductive bone protein. J Bone Joint Surg Am 1995;77:14041417.
14 Boden SD, Schimandle JH, Hutton WC: Evaluation of a bovine-derived osteoinductive bone
protein in a non-human primate model of lumbar spinal fusion. Trans Orthop Res Soc 1996;
21:118.
15 Boden SD, Schimandle JH, Hutton WC, Damien CJ, Benedict JJ, Baranowski C, Collier S: In vivo
evaluation of a resorbable osteoinductive composite as a graft substitute for lumbar spinal fusion.
J Spinal Disord 1997;10:111.
16 Boden SD, Sumner DR: Biologic factors affecting spinal fusion and bone regeneration. Spine
1995;20(suppl):102S112S.
17 Boden SD, Titus L, Hair G, Liu Y, Viggeswarapu M, Nanes MS, Baranowski C: 1998 Volvo Award
in basic sciences: Lumbar spine fusion by local gene therapy with a cDNA encoding a novel
osteoinductive protein (LMP-I). Spine 1998;23:24862492.
18 Boonen S, Broos P, Dequeker J, Bouillon R: The prevention or treatment of age-related osteo
porosis in the elderly by systemic recombinant growth factor therapy (rhIGF-I or rhTGF beta):
A perspective. J Intern Med 1997;242:285290.
19 Burns BH: An operation for spondylolisthesis. Lancet 1933;224:i233.
20 Calandruccio RA, Benton BF: Anterior lumbar fusion. Clin Orthop 1964;35:6368.
21 Caplan AI: Mesenchymal stem cells. J Orthop Res 1991;9:641650.
22 Centrella M, McCarthy TL, Canalis E: Current concepts review: Transforming growth factor-beta
and remodeling of bone. J Bone Joint Surg Am 1991;73:14181428.
23 Cole BJ, Yasko A, Browne MA: The ability of recombinant human bone morphogenetic protein to
heal segmental bone defects A comparison of physiologic carriers to autologous bone graft.
J Bone Miner Res 1992;7(suppl):S251.
24 Cook SD, Baffes GC, Wolfe MW, Sampath TK, Rueger DC, Whitecloud TS 3rd: The effect of
recombinant human osteogenic protein-1 on healing of large segmental bone defects. J Bone Joint
Surg Am 1994;76:827838.
dramroo
25 Cook SD, Dalton JE, Tan EH, Whitecloud T, Rueger DC: In vivo evaluation of recombinant human
osteogenic protein (rhOP-1) implants as a bone graft substitute for spinal fusions. Spine 1994;19:
16551663.
26 Damien CJ, Christel PS, Benedict JJ, Patat J-L, Guillemin G: A composite of natural coral, colla
gen, bone protein, and basic fibroblast growth factor tested in a rat subcutaneous model. Ann Chir
Gynaecol 1993;82:117128.
27 Damien CJ, Parsons JR, Benedict JJ, Weisman DS: Investigation of a hydroxyapatite and calcium
sulfate composite supplemented with an osteoinductive factor. J Biomed Mater Res 1990;24:
639654.
28 David SM, Murakami T, Tabor OB, Grub HE, Meyer RA Jr, Howard BA, Wozney JM, Hanley EN Jr:
Lumbar spinal fusion using recombinant human bone morphogenetic protein-2 (rhBMP-2): A ran
domized, blinded and controlled study. Trans In Soc Study Lumbar Spine 1995;22:14.
29 Delloye C, Verhelpen M, DHemricourt J, Govaerts B, Bourgois R: Morphometric and physical
investigations of segmental cortical bone autografts and allografts in canine ulnar defects. Clin
Orthop 1992;282:273292.
30 Fairbank JCT, Davies JB, Mbaot JC, OBrien JP: The Oswestry low back pain disability ques
tionnaire. Physiotherapy 1980;66:271273.
31 Fang J, Zhu YY, Smiley E, Bonadio J, Rouleau JP, Goldstein SA, McCauley LK, Davidson BL,
Roessler BJ: Stimulation of new bone formation by direct transfer of osteogenic plasmid genes.
Proc Natl Acad Sci USA 1996;93:57535758.
32 Fernyhough JC, Schimandle JH, Weigel MC, Edwards CC, Levine AM: Chronic donor site pain
complicating bone graft harvesting from the posterior iliac crest for spinal fusion. Spine 1992;
17:14741480.
33 Fischgrund JS, James SB, Chabot MC, Hankin R, Herkowitz HN, Wozney JM, Shirkhoda A:
Augmentation of autograft using rhBMP-2 and different carrier media in the canine spine fusion
model. J Spinal Disord 1997;10:467472.
34 Fraser RD: Interbody, posterior, and combined lumbar fusions. Spine 1995;20(suppl):
S167S177.
35 Gerhart TN, Kirker-Head CA, Kriz MJ, Holtrop ME, Hennig GE, Hipp J, Schelling SH, Wang E:
Healing segmental femoral defects in sheep using recombinant human bone morphogenetic pro
tein. Clin Orthop 1993;293:317326.
36 Goldberg VM, Stevenson S: Natural history of autografts and allografts. Clin Orthop 1987;225:
716.
37 Goldberg VM, Stevenson S, Shaffer JW: Biology of autografts and allografts; in Friedlaender GE,
Goldberg VM (eds): Bone and Cartilage Allografts. Park Ridge, American Academy of Orthopaedic
Surgeons, 1991, pp 312.
38 Goulet JA, Senunas LE, DeSilva GL, Greenfield MLVH: Autogenous iliac crest bone graft. Clin
Orthop 1997;339:7681.
39 Grevitt M, Khazim R, Webb J, Mulholland R, Shepperd J: The short form-36 health survey ques
tionnaire in spine surgery. J Bone Joint Surg Br 1997;79:4852.
40 Hanisch O, Tatakis DN, Boskovic MM, Rohrer MD, Wikesjo UM: Bone formation and reosseoin
tegration in peri-implantitis defects following surgical implantation of rhBMP-2. Int J Oral
Maxillofac Implants 1997;12:604610.
41 Hanisch O, Tatakis DN, Rohrer MD, Wohrle PS, Wozney JM, Wikesjo UM: Bone formation and
osseointegration stimulated by rhBMP-2 following subantral augmentation procedures in nonhu
man primates. Int J Oral Maxillofac Implants 1997;12:785792.
42 Harmon PH: Anterior extraperitoneal lumbar disk excision and vertebral fusion. Clin Orthop
1960;18:169198.
43 Heiple KG, Chase SW, Herndon CH: A comparative study of the healing process following
different types of bone transplantation. J Bone Joint Surg Am 1963;45:15931616.
44 Hogan BLM: Bone morphogenetic proteins: Multi-functional regulators of vertebrate develop
ment. Genes Dev 1996;10:15801594.
45 Holliger EH, Trawick RH, Boden SD, Hutton WC: Morphology of the lumbar intertransverse
process fusion mass in the rabbit model: A comparison between two bone graft materials rhBMP-2
and autograft. J Spinal Disord 1996;9:125128.
BMP Update 11
dramroo
46 Kim SS, Denis F, Lonstein JE, Winter RB: Factors affecting fusion rate in spondylolisthesis. Spine
1990;15:977984.
47 Kozak JA, Heilman AE, OBrien JP: Anterior lumbar fusion operations: Technique and graft mate
rials. Clin Orthop 1994;300:4551.
48 Kumar A, Kozak JA, Doherty BJ, Dickson JH: Interspace distraction and graft subsidence after
anterior lumbar fusion with femoral strut allograft. Spine 1993;18:23932400.
49 Kuslich SD, Ulstrom CL, Griffith SL, Ahern JW, Dowdle JD: The Bagby and Kuslich method of
lumbar interbody fusion. Spine 1998;23:12671279.
50 Laurie SWS, Kaban LB, Mulliken JB, Murray JE: Donor-site morbidity after harvesting rib and
iliac bone. Plast Reconstr Surg 1984;73:933938.
51 Martin GJ, Boden SD, Morone MA, Moskovitz PA: Posterolateral intertransverse process spinal
fusion arthrodesis with rhBMP-2 in a non-human primate: Important lessons learned regarding
dose, carrier, and safety. J Spinal Disord 1999;12:179186.
52 McHorney CA, Ware JE, Lu R, Sherbourne CD: The MOS 36-item short form health survey
(SF-36). III. Tests of data quality, scaling assumptions, and reliability across diverse groups. Med Care
1994;32:4066.
53 Muschler GF, Hyodo A, Manning T, Kambic H, Easley K: Evaluation of human bone morpho
genetic protein 2 in a canine spinal fusion model. Clin Orthop 1994;308:229240.
54 Rawlinson N: Morbidity after anterior cervical decompression and fusion: The influence of the
donor site on recovery, and the results of a trial of surgibone compared to autologous bone. Acta
Neurochir 1994;131:106118.
55 Ray CD: Threaded titanium cages for lumbar interbody fusions. Spine 1997;22:667680.
56 Sacks S: Anterior interbody fusion of the lumbar spine: Indications and results in 200 cases. Clin
Orthop 1966;44:163170.
57 Sandhu HS, Kabo IM, Turner AS, Liu D, Kanim LEA, Toth IM: RhBMP-2 augmentation of tita
nium fusion cages for experimental anterior lumbar fusion. Proceedings of the 11th Annual
Meeting of the North American Spine Society. Rosemont, North American Spine Society, 1996,
p 47.
58 Sandhu HS, Kanim LE, Kabo JM, Toth JM, Zeegen EN, Liu D, Delamarter RB, Dawson EG:
Effective doses of recombinant bone morphogenetic protein-2 in experimental spinal fusion.
Spine 1996;21:21152122.
59 Sandhu HS, Kanim LE, Kabo JM, Toth JM, Zeegan EN, Liu D, Seeger LL, Dawson EG:
Evaluation of rhBMP-2 with an OPLA carrier in a canine posterolateral (transverse process)
spinal fusion model. Spine 1995;20:26692682.
60 Sandhu HS, Kanim LE, Toth JM, Kabo JM, Liu D, Delamarter RB, Dawson EG: Experimental
spinal fusion with recombinant human bone morphogenetic protein-2 without decortication of
osseous elements. Spine 1997;22:11711180.
61 Shimandle JH, Boden SO, Hutton WC: Experimental spinal fusion with recombinant human bone
morphogenetic protein-2 (rhBMP-2). Spine 1995;20:13261337.
62 Scene CL, Freeze A, Weil RJ, Marquette PJ: Analysis of harvest morbidity and radiographic out
come using autograft for anterior cervical fusion. Spine 1997;22:22222227.
63 Schuster JJ: CRC Handbook of Sample Size: Guidelines for Clinical Trial. Boca Raton, CRC
Press, 1990.
64 Sorenson KH: Anterior interbody lumbar spine fusion for incapacitating disc degeneration and
spondylolisthesis. Acta Orthop Scand 1978;49:269277.
65 Stauffer RN, Coventry MB: Anterior interbody lumbar spine fusion. J Bone Joint Surg Am
1972;54:756768.
66 Stewart AL, Hays RD, Ware JE: The MOS short form general health survey (SF-36). Med Care
1988;26:724732.
67 Urist MR: Bone. Formation by autoinduction. Science 1965;150:893899.
68 Urist MR, Strates BS: Bone morphogenetic protein. J Dent Res 1971;50:13921406.
69 Ware JE, Sherbourne CD: The MOS 36-item short form health survey (SF-36). I. Conceptual
framework and item selection. Med Care 1992;30:473483.
70 Wozney JM: Bone morphogenetic proteins and their gene expression; in Noda M (ed): Cellular
and Molecular Biology of Bone. Boston, Academic Press, 1993, pp 131167.
dramroo
71 Yamaguchi A, Katagiri T, Ikeda T, Wozney JM, Rosen V, Wang EA, Kahn AJ, Suda T, Yoshiki S:
Recombinant human bone morphogenetic protein stimulates osteoblastic maturation and inhibits
myogenic differentiation in vitro. J Cell Biol 1991;113:681687.
72 Younger EM, Chapman MW: Morbidity at bone graft donor sites. J Orthop Trauma 1989;3:
192195.
73 Zdeblick TA, Ghanayem AJ, Rapoff AJ, Swain C, Bassett T, Cooke ME, Markel M: Cervical inter-
body fusion cages: An animal model with and without bone morphogenetic protein. Spine
1998;23:758766.
Brian R. Subach, MD, Neurosurgery
550 Peachtree Street/ Suite 806, Atlanta, GA 30308-2247 (USA)
Tel. 1 404 686 8101, Fax 1 404 686 4805
BMP Update 13
dramroo
Cellular and Genetic Approaches for

Spinal Fusions
Gregory A. Helma, b, D. Greg Andersonc
Departments of aNeurosurgery, bBiomedical Engineering and cOrthopedic Surgery,
University of Virginia Health System, Charlottesville, Va., USA
A variety of evolving technologies are currently being evaluated in pre

clinical studies to promote tissue repair and/or regeneration in the spinal region.
Cellular and genetic techniques to induce bone formation for interbody or pos
terolateral spinal arthrodesis are an attractive approach, since this technology
could be employed through a minimally invasive approach with decreased
morbidity and potentially higher fusion rates compared to traditional open pro
cedures. Tissue engineering techniques are also being evaluated for disc repair
and regeneration using various growth and differentiation factors, mesenchy
mal stem cells (MSCs) and genetic therapies. Percutaneous soft tissue stabi
lization techniques are also within the scope of current technologies, potentially
allowing stabilization of the spine by the induction of ligamentous tissues, thus
avoiding extensive spinal arthrodesis procedures. Clearly, tissue-engineering
techniques will continue to evolve and certainly lead to more effective, and less
invasive, procedures for the treatment of traumatic, neoplastic and degenerative
spine problems.
Stem Cell Technologies
The direct application of osteoinductive stem cells on a bioresorbable

matrix, or scaffold, may be a useful strategy for engineering bone and soft tis
sues in the paraspinal region. Stem cells have several advantages for inducing
osteogenesis compared with the application of bone growth factors alone.
Pluripotent stem cells have the capacity to differentiate into the desired cell type
and form appropriate matrix elements under the control of various growth and
dramroo
differentiation factors. By supplying stem cells to the site, the growth factor
response is not dependent solely on the availability of local stem cell popula
tions, which may be diminished due to senility, radiation, medical factors, or
medications which might alter the intrinsic stem cell population. Therefore,
direct implantation of pluripotent cells with or without growth factors has the
potential to yield more rapid and uniform bony healing. Osteoprogenitor cells
have been isolated from the bone marrow of rats, rabbits, dogs, and humans, as
well as nonmarrow locations such as adipose tissue. These cells can be isolated
and expanded in tissue culture, prior to implantation into a target location. Stem
cells also have the potential to be genetically altered to express various growth
factors or other therapeutic genes prior to implantation.
A variety of studies have demonstrated the utility of using stem cells to
achieve bone formation. Bruder et al. [1] placed autologous canine MSCs
obtained from bone marrow onto porous cylinders composed of hydroxyapatite
and tricalcium phosphate. The implant was subsequently grafted into critical-
sized femoral defects. Implants without MSCs produced atrophic nonunions in
all treated animals. Implants containing the MSCs produced lamellar and
woven bone within the carrier and resulted in a solid union at the site of the
defect. Bruder et al. [2] also successfully achieved bone induction using human
MSCs on a ceramic carrier in athymic nude rats. Using radiography, biome
chanical testing, and histologic analysis, the human MSCs were capable of
healing critical-sized femoral defects. Quarto et al. [3] studied the use of autol
ogous bone marrow stromal cells delivered on a macroporous hydroxyapatite
carrier for the healing of large bony defects (4.0 cm) in a small human clinical
series. The implant was placed within the defect, and the fracture was stabilized
with external fixation. In each instance, the composite implant was able to
achieve successful union of the bony defect, thus producing the first direct
evidence that tissue engineering of bone can be successfully applied to humans
in a clinical setting.
Bone Morphogenetic Protein Gene Therapy
Bone Morphogenetic Protein Overview

Bone morphogenetic proteins (BMPs) are a group of secreted proteins that
are members of the transforming growth factor- (TGF-) superfamily based
on their high degree of homology within the C-terminal seven-cysteine region
[47]. These proteins were originally recognized for their ability to form ectopic
bone by inducing primitive mesenchymal cell chemotaxis, proliferation, and
differentiation into chondrocytes and osteoblasts [813]. The ectopic bone is
formed primarily through endochondral mechanisms, recapitulating many of
Cellular and Genetic Approaches 15

dramroo
the events seen during embryonic development [7, 10, 11, 1315]. However,
when delivered in high concentrations, direct or intramembranous bone forma
tion can also be induced at the treatment site. Other BMPs have been found
experimentally to induce the formation of various other tissue types, including
cartilage, tendon, and ligament [1624]. Molecular cloning techniques have
allowed the production of certain BMPs in large quantities, which has con
tributed to the rapid advancement of our collective understanding of BMP
biology and potential for clinical applications [7].
BMPs are initially synthesized as monomeric precursors that, after prote
olytic cleavage, undergo dimerization through disulfide bonding [15, 25]. These
dimers, which can be either homologous or heterologous, are released extracellu
larly and act by binding to type I and II receptors [2634]. Receptor binding and
subsequent transphosphorylation leads to the phosphorylation and activation of
various cytoplasmic Smad proteins [3540]. These activated Smad proteins then
translocate to the nucleus and initiate cellular responses, including chemotaxis,
proliferation, and differentiation.
Gene Therapy Overview

Gene-based therapies attempt to deliver specific genes, known as the trans-
gene, to target cells to change the existing physiologic state or disease process
[41, 42]. Delivery of the genetic material to the target cell can be accomplished
by both nonviral and viral vectors. Nonviral or synthetic delivery techniques
include the use of molecular conjugates, liposomes, naked DNA, plasmids,
electroporation, or the incorporation of the transgenes with viral proteins.
To date, the majority of these techniques are limited by their relatively low cel
lular transduction rates. In contrast, viral vectors can be highly efficient in their
delivery of genetic material. Some of the more common viruses used for gene
therapy include the adenovirus, herpes virus, retrovirus, and adeno-associated
viruses (AAV). These vectors are often genetically modified so that they are
replication-deficient, through the removal of specific portions of the viral
genome, which allow for the insertion of a therapeutic transgene. Although
gene therapy as a concept continues to capture the imagination of basic scien
tists and clinical researchers, the advancement of the techniques toward human
use has proven difficult [4144]. Technical hurdles such as insufficient trans-
gene expression, an inability to circumvent the host immune system, and a fail
ure to achieve long-term transgene expression continue to present a challenge
for many clinical applications. Fortunately, bone induction using BMPs may be
an ideal application of gene therapy because long-term expression of the BMP
is not required to achieve therapeutic bone formation. In fact, the elimination
of BMP expression may be advantageous in that tissue overgrowth or toxicity
is less likely to be encountered.
Helm/Anderson 16
dramroo
Viral Vectors
Adenoviral Vectors. Adenoviruses are double-stranded DNA viruses
which bind to specific cell surface receptors, enter the cells by endocytosis, and
subsequently release their DNA into the cytoplasm [45, 46]. The viral genome
is divided into immediate early genes, early genes, and late genes according to
the time in which the genes are expressed. The immediate early genes activate
early gene transcription, while the early genes are involved both in subsequent
viral replication and host immune evasion. The late genes code for the aden
oviruses structural proteins [46]. Most adenoviral vectors studied to date are
derived from the adenovirus serotype-5, which are rendered replication-
defective by deletion of the E1 region. The E3 region is often deleted as well to
make room for larger transgenes. First generation adenoviral vectors can
accommodate up to 8 kb of foreign DNA.
The advantages of adenoviral vectors include their ability to be produced in
high titers, their extrachromosomal life cycle, which reduces the risk of insertional
mutagenesis, and their ability to transfect numerous cell types [46, 47]. There are
several potential disadvantages of the adenoviral vector for gene therapy. Because
the virus does not integrate into the cellular genome, the length of gene expression
is limited and therapeutic genes are not passed to the progeny of the transduced
cells. Perhaps the most problematic issue with adenoviral vectors, however, is the
robust humoral and cellular immune response that can occur at the treatment site
resulting in reduced transgene expression. Intense research efforts are currently
directed at blunting the immune response by alterations to the adenoviral vector
such as deleting the viral DNA polymerase gene (pol adenoviral vectors) or
completely eliminating the viral genome (gutless adenoviral vectors).
AAV Vectors. AAV, which are defective single-stranded DNA parvoviruses,
are also attractive vectors for BMP gene therapy studies [48]. AAV vectors have
the ability to integrate stably into the target cells genome, transduce a variety of
cell types, maintain high levels of gene expression, transfect both proliferating
and quiescent cells, and be generated in high titers. Numerous studies have
demonstrated that AAV can efficiently transduce muscle and other cells in vivo
with little inflammatory response and no evidence of insertional mutagenesis.
The production of AAV vectors was initially fraught with numerous technical
difficulties; however, current techniques of vector production lead to a high yield
of recombinant AAV, completely free of wild-type AAV.
Retroviral Vectors. Retroviral vectors enter target cells through interac
tions between the viral envelope proteins and cell surface glycoproteins.
Importantly, retroviruses contain single-stranded RNA. Once the viral genome
is released into the cytoplasm, retroviral reverse transcriptase produces a double-
stranded DNA copy, which subsequently integrates into the host genome
during mitosis. One disadvantage to using retroviral vectors is that retroviruses

dramroo
can only integrate their genetic material into proliferating cells [49]. Therefore,
it may be difficult to achieve adequate cellular transduction at sites requiring
bone induction by direct injection. Other major disadvantages of retroviral
vectors are their low infectivity and instability of the virions. The utilization of
retroviruses for BMP gene therapy will most likely be limited to ex vivo
approaches, such as for the genetic modification of MSC populations.
Herpes Viral Vectors. Herpes viruses are double-stranded DNA viruses
which can cause significant human pathology, including cold sores and encephali
tis. Gene therapy studies utilizing herpes viral vectors typically utilize geneti
cally modified herpes simplex type 1. In the normal life cycle of these viruses,
the virion fuses with the cell membrane and is transported to the nucleus where,
after several phases of gene transcription, the cell lyses and releases progeny
viral particles. Herpes viruses also have the ability to enter a latency phase, dur
ing which time the viral genome is not actively transcribed. This latency phase
can be lost during cell division [46]. Herpes viral vectors have the advantage of
being able to accommodate up to 40 kb of foreign DNA and can be utilized to
insert foreign DNA into a variety of cell populations, including myocytes, with
limited toxicity. Herpes viral vectors may, therefore, be a reasonable vector for
BMP gene therapy [50].
Nonviral Vectors
Direct Plasmid Injection. Wolff [51] was the first to show that the direct
intramuscular injection of plasmid DNA could lead to low level, short-term
gene expression. Other researchers have now demonstrated the successful
transduction of the myocardium, brain, thyroid, and various tumors using this
technique [5255]. Inadequate cellular transduction rates can be significantly
increased by pretreatment of the injected area with hypertonic saline or bupi
vacaine [56]. Systemic gene expression can also be obtained by the direct intra
venous injection of naked DNA in adults, although rapid degradation of the
DNA prior to reaching the target cell remains problematic in the absence of a
delivery system, such as cationic liposomes [57]. Currently, no published stud
ies have demonstrated successful bone formation using direct injections of
osteogenic plasmid into either orthotopic or heterotopic sites.
Electroporation. The diffusion of extracellular DNA into a cell in vitro
and in vivo can be significantly increased by permeabilizing the cells mem
brane using short, high intense electric pulses. The technique essentially opens
small pores in the cells membrane, through which molecules can diffuse down
concentration gradients. When the pores spontaneously close, the DNA is sealed
within the cells cytoplasm, where it can be transported to the nucleus [58]. This
technique can increase the transduction rate over a 1,000-fold compared to
direct plasmid injection, and has been utilized to transduce liver, melanoma,
Helm/Anderson 18
dramroo
skin, and muscle cells [59, 60]. The ideal parameters for cellular transduction
appear to vary between tissues. In addition, only short-term gene expression has
been achieved. The utilization of this approach for the delivery of osteogenic
genes has yet to be defined.
Gene Gun. Another interesting technology which is currently under investi
gation for transducing cells with foreign DNA is the gene gun. The technique
involves the coating of gold particles with plasmid DNA, which are subsequently
bombarded into the tissue of interest [61]. Under optimal conditions, the gene gun
can be utilized to transduce 1020% of the cells at the treatment site. Although
gene expression of up to 60 days has been achieved, the depth of tissue penetra
tion is limited to 0.5 mm [62]. In addition, low levels of gene expression are
generally achieved, which is problematic for BMP gene therapy, where relatively
high levels of local BMP expression are required for tissue induction.
Liposomes. Liposomes are commonly used to deliver DNA to cells in
vitro [63]. Cationic derivatives of diacylglycerol and cholesterol, lipid deriva
tives of polyamines, and quaternary ammonium detergents are typically used
to form the cationic lipid-DNA complexes. The cationic-lipid compounds serve
to decrease the negative changes of the DNA plasmids and facilitate entrance
of the plasmids through the cell membrane. Liposome preparations also
contain neutral or helper lipids, including cholesterol or dioleoylphos
phatidylethanolamine, to improve DNA release from the endosome into the
cytoplasm [64]. Liposomes can transduce a wide variety of cells and tissues,
including vascular endothelium, lung, brain, and skin [65, 66]. The attachment
of cell-specific antibodies to the liposomal membrane may improve tissue
specificity of this transduction technique. Intense research efforts are currently
directed at improving liposome production and delivery, which may render
these vectors ideal for BMP gene delivery in the near future.
Polymer-DNA Complexes. High-molecular-weight cationic polymers, such
as poly-L-lysine, poly-L-ornithine, polyethylenimine, and chitosan, can improve
DNA delivery to cells via nonspecific absorptive uptake [67, 68]. Various syn
thetic polymers can also be utilized to improve cellular transduction rates and
can be designed to be biodegradable, thermosensitive, and biocompatible [69].
Polymers can also be constructed with targeting ligands, such as antibodies,
transferrin, and asialoglycoprotein to improve tissue specificity. Additional
modifications to these molecules can also be made to improve cellular uptake
and cytoplasmic trafficking of the therapeutic gene.
Direct BMP Gene Therapy

The promotion of osteogenesis through the direct injection of adenoviral
vectors in vivo has been successfully achieved in both immunosuppressed and
immune-competent animals [70, 71]. All immune-competent animals tested

dramroo
with BMP adenoviral vectors, to date, have demonstrated evidence of an
immune response at the treatment site. The immune response appears to be both
humoral and cellular, and is directed against both the injected adenoviral parti
cles and cells transduced by the adenoviral vector. It remains unclear whether
there is a major immune response directed against the foreign BMP (i.e. human
BMP gene in animal models) expressed using gene therapy techniques. A vari
ety of different BMP adenoviral vectors have been shown to induce bone for
mation in athymic nude rodents, including Ad-BMP-2, Ad-BMP-4, Ad-BMP-6,
Ad-BMP-7 and Ad-BMP-9. In immune-competent rodents, Ad-BMP-6 and
Ad-BMP-9 are able to overcome the host immune response and induce signif
icant bone formation. Figure 1 demonstrates the mechanisms involved in
osteogenesis using direct BMP gene therapy with an adenoviral vector.
In the paraspinal region, direct injection of adenoviral vectors, including
Ad-BMP-2 and Ad-BMP-9, have been shown to induce spinal fusion in athymic
nude rodents, without producing central or lateral stenosis caused by exuberant
bone formation [71, 72]. The fusion mass completely integrates with the adja
cent laminae and spinal processes, without evidence of pseudoarthrosis, sug
gesting that decortication of cortical bone is not required for bony fusion when
BMPs are applied in the paraspinal region. Stereotactically injected BMP vec
tors into the paraspinal musculature is a compelling approach, since multiple,
percutaneous injections could be performed to produce a fusion mass with a
predetermined three-dimensional shape and in locations specific for each patho
logical process. Postprocedural pain would be minimal since the technique
would not require muscular dissection. Patients requiring neural decompression
could be taken to surgery following successful bony fusion, where extensive
laminectomies, facetectomies, and foraminotomies could be performed without
leading to spinal instability. Figure 2 demonstrates a posterolateral spinal fusion
in the lumbar region of a rabbit using percutaneous, direct gene therapy of a
BMP-6 adenoviral vector, demonstrating significant bone formation with
excellent union with the host transverse processes.
Ex vivo BMP Gene Therapy

Another approach that is currently being investigated by several research
groups is ex vivo gene therapy, a technique in which osteogenic genes are
inserted into cells in tissue culture, and the genetically altered cells are subse
quently implanted into regions requiring bone formation in experimental ani
mals [73]. The cellular implants express and secrete bone morphogens, which
in turn induce the osteogenic response. Lieberman et al. [74, 75] demonstrated
that a murine bone marrow stromal cell line, which had been transduced with
BMP-2 cDNA using an adenoviral vector, could induce both heterotopic and
orthotopic bone formation in severe combined immune-deficient (SCID) mice.
Helm/Anderson 20
dramroo
Fig. 1. Diagram showing mechanisms
of bone formation using direct, percuta
neous injection of an adenoviral vector con
taining the BMP-9 gene. Expression of
BMP-9 by the transduced cells leads to
migration, proliferation, and differentiation
of MSCs, ultimately leading to successful
osteogenesis.
Fig. 2. Three-dimensional CT recon

struction of rabbit lumbar spine treated with
bilateral, percutaneous injections of a BMP-6
adenoviral vector using a posterolateral
approach, demonstrating significant bone
induction and a successful transverse process
spinal fusion. The data demonstrates that
successful osteogenesis can be achieved in
immune-competent rabbits using current
gene therapy technologies.

dramroo
In a similar study, Riew et al. [76] transduced marrow-derived MSCs with the
BMP-2 gene and autologously reimplanted the genetically modified cells in the
paraspinal region in rabbits. One of the five treated rabbits demonstrated as suc
cessful intertransverse process spinal fusion, which was assessed radiographi
cally and histologically. Although only 20% of the treated animals showed
significant bone formation, this study clearly demonstrated the potential of
ex vivo gene therapy techniques. This group subsequently demonstrated that the
rate of bone formation could be increased to 100% in the treated region by
implanting the cells 7 days following viral transfection [77].
Musgrave [78] transduced mesenchymal cell populations obtained from
human muscle with the BMP-2 gene using an adenoviral vector. The trans
duced cells were implanted into SCID mice and demonstrated successful
ectopic bone formation. Lee et al. [79] inserted the BMP-2 gene into muscle-
derived mesenchymal cells in mice and implanted the cells into mouse critical-
sized cranial defects. The transduced cells significantly increased the healing
rate of defects compared to control cells. In addition, fluorescent in situ
hybridization was utilized to demonstrate incorporation of the transduced cells
into the induced bone. In another compelling study, Turgeman [80] isolated
human MSCs from the bone marrow of normal patients, as well as patients
suffering from osteoporosis. The cells were transduced with the BMP-2 gene
and subsequently grafted into ectopic and orthotopic locations, leading to suc
cessful osteogenesis. Utilizing a retroviral BMP-2 vector, Laurencin et al. [81]
demonstrated successful heterotopic bone induction by BMP-2-transduced
cells delivered on a PLAGA-HA scaffold in a SCID mouse model. The expres
sion of the human bone morphogenetic protein-7 gene by periosteal-derived
rabbit mesenchymal cells has also been shown to induce bone and articular
cartilage repair in a rabbit knee osteochondral defect model [82]. In an elegant
study, Moutsatsos et al. [83] genetically engineered a murine MSC line to
secrete BMP-2 in a regulated fashion. The BMP-2 gene was under control of a
doxycycline-responsive promoter, such that the presence of doxycycline in
vitro and in vivo would downregulate BMP-2 expression. In vivo implantation
of this cell line led to both orthotopic and ectopic bone formation in a regulated
fashion, suggesting that long-term regulation of bone induction may be
possible.
Our lab has recently demonstrated that human MSCs transduced with the
BMP-9 gene can induce robust bone formation in athymic nude animals. In an
ectopic model, the implanted cells survive long-term, as assessed by an anti
human mitochondrial stain, and contribute to the cellular composition of the
ectopic bone. Local host stem cells at the injection site are also stimulated by
the secreted BMP, and differentiate into chondrocytes and osteoblasts at the
treatment site. The genetically modified cells were also capable of forming
Helm/Anderson 22
dramroo
significant bone formation in the paraspinal region following percutaneous
injection. The fusion mass integrated completely with the adjacent host spine,
similar to the direct BMP adenoviral vector treatment sites, without evidence of
posttreatment neural compression (unpubl. data).
In another set of interesting studies, Boden et al. [84, 85] have reported a
novel ex vivo gene therapy technique which utilizes the insertion of the osteogenic
LMP-1 gene into allogenic bone marrow cells. These investigators demonstrated
significant bone formation by the transduced cells in the paraspinal region of
rodents, in spite of relatively low transduction rates. LMP-1 gene therapy is
unique in that it is thought to induce the secretion of a variety of osteogenic
growth factors, which in turn stimulate bone formation.
These ex vivo techniques have the advantage of not only expressing
osteogenic morphogens, but also supplying the treated region with bone pre
cursor cells, which may be of limited supply at the treatment site. For example,
it is unclear whether pluripotent stem cells are uniformly present throughout the
body in the adult human, which could make direct BMP or BMP gene therapy
treatments ineffective. Also, the number of MSCs may decrease with age,
which might decrease the physiologic activity of BMPs. The introduction of
stem cells, which are genetically modified to secrete bone morphogens, is,
therefore, a compelling approach. The harvest and expansion of autologous
stem cells for widespread human use may be hampered by high costs, cellular
contamination, and other technical difficulties. Therefore, other approaches
such as the genetic modification of traditional bone grafts (which contain
cellular precursors such as stem cells and osteoblasts) at the time of surgery
may be a more reasonable near-term approach.
Future Research and Development
Recent studies have clearly demonstrated the potential utility of MSCs and
BMP gene therapy for the promotion of bone formation for spinal applications.
However, the field of molecular spine surgery is certainly in its infancy. Many
of the techniques are under continuing development and may require refine
ment prior to clinical application. The establishment of an allogenic or xeno
geneic source of stem cells, which could be modified to attenuate potential host
immune responses, would have significant advantages compared to autologous
cells for clinical application. Various groups are currently studying the use of
genetically modified porcine cells, which are genetically altered to decrease
their expression of foreign antigens. In addition, ex vivo gene therapy approaches
are currently being developed to induce local immunosuppression around the
transplanted cells. The efficacy of BMP gene therapy might also be improved

dramroo
BMP BMP
vector I vector II
Muscle
BMP I cell
homodimers BMP I/II
heterodimers BMP II

homodimers

Fig. 3. Diagram demonstrating the

Host local production of potent BMP het-
stem
cell Bone erodimers using combinational gene therapy
formation techniques with two BMP adenoviral vectors.
Osteocyte
Two homodimeric and one heterodimeric
Chondrocyte species are produced, which may be more
Osteoblast effective at promoting bone formation than
treatment with a single BMP vector.
by achieving the expression of a cocktail of different growth factors at the treat

ment site. Several growth factors, including BMP-6 and TGF-3, have been
shown to have synergistic effects on stem cells in vitro. In addition, the gener
ation of potent BMP heterodimers could be produced using a combination of
vectors, which may have improved osteogenic effects compared to their respec
tive BMP homodimers alone (fig. 3). Finally, the optimal gene delivery system
for osteogenic genes has certainly not been firmly established for human use.
Although viral vectors have clearly been the most effective approach in animal
studies, their use in humans may be limited by their antigenicity. Therefore, the
development of novel nonviral approaches which are able to achieve short-term,
high-level transgene expression would be ideal for human use. It is anticipated
that many of these issues will be addressed in the near future and that cutting
edge tissue engineering technologies can be successfully applied to treat human
spine pathology.
Acknowledgment
This work was supported by an NIH grant (1 RO1 AR/AI4648801A2) and Medtronic-
Sofamor Danek; we thank Mona Banton for her editorial support.
Helm/Anderson 24
dramroo
References
1 Bruder SP, Kurth AA, Shea M, Hayes WC, Jaiswal N, Kadiyala S: Bone regeneration by implantation
of purified, culture-expanded human mesenchymal stem cells. J Orthop Res 1998;16:155162.
2 Bruder SP, Kraus KH, Goldberg VM, Kadiyala S: The effect of implants loaded with autologous
mesenchymal stem cells on the healing of canine segmental bone defects. J Bone Joint Surg Am
1998;80:985996.
3 Quarto R, Mastrogiacomo M, Cancedda R, Kutepov SM, Mukhachev V, Lavroukov A, Kon E,
Marcacci M: Repair of large bone defects with the use of autologous bone marrow stromal cells.
N Engl J Med 2001;344:385386.
4 Celeste AJ, Iannazzi JA, Taylor RC, Hewick RM, Rosen V, Wang EA, Wozney JM: Identification
of transforming growth factor beta family members present in bone-inductive protein purified
from bovine bone. Proc Natl Acad Sci USA 1990;87:98439847.
5 Sampath TK, Reddi AH: Homology of bone-inductive proteins from human, monkey, bovine, and
rat extracellular matrix. Proc Natl Acad Sci USA 1983;80:65916595.
6 Sampath TK, Nathanson MA, Reddi AH: In vitro transformation of mesenchymal cells derived
from embryonic muscle into cartilage in response to extracellular matrix components of bone.
7 Wozney JM, Rosen V, Celeste AJ, Mitsock LM, Whitters MJ, Kriz RW, Hewick RM, Wang EA:
Novel regulators of bone formation: Molecular clones and activities. Science 1988;242:15281534.
8 Urist MR: Bone: Formation by autoinduction. Science 1965;150:893899.
9 Urist MR, Silverman BF, Buring K, Dubuc FL, Rosenberg JM: The bone induction principle. Clin
Orthop 1967;53:243283.
10 Urist MR, Dowell TA, Hay PH, Strates BS: Inductive substrates for bone formation. Clin Orthop
1968;59:5996.
11 Urist MR: A morphogenetic matrix for differentiation of bone tissue. Calcif Tissue Res 1970
(suppl):98101.
12 Urist MR, Strates BS: Bone morphogenetic protein. J Dent Res 1971;50:13921406.
13 Urist MR, Sato K, Brownell AG, Malinin TI, Lietze A, Huo YK, Prolo DJ, Oklund S, Finerman GA,
DeLange RJ: Human bone morphogenetic protein (hBMP). Proc Soc Exp Biol Med 1983;
173:194199.
14 Yamaguchi A, Katagiri T, Ikeda T, Wozney JM, Rosen V, Wang EA, Kahn AJ, Suda T, Yoshiki S:
Recombinant human bone morphogenetic protein-2 stimulates osteoblastic maturation and
inhibits myogenic differentiation in vitro. J Cell Biol 1991;113:681687.
15 Reddi AH, Wientroub S, Muthukumaran N: Biologic principles of bone induction. Orthop Clin
North Am 1987;18:207212.
16 Helm GA, Li JZ, Alden TD, Hudson SB, Beres EJ, Cunningham M, Mikkelsen MM, Pittman DD,
Kerns KM, Kallmes DF: A light and electron microscopic study of ectopic tendon and ligament
formation induced by bone morphogenetic protein-13 adenoviral gene therapy. J Neurosurg 2001;
95:298301.
17 Aspenberg P, Forslund C: Enhanced tendon healing with GDF 5 and 6. Acta Orthop Scand 1999;
70:5154.
18 Chang SC, Hoang B, Thomas JT, Vukicevic S, Luyten FP, Ryba NJ, Kozak CA, Reddi AH,
Moos M Jr: Cartilage-derived morphogenetic proteins. New members of the transforming growth
factor-beta superfamily predominantly expressed in long bones during human embryonic develop
ment. J Biol Chem 1994;269:2822728234.
19 Chen P, Carrington JL, Hammonds RG, Reddi AH: Stimulation of chondrogenesis in limb bud
mesoderm cells by recombinant human bone morphogenetic protein 2B (BMP-2B) and modula
tion by transforming growth factor beta 1 and beta 2. Exp Cell Res 1991;195:509515.
20 Evans CHPD, Robbins PDP: Genetically augmented tissue engineering of the musculoskeletal
system. Clin Orthop 1999;367S:S410S418.
21 Grande DAP, Breitbart ASMD, Mason JP, Paulino CMD, Laser JBS, Schwartz REMD: Cartilage
tissue engineering: Current limitations and solutions. Clin Orthop 1999;367S:S176S185.
22 Kapur SP, Reddi AH: Chondrogenic potential of mesenchymal cells elicited by bone matrix in
vitro. Differentiation 1986;32:252259.

dramroo
23 Reddi AH: Extracellular bone matrix dependent local induction of cartilage and bone. J Rheumatol
Suppl 1983;11:6769.
24 Sampath TK, Reddi AH: Distribution of bone inductive proteins in mineralized and demineralized
extracellular matrix. Biochem Biophys Res Commun 1984;119:949954.
25 Reddi AH: Bone morphogenetic proteins: An unconventional approach to isolation of first mam
malian morphogens. Cytokine Growth Factor Rev 1997;8:1120.
26 Feng XH, Derynck R: Ligand-independent activation of transforming growth factor (TGF) beta
signaling pathways by heteromeric cytoplasmic domains of TGF-beta receptors. J Biol Chem
1996;271:1312313129.
27 Feng XH, Filvaroff EH, Derynck R: Transforming growth factor-beta (TGF-beta)-induced
down-regulation of cyclin A expression requires a functional TGF-beta receptor complex.
Characterization of chimeric and truncated type I and type II receptors. J Biol Chem 1995;270:
2423724245.
28 Chen RH, Moses HL, Maruoka EM, Derynck R, Kawabata M: Phosphorylation-dependent inter
action of the cytoplasmic domains of the type I and type II transforming growth factor-beta recep
tors. J Biol Chem 1995;270:1223512241.
29 Chen RH, Derynck R: Homomeric interactions between type II transforming growth factor-beta
receptors. J Biol Chem 1994;269:2286822874.
30 Derynck R: TGF-beta-receptor-mediated signaling. Trends Biochem Sci 1994;19:548553.
31 Lawler S, Candia AF, Ebner R, Shum L, Lopez AR, Moses HL, Wright CV, Derynck R: The
murine type II TGF-beta receptor has a coincident embryonic expression and binding preference
for TGF-beta 1. Development 1994;120:165175.
32 Ebner R, Chen RH, Lawler S, Zioncheck T, Derynck R: Determination of type I receptor speci
ficity by the type II receptors for TGF-beta or activin. Science 1993;262:900902.
33 Ebner R, Chen RH, Shum L, Lawler S, Zioncheck TF, Lee A, Lopez AR, Derynck R: Cloning of
a type I TGF-beta receptor and its effect on TGF-beta binding to the type II receptor. Science
1993;260:13441348.
34 Chen RH, Ebner R, Derynck R: Inactivation of the type II receptor reveals two receptor pathways
for the diverse TGF-beta activities. Science 1993;260:13351338.
35 Zhang Y, Derynck R: Regulation of Smad signalling by protein associations and signalling
crosstalk. Trends Cell Biol 1999;9:274279.
36 Feng XH, Lin X, Derynck R: Smad2, Smad3 and Smad4 cooperate with Sp1 to induce p15(Ink4B)
transcription in response to TGF-beta. EMBO J 2000;19:51785193.
37 Derynck R, Zhang Y, Feng XH: Smads: Transcriptional activators of TGF-beta responses. Cell
1998;95:737740.
38 Derynck R: SMAD proteins and mammalian anatomy. Nature 1998;393:737739.
39 Lawler S, Feng XH, Chen RH, Maruoka EM, Turck CW, Griswold-Prenner I, Derynck R: The type
II transforming growth factor-beta receptor autophosphorylates not only on serine and threonine
but also on tyrosine residues. J Biol Chem 1997;272:1485014859.
40 Filvaroff E, Erlebacher A, Ye J, Gitelman SE, Lotz J, Heillman M, Derynck R: Inhibition of TGF-
beta receptor signaling in osteoblasts leads to decreased bone remodeling and increased trabecu
lar bone mass. Development 1999;126:42674279.
41 Miller AD: Human gene therapy comes of age. Nature 1992;357:455460.
42 Trapnell BC, Gorziglia M: Gene therapy using adenoviral vectors. Curr Opin Biotechnol 1994;5:
617625.
43 Marcel T, Grausz JD: The TMC Worldwide Gene Therapy Enrollment Report, end 1996. Hum
Gene Ther 1997;8:775800.
44 Anderson WF: Gene therapy. Sci Am 1995;273:124128.
45 Seth P, Willingham MC, Pastan I: Adenovirus-dependent release of 51Cr from KB cells at an acidic
pH. J Biol Chem 1984;259:1435014353.
46 Kramm C, Sena-Esteves M, Barnett FH, et al: Gene therapy for brain tumors. Brain Pathol 1995;5:
345381.
47 Le Gal La Salle G, Robert JJ, Berrard S, Ridoux V, Stratford-Perricaudet LD, Perricaudet M,
Mallet J: An adenovirus vector for gene transfer into neurons and glia in the brain. Science 1993;
259:988990.
Helm/Anderson 26
dramroo
48 Fisher KJ, Jooss K, Alston J, Yang Y, Haecker SE, High K, Pathak R, Raper SE, Wilson JM:
Recombinant adeno-associated virus for muscle directed gene therapy. Nat Med 1997;3:
306312.
49 Culver KW: Gene therapy for malignant neoplasms of the CNS. Bone Marrow Transplant 1996;
18(suppl 3):S6S9.
50 Coffin RS, Howard MK, Cumming DV, Dollery CM, McEwan J, Yellon DM, Marber MS,
MacLean AR, Brown SM, Latchman DS: Gene delivery to the heart in vivo and to cardiac
myocytes and vascular smooth muscle cells in vitro using herpes virus vectors. Gene Ther 1996;3:
560566.
51 Wolff JA: Direct gene transfer into mouse muscle in vivo. Science 1990;247:14651468.
52 Nomura T: Gene expression and antitumor effects following direct interferon (IFN)-gamma gene
transfer with naked plasmid DNA and DC-chol liposome complexes in mice. Gene Ther 1999;6:
121129.
53 Ardehali A: Direct gene transfer into donor hearts at the time of harvest. J Thorac Cardiovasc Surg
1995;109:716720.
54 Schwartz B: Gene transfer by naked DNA into adult mouse brain. Protein Sci 1996;5:488494.
55 Sikes ML: In vivo gene transfer into rabbit thyroid follicular cells by direct DNA injection. Hum
Gene Ther 1994;5:837844.
56 Davis HL: Direct gene transfer into skeletal muscle in vivo: Factors affecting efficiency of trans
fer and stability of expression. Hum Gene Ther 1993;4:151159.
57 Zhang G: High levels of foreign gene expression in hepatocytes after tail vein injections of naked
plasmid DNA. Hum Gene Ther 1999;10:17351737.
58 Mir LM: Introduction of definite amounts of nonpermeant molecules into living cells after elec
tropermeabilization: Direct access to the cytosol. Exp Cell Res 1988;175:1525.
59 Titomirov AV: In vivo electroporation and stable transformation of skin cells of newborn mice by
plasmid DNA. Biochim Biophys Acta 1991;1088:131134.
60 Aihara H, Miyazaki J: Gene transfer into muscle by electroporation in vivo. Nat Biotechnol 1998;
16:867870.
61 Yang NS: In vivo and in vitro gene transfer to mammalian somatic cells by particle bombardment.
62 Zelenin AV: Bacterial beta-galactosidase and human dystrophin genes are expressed in mouse
skeletal muscle fibers after ballistic transfection. FEBS Lett 1997;414:319322.
63 Felgner PL: Lipofection: A highly efficient, lipid-mediated DNA-transfection procedure. Proc
Natl Acad Sci USA 1987;84:74137417.
64 Farhood H: The role of dioleoyl phosphatidylethanolamine in cationic liposome mediated gene
transfer. Biochim Biophys Acta 1995;1235:289295.
65 Ono T: Plasmid DNAs directly injected into mouse brain with lipofectin can be incorporated and
expressed by brain cells. Neurosci Lett 1990;117:259263.
66 Raz E: Intradermal gene immunization: The possible role of DNA uptake in the induction of
cellular immunity to viruses. Proc Natl Acad Sci USA 1994;91:95199523.
67 Nishikawa M: Nonviral vectors in the new millennium: Delivery barriers in gene transfer. Hum
Gene Ther 2001;12:861870.
68 Boussif O: A versatile vector for gene and oligonucleotide transfer into cells in culture and in vivo:
Polyethylenimine. Proc Natl Acad Sci USA 1995;92:72977301.
69 Maheshwari A: Soluble biodegradable polymer-based cytokine gene delivery for cancer treatment.
Mol Ther 2000;2:121130.
70 Varady P, Li JZ, Cunningham M, Beres EJ, Das S, Engh J, Alden TD, Pittman DD, Kerns KM,
Kallmes DF, Helm GA: Morphologic analysis of BMP-9 gene therapy-induced osteogenesis. Hum
Gene Ther 2001;12:697710.
71 Alden TD, Pittman DD, Beres EJ, Hankins GR, Kallmes DF, Wisotsky BM, Kerns KM, Helm GA:
Percutaneous spinal fusion using bone morphogenetic protein-2 gene therapy. J Neurosurg 1999;
90:109114.
72 Helm GA, Alden TD, Beres EJ, Hudson SB, Das S, Engh JA, Pittman DD, Kerns KM, Kallmes DF:
Use of bone morphogenetic protein-9 gene therapy to induce spinal arthrodesis in the rodent.
J Neurosurg 2000;92:191196.

dramroo
73 Gazit D, Turgeman G, Kelley P, Wang E, Jalenak M, Zilberman Y, Moutsatsos I: Engineered
pluripotent mesenchymal cells integrate and differentiate in regenerating bone: A novel cell-
mediated gene therapy. J Gene Med 1999;1:121133.
74 Lieberman JR, Le LQ, Wu L, Finerman GA, Berk A, Witte ON, Stevenson S: Regional gene ther
apy with a BMP-2-producing murine stromal cell line induces heterotopic and orthotopic bone
formation in rodents. J Orthop Res 1998;16:330339.
75 Lieberman JR, Daluiski A, Stevenson S, Wu L, McAllister P, Lee YP, Kabo JM, Finerman GA,
Berk AJ, Witte ON: The effect of regional gene therapy with bone morphogenetic protein-2
producing bone-marrow cells on the repair of segmental femoral defects in rats. J Bone Joint Surg
Am 1999;81:905917.
76 Riew KD, Wright NM, Cheng S, Avioli LV, Lou J: Induction of bone formation using a recombi
nant adenoviral vector carrying the human BMP-2 gene in a rabbit spinal fusion model. Calcif
Tissue Int 1998;63:357360.
77 Cheng SL, Lou J, Wright NM, Lai CF, Avioli LV, Riew KD: In vitro and in vivo induction of bone
formation using a recombinant adenoviral vector carrying the human BMP-2 gene. Calcif Tissue
Int 2001;68:8794.
78 Musgrave DS: Human skeletal muscle cells in ex vivo gene therapy to deliver bone morphogenetic
protein-2. J Bone Joint Surg Br 2002;84:120127.
79 Lee JYMD, Musgrave DMD, Pelinkovic DMD, Fukushima KMDP, Cummins JB, Usas AMD,
Robbins PP, Fu FHMD, Huard JP: Effect of bone morphogenetic protein-2-expressing muscle-
derived cells on healing of critical-sized bone defects in mice. J Bone Joint Surg Am 2001;83-A:
10321039.
80 Turgeman G: Engineered human mesenchymal stem cells: A novel platform for skeletal cell medi
ated gene therapy. J Gene Med 2001;3:240251.
81 Laurencin CT: Poly(lactide-co-glycolide)/hydroxyapatite delivery of BMP-2-producing cells:
A regional gene therapy approach to bone regeneration. Biomaterials 2001;22:12711277.
82 Mason JM, Breitbart AS, Barcia M, Porti D, Pergolizzi RG, Grande DA: Cartilage and bone regen
eration using gene-enhanced tissue engineering. Clin Orthop 2000;379(suppl):S171S178.
83 Moutsatsos IK, Turgeman G, Zhou S, Kurkalli BG, Pelled G, Tzur L, Kelley P, Stumm N, Mi S,
Muller R, Zilberman Y, Gazit D: Exogenously regulated stem cell-mediated gene therapy for bone
regeneration. Mol Ther 2001;3:449461.
84 Boden SD, Titus L, Hair G, Liu Y, Viggeswarapu M, Nanes MS, Baranowski C: Lumbar spine
fusion by local gene therapy with a cDNA encoding a novel osteoinductive protein (LMP-1).
Spine 1998;23:24862492.
85 Boden SD, Liu Y, Hair GA, Helms JA, Hu D, Racine M, Nanes MS, Titus L: LMP-1, a LIM-
domain protein, mediates BMP-6 effects on bone formation. Endocrinology 1998;139:
51255134.
Gregory A. Helm, MD, PhD
Department of Neurosurgery, University of Virginia Health System
PO Box 800212, Charlottesville, VA 22908 (USA)
Tel. 1 434 924 1642, Fax 1 434 924 2994, E-Mail [email protected]
Helm/Anderson 28
dramroo
Bone Substitutes
Karin R. Swartz, Gregory R. Trost
University of Wisconsin-Madison, Madison, Wisc., USA
There has long been a need to minimize the morbidity associated with
spinal fusions. Bone substitutes are the most recent focus of this process, in an
attempt to minimize the morbidities of donor site harvesting without compro
mising fusion. Autograft bone, consisting primarily of cancellous and some cor
tical bone, is the current standard for spinal fusion constructs. However,
harvesting autograft usually requires a second incision and involves a risk of
associated morbidities, such as infection, blood loss, hematoma formation, neu
rologic and vascular damage, pain, joint destabilization, and fractures.
Furthermore there is a limit to the size, shape, and volumes of harvested auto
graft. Methods for strengthening the spine and improving nonunion/delayed
healing rates are actively being pursued.
The ideal graft should be biocompatible, osteoinductive, and resorbable.
This allows for incorporation with the surrounding tissues, including vascular
ization [1, 2]. An ideal graft should be malleable, but once in vivo, should not
flow out of the graft bed [3]. It must restore the spine to its functional state [2]
with pain relief [4], allowing for decreased disability. Furthermore, an ideal
graft must be sterile, to abolish the risk of disease transmission. Finally, it
should be easy to use and cost-effective.
Physiology
Bone development occurs via several processes, specifically osteogenesis,

osteoinduction, and osteoconduction [2]. Osteogenesis is the creation of bone,
primarily occurring through the bone-building capabilities of osteoblasts, the
deposition by osteoblasts of rigid extracellular matrix (ECM), and the eventual
mineralization of the ECM into bone. The osteoblastic activity is normally
balanced by the bone-resorbing function of osteoclasts, which also allows for
dramroo
remodeling. Bone remodeling, a balance between bone formation (osteoblastic)
and bone resorption (osteoclastic) functions, is important for skeletal growth as
well as maintenance of normal bone structure. Various growth factors (GFs)
have been implicated in the function of maintaining/healing bone, including
transforming growth factor- (TGF-), platelet-derived growth factor, insulin-
like growth factors, fibroblast growth factors, and bone morphogenetic proteins
(BMPs) [3].
Osteoinduction involves active recruitment of mesenchymal stem cells
and facilitation of their differentiation along the osteoblastic lineage [2].
Osteoinduction more commonly occurs in cancellous more so than cortical
bone, and appears to be under the control of BMPs, as opposed to the other
GFs, which are not osteoinductive. Osteoconduction, which involves the ingrowth
of cells to create and support a stable bony environment, is likewise achieved
more readily through cancellous than cortical bone.
Bone Allograft
Cadaveric sources of bone graft are available autograft substitutes that

have undergone sterilization and processing. Tissue processing can weaken the
cadaveric bone graft and affect healing properties. Furthermore, despite pro
cessing, there is a small but recognized concern over the possibility of disease
transmission. Grafts are typically a combination of both cancellous and cortical
bone cancellous for its osteoinductive and osteoconductive properties, corti
cal for its load-bearing capacity. Allografts come in many different sizes
and shapes. As the number of procedures using allograft continues to rise, the
availability will become limited, thus prompting searches for alternatives.
Alternatives to bone banking involve products of tissue engineering,
consisting of natural or synthetic scaffolds with GFs, and of GF therapeutic
techniques that involve direct implantation of cytokines, transduction of genes
encoding cytokines with osteogenic capability, or of transplantation of cultured
osteogenic cells [5].
Tissue Engineering
Tissue engineering applications are potential approaches for the future of

bone repair strategies. These applications rely on GFs for development of new
tissue through chemotaxis, proliferation, differentiation, and new tissue forma
tion [3]. The development of carriers, delivery systems, and specific GFs is
underway.
Swartz/Trost 30
dramroo
Table 1. Commercially available substitutes [adapted from 2]
OrthoBlast Heat-sensitive copolymer with cancellous bone chips and DBM,

available as injectable paste or putty; osteoconductive, bioresorbable, limited
osteoinduction
DynaGraft Heat-sensitive copolymer with DBM, available as injectable gel, matrix,
or putty; osteoconductive, bioresorbable, limited osteoinduction
ProOsteon Coral HA composite; available in granular or block form;
osteoconductive and bioresorbable
Grafton DBM combined with glycerol, available as gel; osteoconductive,
bioresorbable, limited osteoinduction
OSTEOSET Surgical grade calcium sulfate; available as various-sized pellets;
osteoconductive, bioresorbable
AlloMatrix DBM with surgical grade calcium sulfate powder; available as injectable or
formable putty; osteoconductive, bioresorbable, limited
osteoinduction
Collagraft Mixture of HA, tricalcium phosphate, and bovine collagen; available as strip
configurations; osteoconductive, bioresorbable, limited
osteoinduction when mixed with bone marrow
Vitoss Surgical grade tricalcium phosphate; available as blocks or morsels;
osteoinductive, bioresorbable
DBM Demineralized bone matrix; HA hydroxyapatite.
Carriers
Carrier substances can be divided into four major categories: inorganic
materials (bioglasses and ceramics), synthetic polymers (especially poly
methylmethacrylate), natural polymers (primarily collagen), and composites
(various) (table 1) [3]. Collagen combined with hydroxyapatite is one of the
most widely used first-generation composite carriers, which is both flexible and
able to bond directly with bone [3]. As scaffolds, carriers must act as permis
sive environments for osteoconduction to be successful [3] (table 2).
Hyaff-11
Hyaff-11 is a new semisynthetic biopolymer that can be fabricated into
porous scaffolds for tissue reconstruction. It is a benzyl ester derivative of the poly
saccharide hyaluronic acid, a natural component of the ECM. Hyaluronic acid
plays a role in proteoglycan organization, cell differentiation, and wound healing.
It has well-established biocompatibility, and is biodegradable in 812 weeks in
vivo [6]. In a study by Kim and Valenti [6], Hyaff-11 was used as scaffolding for
the delivery of rhBMP-2, and showed low constitutive levels of release of BMP-2,
with good retention of the morphogenetic protein at the site, with success based on
the resultant increased expression of osteoblasts with bone formation.
Bone Substitutes 31
dramroo
Table 2. Pros and cons of carrier substances [adapted from 3]
Substance Inorganic materials Synthetic polymers Natural polymers Composites
Pros Similar structureto bone Reproducible manufacture Excellent biocompatibility Benefits from
Resorbable or Readily tailored May have natural affinity different materials
nonresorbable Controlled release for GFs exploited
Affinity for BMPs Ease of sterilization
Cons Brittle/difficult Breakdown products Pathogen transmission Complex manufacture
to mold might be inflammatory Sterilization
May be exothermic Solvents or cross-linkers Potential immunogenicity
might denature proteins
Nonphysiologic material
Examples Porous coralline HA Poly (-hydroxy acid) Fibrin glue Collagen-TCP
B-tricalcium Polypropylene fumarate Collagen (type I) Collagen-HA
phosphates (B-TCP) Polyanhydrides Chitosan TCP-cellulose
Hyaluronic acid Polyphosphazenes Hyaluronic acid
Calcium phosphate Polozamers Gelatin
cements DBM
Metals
Calcium sulfates
HA Hydroxyapatite; DBM demineralized bone matrix.
PLA-DX-PEG
Saito et al. [5] developed a biodegradable synthetic polymer to deliver
BMP-2. PLA-DX-PEG (poly-D,L-lactic acid-p-dioxanone-polyethylene
glycol) is available as a block copolymer that is biocompatible and biodegrad
able. It offers an alternative to bone banking with no disease transmission, no
evidence of immunoreaction or foreign body reaction, and is more effective
than collagen-BMP-2 systems in regard to calcium content of formed ossi
cles. Also, it swells a little when in contact with water and thus fills the dead
space at the bone-implant interface and provides a layer of scaffolding for
new bone [5].
Polymethylmethacrylate
Polymethylmethacrylate is the only available cement with reports of clinical
application and experience for vertebroplasty in the treatment of osteoporosis [1].
Polyetheretherketone
Polyetheretherketone (PEEK) is a radiolucent, biocompatible synthetic
polymer that purports similar stiffness to cortical bone. It is a polyaromatic
Swartz/Trost 32
dramroo
semicrystalline thermoplastic which is available as a cage (Stryker Solis) or
in a pellet form for injection molding or machined cage (Invibio PEEK
OPTIMA). When in cage form, it can be packed with either auto- or allograft
to facilitate fusion.
Poly(L-Lactic Acid)
Poly(L-lactic acid) cages are radiolucent bioabsorbable cages that
demonstrate higher elasticity and thus less stiffness than metal cages. They have
been proven mechanically sufficient and resorb after providing support for an
average of 6 months [7].
Demineralized bone matrix products (ex: Osteofil) are other osteoinduc
tive graft substitutes. No current objective clinical data [8] is available.
Delivery Systems
Techniques for application of GFs are many, with the most commonly
cited described next: (1) delivery of DNA encoding a GF or in vivo gene
therapy, (2) delivery of a cell to produce the GF or ex vivo gene therapy also
called cell therapy, and (3) delivery of the protein itself via some carrier
matrix [3].
The latter of these has advanced the furthest, with ongoing attempts to
discover optimal carriers for GFs. These carriers would allow controlled release
(appropriate kinetics) with therapeutic dosing and would act as sturdy scaffolds
[3], thus encouraging cells to migrate and differentiate (osteoconduction and
osteoinduction), as well as allow ingrowth of blood vessels, passage of nutri
ents, and expulsion of waste products [3]. Degradation and breakdown of the
carrier matrix should either be inert or beneficial to healing, and should be
synchronized with the rate of bone regeneration: too slow, and bone growth/
remodeling is retarded and carries a risk of encapsulation, and too fast, and the
definition of optimal shape may be lost [3].
Gene therapy, in its most generic terms, can be divided into in vivo or
ex vivo techniques to deliver a gene sequence for the osteoinductive factor (rather
than the factor itself) [9]. Both involve insertion of foreign cDNA directly into
a host (in vivo) or into donor cells (ex vivo) via a vector, ultimately forming
cells whose primary purpose is to produce a desired GF [3]. In the in vivo
model, also known as in situ therapy, a vector is directly introduced into host
cells inside the body, the vector acting as the carrier. In the ex vivo model, also
known as cell therapy, the cells can either be harvested from the patient or
another donor; that cell then becomes the carrier for the secreted GF (using
tissue culture expansion and gene transduction), and then is delivered into
the host [3, 10]. The success of gene therapy relies on accurate transduction,
transcription, translation, and expression (table 3).
Bone Substitutes 33
dramroo
Table 3. Pros and cons of gene therapy techniques
Pros Cons
Ex vivo No viral particle transmission More technically demanding

No DNA complexes Cost
Higher efficiency of transduction
In vivo Easier Difficult targeting specific cells
No harvesting Transduction efficiency
Potentially lower cost Vector diffusion from site
Host response to viral proteins
Growth Factors
The current trend in osteoinductive bone graft substitutes for minimally
invasive spine fusions includes providing GF cytokines in addition to the scaf
folding of the carriers. Categories of factors used in these substitutes include
purified bone GFs, recombinant bone GFs, and gene therapy-delivered GFs.
The working current hypothesis is that a burst followed by sustained factor
release is ideal, allowing for cellular migration to the area, with eventual com
mitment to osteoblastic lineages with propagation of cellular numbers [3].
An example of a GF invested as an osteoinductive bone graft is BMP.
Bone Morphogenetic Proteins

BMPs have been proposed to revolutionize bone surgery. BMPs are
members of the TGF- family, which guides the differentiation of stem cells
down specific lineages (i.e., bone, cartilage) [6, 8, 1117]. Urist [18] initially
discovered that devitalized bone could induce ectopic bone in rodents, and then
went on to isolate the protein from demineralized bone matrix; others went on to
clone these bone-inducing substances, and currently over forty have been iden
tified [19]. They act locally, and thus must be delivered directly to the site
needed. BMPs trigger osteoblastic differentiation of marrow mesenchymal stem
cells, preosteoblasts, and undifferentiated mesenchymal cells [6, 8, 1117].
BMP-2 is the most widely studied, and is one of the most potent osteoin
ductive of the BMPs [6, 8, 1114, 16, 18, 20, 21]. Other chondrogenic/potentially
osteogenic morphogenetic proteins include BMP-4 [13], BMP-6 [17], BMP-7
(aka osteogenic protein-1, OP-1) [15], and BMP-9 [11]. BMPs have many roles
besides osteogenesis: BMP-4 has been implicated as proapoptotic in somatic
development [22], BMP-9 in fetal and adult liver function [23] as well as in
CNS development [24], BMP-2, BMP-7 and BMP-4 in CNS development [25],
and BMP-4 and BMP-7 in lung development [26].
Swartz/Trost 34
dramroo
Sustained production of high levels of activated BMP-2 using recombinant
adenovirus type 5 (Ad5BMP-2) has been achieved by Olmsted et al. [14]. In their
research, they were able to elicit BMP synthesis upon infection even by non
osteoprogenitor cells, which was 3-fold more active than equivalent concentrations
of rhBMP-2 [14].
The issue of physical containment of BMP still needs to be optimized, such
that it remains in place, without causing unwanted bone growth that could
potentially cause neurovascular compromise.
Preclinical Data
Animal models, especially small animals, have shown success with treat
ment of fractures and segmental defects [3]; Murata et al. [27] showed successful
rapid bone growth from an onlay over bone without periosteum, using rhBMP-2
and bioabsorbable atelocollagen in rats. Ripamonti et al. [28] used BMP-7
(OP-1) to grow bone over a cranial defect in baboons. Boden [29] developed a
rabbit model to characterize the healing process, and sequenced a novel cDNA
encoding for LMP-1 (LIM-mineralizing protein-1, an intracellular protein which
induces osteoblastic differentiation); when implanted locally in the bone marrow
transfected with LMP-1 cDNA, he reported 100% fusion. Majumdar et al. [11,
30] isolated multipotential mesenchymal cells from the bone marrow and cul
tured them in the presence of TGF-3 and recombinant human BMPs (rhBMP-2
and rhBMP-9) causing chondrogenic differentiation, a technique that could allow
for delivery of chondrogenic GFs to the site of articular cartilage repair. Noshi et
al. [16] further advanced the bone-forming capability of a composite using mar
row mesenchymal stem cells and porous hydroxyapatite by adding rhBMP-2 to
the composite, which resulted in accelerated bone formation.
Clinical Trials
Boden et al. [20] were the first to demonstrate the clinical efficacy of
rhBMP-2 for single-level spinal arthrodesis, comparing tapered titanium fusion
cages packed with either rhBMP-2 (1.5 mg/ml) soaked onto bovine collagen
sponge or with morselized iliac autograft. Success was based on shorter hospi
tal stays, shorter operative time, less blood loss, and improved fusion in the
cage-BMP construct.
Sandhu [8] performed anterior lumbar interbody fusions using combined
threaded titanium cages filled with autograft and compared with cages filled
with rhBMP-2, and found the cages filled with rhBMP-2 had better outcomes,
including high fusion rates, shorter operative time, and shorter hospital stays.
Bone Substitutes 35
dramroo
Table 4. Comparisons of autograft, allograft, and substitutes
Bone graft Structure/ Osteoconduction Osteoinduction Osteogenesis Price

strength
Autograft
Cancellous No
Cortical
Allograft
Cancellous
Frozen No No
Freeze-dry No No
Cortical
Frozen No No
Freeze-dry No No
Substitutes
OrthoBlast No No 10 cm3: USD 850/
DynaGraft No No 5 cm3 USD 470
ProOsteon 500R ? No No 10 cm3: USD 850/
Grafton No No 5 cm3 USD 470
OSTEOSET No No No 10 cm3: USD 1,150
AlloMatrix No No 6 strips: USD 1,109/
Collagraft No No 3 strips: USD 683
Vitoss No No No 10 cm3: USD 525
Friedlaender et al. [15] evaluated rhOP-1 (BMP-7), comparing clinical and

radiographic results in the treatment of tibial nonunion in a controlled, prospec
tive, randomized, and partially blinded clinical trial. They inserted intramedullary
rods packed with either rhOP-1 in a type I collagen carrier or with fresh bone
autograft. They determined there were no statistically significant differences in
the results at 9 months, with 81 and 85% clinical success (respective) and 75
and 84% radiographic success (respective). Interestingly, postoperative
osteomyelitis rates at the nonunion site were greater in the autograft group
(21% in autograft patients, 3% in OP-1 patients).
Other Issues
Different anatomical sites may require different substitutes. Special con
sideration must be given to variables of the involved kinetics with different
locations, the effects of medications, the impact of disease processes, and of
aging on the efficacy of function [2, 3, 6].
Objective measures determining utility of any bone-stabilizing product must
certainly include cost containment issues centered around operating room func
tions, including time required for the procedure, duration of needed anesthesia,
supplies used/needed, and length of stay in recovery and in hospital (table 4).
Swartz/Trost 36
dramroo
As always, success of fusion depends on many factors, but is maximized
by the basic tenets of properly prepared sites, exposed cancellous bone,
mechanical loading and stability.
References
1 Heini PF, Berlemann U: Bone substitutes in vertebroplasty. Eur Spine J 2001;10:S205S213.

2 Greenwald AS, Boden SD, Goldberg VM, Khan Y, Laurencin CT, Rosier RN, for the Committee
on Biological Implants: Bone-graft substitutes: Facts, fictions, and applications. J Bone Joint Surg
Am 2001;83:S98S103.
3 Li RH, Wozney JM: Delivering on the promise of bone morphogenetic proteins. Trends
Biotechnol 2001;19:255265.
4 Jackson DW, Scheer MJ, Simon TM: Cartilage substitutes: Overview of basic science and treat
ment options. J Am Acad Orthop Surg 2001;9/1:3752.
5 Saito N, Okada T, Horiuchi H, Murakami N, Takahashi J, Nawata M, Ota H, Nozaki K, Takaoka K:
A biodegradable polymer as a cytokine delivery system for inducing bone formation. Nat
Biotechnol 2001;19:332335.
6 Kim HD, Valenti RF: Retention and activity of BMP-2 in hyaluronic acid-based scaffolds in vitro.
J Biomed Mater Res 2002;59:573584.
7 Van Dijk M, Smit TH, Sugihara S, Burger EH, Wuisman PI: The effect of cage stiffness on the
rate of lumbar interbody fusion. Spine 2002;27:682688.
8 Sandhu HS: Anterior lumbar interbody fusion with osteoinductive growth factors. Clin Orthop
2000;371:5660.
9 Boden SD, Hair GA, Viggeswarapu M, Liu Y, Titus L: Gene therapy for spine fusion. Clin Orthop
2000;379S:S225S233.
10 Oakes DA, Lieberman JR: Osteoinductive applications of regional gene therapy: Ex vivo gene
transfer. Clin Orthop 2000;379S:S101S112.
11 Majumdar MK, Wang E, Morris EA: BMP-2 and BMP-9 promote chondrogenic differentiation of
human multipotential mesenchymal cells and overcome the inhibitory effects of IL-1. J Cell
Physiol 2001;189:275284.
12 Kingsley DM: What do BMPs do in mammals? Clues from the mouse short-ear mutation. Trends
Genet 1994;10:1621.
13 Akiyama H, Shukunami C, Nakamura T, Hiraki Y: Differential expressions of BMP family genes
during chondrogenic differentiation in mouse ATDC5 cells. Cell Struct Funct 2000;25:195204.
14 Olmsted EA, Blum JS, Rill D, Yotnda P, Gugala Z, Lindsey RW, Davis AR: Adenovirus-mediated
BMP2 expression in human bone marrow stromal cells. J Cell Biochem 2001;82:1121.
15 Friedlaender GE, Perry CR, Cole JD, Cook SD, Cierny G, Muschler GF, Zych GA, Calhoun JH,
LaForte AJ, Yin S: Osteogenic protein-1 (bone morphogenetic protein-7) in the treatment of tibial
nonunions. J Bone Joint Surg Am 2001;83:S151S158.
16 Noshi T, Yoshikawa T, Dohi Y, Ikeuchi M, Horiuchi K, Ichijima K, Sugimura M, Yonemasu K,
Ohgushi H: Recombinant human bone morphogenetic-2 potentiates the in vivo osteogenic ability
of marrow/hydroxyapatite composites. Artif Organs 2001;25:201208.
17 Boskey AL, Paschalis EP, Binderman I, Doty SB: BMP-6 accelerates both chondrogenesis and
mineral maturation in differentiating chick limb-bud mesenchymal cell cultures. J Cell Biochem
2002;84:509519.
18 Urist MR: Bone: Formation by autoinduction. Science 1965;150:893899.
19 Wang EA, Rosen V, Cordes P, Hewick RM, Kriz MF, Luxenberg DP, Sibley BS, Wozney JM:
Purification and characterization of other distinct bone-inducing factors. Proc Natl Acad Sci USA
1988;85:94849488.
20 Boden SD, Zdeblick TA, Sandhu HS, Heim SE: The use of rhBMP-2 in interbody fusion
cages: Definitive evidence of osteoinduction in humans: A preliminary report. Spine 2000;25:
376381.
Bone Substitutes 37
dramroo
21 Subach BR, Haid RW, Rodts GE, Kaiser MG: Bone morphogenetic protein in spinal fusion:
Overview and clinical update. Neurosurg Focus 2001;10/4:16.
22 Sanders EJ, Parker E: Ablation of axial structures activates apoptotic pathways in somite cells of
the chick embryo. Anat Embryol 2001;204:389398.
23 Miller AF, Harvey SA, Thies RS, Olson MS: Bone morphogenetic protein-9. An autocrine/
paracrine cytokine in the liver. J Biol Chem 2000;275:1793717945.
24 Lopez-Coviella I, Berse B, Krauss R, Thies RS, Blusztajn JK: Induction and maintenance of the
neuronal cholinergic phenotype in the CNS by BMP-9. Science 2000;289:313316.
25 Jordan J, Bottner M, Schluesener HJ, Unsicker K, Krieglstein K: Bone morphogenetic proteins:
Neurotrophic roles for midbrain dopaminergic neurons and implications of astroglial cells. Eur J
Neurosci 1997;9:16991709.
26 De Caestecker M, Meyrick B: Bone morphogenetic proteins, genetics, and the pathophysiology of
primary pulmonary hypertension. Respir Res 2001;2/4:193197.
27 Murata M, Maki F, Sato D, Shibata T, Arisue M: Bone augmentation by onlay implant using
recombinant human BMP-2 and collagen on adult rat skull without periosteum. Clin Oral Impl
Res 2000;11:289295.
28 Ripamonti U, van den Heever B, Sampath TK, Tucker MM, Rueger DC, Redoli AH: Complete
regeneration of bone in the baboon by recombinant human osteogenic protein-1 (hOP-1, bone
morphogenetic protein-7). Growth Factors 1996;13:273289.
29 Boden SD: Biology of lumbar spine fusion and use of bone graft substitutes: Present, future, and
next generation. Tissue Eng 2000;6:383399.
30 Majumdar MK, Banks V, Peluso DP, Morris EA: Isolation, characterization, and chondrogenic
potential of human bone marrow-derived multipotential stromal cells. J Cell Physiol 2000;185:
98106.
Gregory R. Trost, MD
Department of Neurological Surgery
University of Wisconsin-Madison
600 Highland Ave, Madison, WI 53792-3232 (USA)
Tel. 1 608 263 1410, E-Mail [email protected]
Swartz/Trost 38
dramroo
Resorbable Technology for
Spinal Stabilization
Daniel K. Resnicka, Joseph T. Alexanderb, William C. Welchc

a
Department of Neurological Surgery, University of Wisconsin, K4/834 Clinical
Science Center, Madison, Wisc.,
b
Department of Neurological Surgery, Wake Forest University, Winston-Salem,
N.C., and
c
Department of Neurological Surgery, University of Pittsburgh, Pittsburgh, Pa., USA
Incredible advances in the surgeons ability to restore stability to the unsta

ble spine have been made in the last decade. Using metal implants, surgeons can
now effectively stabilize any motion segment of the spine. With the concomitant
use of autograft or allograft bone, high fusion rates can be achieved in the vast
majority of cases. There are, however, drawbacks to the use of such systems.
Autograft bone is not always available for a given application and autograft
harvest is associated with nontrivial morbidity in a substantial fraction of
patients [1]. The use of allograft bone is associated with lower fusion rates than
autograft bone in most cases. The use of allograft is often limited by both cost
and availability. Machining of allograft bone to predetermined shapes is a tech
nology- and labor-intensive process, and processing of the bone for infection
control and mechanical consistency is expensive. Finally, the use of metallic
implants is associated with late complications related to implant failure, erosion
into soft tissue structures, and degradation of the bone/implant interface.
In response to these drawbacks, several companies have developed resorbable
biomaterials that can serve as temporary fixation devices, structural supports,
and osteoconductive or even osteoinductive conduits for new bone growth.
These products offer the potential advantages of unlimited supply, significant
cost savings, and a reduction in patient morbidity. This chapter is intended to
serve as an overview of some of these new biomaterials and their potential
applications. This overview will be incomplete, as the pace of new development
in this field is staggering. Many of the substances and devices presented are not
FDA approved or are FDA approved for uses outside of the spine. The reader is
dramroo
cautioned to review the FDA-approved indications for the use of any of the
described products prior to the experimental or therapeutic use of any of the
products described.
Resorbable Synthetic Bone Substitutes
Autograft bone remains the gold standard for spinal fusion procedures,
especially lumbar posterolateral fusion. Autograft bone is osteogenic, in that it
contains the cellular components responsible for new bone formation.
Autograft is osteoinductive, in that it contains diffusible proteins which induce
neighboring cells to produce bone. Autograft is also osteoconductive, in that it
has a structure which allows for the migration of osteogenic cells which can
then form new bone [2]. Depending upon the size, shape, and source of the
autograft, mechanical strength may be negligible (cancellous) or substantial
(tricortical iliac crest, fibula, rib). Drawbacks to the use of autograft include
harvest site morbidity as well as limited availability. The use of allograft bone
avoids the problem of graft site morbidity, and allograft humeral, femoral, and
patellar grafts are available for use as large load-bearing struts. Allograft bone
is osteoconductive, and in some preparations may be somewhat osteoinductive
(such as in demineralized bone matrix). Unfortunately, allograft bone is not
always readily available. Allograft is expensive to preshape into readily used
sizes. Finally, some patients may refuse the use of allograft for religious reasons
or because of fear of infection.
Although there are many alternatively prepared allograft products
(Grafton, Dynagraft, Osteofil) and xenograft products (Collagraft, True
Bone Ceramics) available for use which may extend the utility of allograft
use, the manufacture of these products still requires the use of processed donor
(human or animal) bone. In some cases, the extensive preparation of these
products has resulted in unexpected toxicity. For example, Botsman et al. [3]
and Wang et al. [4] recently presented evidence to suggest that the glycerol car
rier in Grafton (Osteotech, Eatontown, N.J., USA) may be toxic in large doses.
This discussion will center on synthetic resorbable bone substitutes.
The use of synthetic bone graft material would negate the risks of harvest
site morbidity, disease transmission, and limited donor availability. The ability
to machine the material would allow for the creation of multiple shapes and
sizes that could be used as off-the-shelf bone substitutes. Substantial cost sav
ings may be realized with reasonably priced implants if the costs of autograft
harvest (estimated to be USD 1,5005,000 for iliac crest [5, 6]) and allograft
harvest and preparation (machined allograft bone wedge approximately
USD 3,000 per level, allograft femoral ring USD 1,2002,400 per level;
Resnick/Alexander/Welch 40
dramroo
source: University of Wisconsin operating room) are avoided. Ideally, these
synthetic bone grafts would be totally absorbed by the body and replaced by
native bone.
Calcium sulfate (plaster of Paris) has been used for bone defect repair
since 1892 [7, 8]. Peltier and Jones [9] used calcium sulfate pellets to fill uni
cameral bone cysts in 26 patients with excellent results, and Coetzee [10] used
similar pellets to treat osseous defects in the skull and facial bones in 110
patients. Recently, Kelly et al. [7] reported the use of Osteoset (Wright Medical
Technology, Tenn., USA) a highly processed form of calcium sulfate for repair
of bone defects in a series of 109 patients with various bone lesions. Osteoset
was used as a bone graft substitute if the surgeon recommended the use of
morsellized graft, the bone void was not intrinsic to the stability of the struc
ture, and the void shape could accommodate the calcium sulfate pellets. Overall
results were excellent, with 100% pellet resorption and 94% new bone growth
(radiographic criteria) noted at 1 year [7]. No spinal fusions were performed
during the study, and other materials were used in addition to the pellets in 65%
of cases. Hadjipavlou et al. [11] recently reported excellent results with calcium
sulfate-filled titanium cages in an interbody fusion model in adult sheep. These
authors obtained roughly equivalent results with cages filled with either autograft
iliac crest or with calcium sulfate. There are at present, however, no published
reports of the use of calcium sulfate for human spinal fusion.
Processed coralline grafts, including hydroxyapatite, have been used exten
sively for the repair of bone defects. Additionally, these products have been used
for spine surgery with a number of different applications. Hydroxyapatite is a
poorly crystalline calcium phosphate compound which is similar in mineral
composition to bone [6]. Calcium phosphate and hydroxyapatite are osteocon
ductive, if the porosity, pore size, and degree of pore interconnectivity are
optimized [6]. Substantial research has established that the ideal pore size is
between 100 and 500 m [12]. Calcium hydroxyapatite is most commonly
formed through processing of coral, which contains calcium phosphate and
calcium carbonate in the form of aragonite. Several natural corals, particularly
Porites and Goniopora, have pore sizes and pore interconnectivity which allow
for osteoconduction. These corals have been harvested, cleansed, and then used
as graft material (Biocoral Inoteb, France), particularly in dental surgery [13].
These corals may also be chemically treated via the replamineform process.
This process results in the conversion of aragonite to hydroxyapatite without
changing the three-dimensional structure of the coral [6]. ProOsteon 200R,
500R, and Interpore porous hydroxyapatite (Interpore Cross International,
Calif., USA) are produced in this fashion. The numbers 200R and 500R refer to
the nominal pore diameter of the crystalline structure, either 200 or 500 m.
The mechanical properties of coralline hydroxyapatite grafts are more similar
Resorbable Technology for Spinal Stabilization 41

dramroo
to cancellous rather than cortical bone. In general, increasing porosity increases
the osteoconductivity of the graft and lowers the compressive strength.
Coralline hydroxyapatite has been used as a bone substitute for both cer
vical interbody as well as lumbar posterolateral fusion in animal, and in some
cases human trials. In 1996, Zdeblick et al. [14] studied the use of ProOsteon
500 R in a multilevel anterior cervical fusion in a goat model with or without
plate fixation. They found that although only 48% of grafts incorporated with
the hydroxyapatite alone, 71% incorporated if anterior cervical plate fixation
was added. However, none of the implants were completely replaced by host
bone at 12 weeks. Biomechanical testing revealed that the hydroxyapatite-fused
segments behaved poorly without plate fixation. However the addition of an
anterior cervical plate improved performance such that segments fused with
hydroxyapatite and a cervical plate behaved similar to segments fused with
allograft (but not autograft) bone and cervical plate in torsion. Hydroxyapatite-
fused segments were inferior to both autograft and allograft in flexion/extension
testing [14]. Boden et al. [15] studied the use of hydroxyapatite in a rabbit
model of lumbar posterolateral fusion. They found that the use of coralline
hydroxyapatite alone resulted in 0/14 solid fusions. Combination of hydroxy
apatite and autograft bone in a 1:1 ratio resulted in a fusion rate of 50% (7/14).
Bozic et al. [16] found that the combination of hydroxyapatite and bone mar
row aspirate resulted in 25% fusion rate in a rabbit model. Addition of electri
cal stimulation (100 A implantable direct current stimulator) significantly
improved fusion rates. Similarly, Baramki et al. [17] noted a significantly lower
fusion rate for hydroxyapatite-grafted animals (50%) than autograft-grafted
animals (100%) in a sheep model of posterolateral fusion.
Despite these rather lukewarm animal results, Thalgott et al. [18] used
hydroxyapatite (ProOsteon 200R) in combination with an anterior cervical
plate for cervical interbody fusions in a group of 26 patients. In this retrospec
tive series of 26 patients, 41 disc spaces were considered incorporated at
24 months. Radiographic fusion was impossible to judge because the graft
material was not resorbed at 24 months and remained more radio-dense than
native bone. The criteria used for the determination of incorporation was loss
of a radiolucent line around the graft. The validity of this radiographic criterion
for the determination of bony fusion is questionable. Iseda et al. [19] found that
the appearance or disappearance radiolucent stripe around a hydroxyapatite
graft did not correlate with changes in the 99mTc-HMDP uptake ratio following
surgery in a small group of patients.
Tricalcium phosphate is a synthetically produced ceramic which has a
porous structure amenable to osteoconduction. It has been used alone and in
combination with hydroxyapatite as a bone graft substitute. It is completely
resorbed by the body and is gradually replaced by new bone formation [1].
dramroo
It is virtually identical to bone in terms of its mineral composition, and no
reports of systemic toxicity exist. A comparative study by Emery et al. [20]
demonstrated that the use of tricalcium phosphate (pore size 400 m) resulted
in superior histological results compared to hydroxyapatite or calcium carbon
ate in a canine interbody fusion model. Toth et al. [21] studied the use of
different -tricalcium phosphate/hydroxyapatite preparations in a goat cervical
interbody fusion model and found that -tricalcium phosphate performed as
well or better than autograft at 3 and 6 months following fusion (radiographic,
histological, dual-energy x-ray absorptiometry, and biomechanical testing).
There was a definite increase in the bony union rate at 3 months with increas
ing porosity (0% union rate for autograft and 30% porous ceramic, 67% for
50% porous ceramic, and 83% for 70% porous ceramic). Delecrin et al. [1]
used a similar compound (Triosite Zimmer, France) in a study of 58 patients
undergoing surgery for scoliosis. In a prospective randomized study, these
authors found that patients who received the ceramic graft had less blood loss
and had no graft site complications. The authors noted that half of the autograft
group experienced pain at the donor site 6 months following surgery. There
were no significant differences between autograft and ceramic groups in terms
of radiographic appearance or functional outcome at a minimum follow-up of
24 months (mean follow-up 48 months) [1].
A pure -tricalcium phosphate synthetic bone product has been recently
introduced (Vitoss, Orthovita, Pa., USA). Vitoss has a unique crystalline
structure, imparted by the manufacturing process, which increases its porosity
significantly. The porosity of Vitoss is between 88 and 92%, compared to an
approximate 55% porosity of hydroxyapatite-calcium carbonate. The increased
porosity of this product allows for the rapid penetration of blood and blood
products as well as for the migration of osteoblasts. Animal experiments using
a humeral defect model have confirmed that the tricalcium phosphate is rapidly
and nearly completely replaced by native bone in a matter of weeks (fig. 1, 2).
The formation of bone in the humeral model is more rapid and more complete
than that seen with coralline hydroxyapatite bone substitutes (fig. 1). In time, the
graft is completely resorbed and replaced by new bone. When combined with
autogenous bone marrow, the osteoinductive properties of the tricalcium phos
phate are superior to demineralized bone matrix in a rat Urist pouch test [22].
A similar -tricalcium phosphate compound has been developed (but is not yet
FDA approved) that has biomechanical properties similar to cortical bone, can
be machined and shaped easily, and can even be delivered in a semiliquid or
putty form (Cortoss, Orthovita, Pa., USA).
Several other types of compounds have been used as synthetic bone substi
tutes. Madawi et al. [23] studied a bioactive osteoconductive polymer made of
methyl methacrylate, calcium gluconate, and polyamide fibers. These authors

dramroo
36 Bone at center of implant
Vitoss scaffold
30 Average adjacent bone
Volume (%)
24
18
12
0
0 13 26 39 52
Time (weeks)
Fig. 1. Rapid absorption of Vitoss with regrowth of new bone within the graft.
a b
Fig. 2. Vitoss tricalcium phosphate bone substitute in a canine humeral defect model.
The implant is almost completely resorbed at 6 weeks postimplant (a) and is indistinguish
able from normal bone radiographically and histologically at 12 weeks (b).
noted favorable clinical results with this compound in trial of cervical interbody
fusion; however, the graft never radiographically incorporated into native bone.
Schulte et al. [24] used a bioglass-polyurethane composite to replace vertebral
bodies following corpectomy for metastatic lesions in 5 patients. All patients
also underwent titanium plate fixation. No instances of implant failure were
noted. Roessler et al. [25] reported sobering results in their clinical study of a
resorbable compound made up of polyethylene oxide and polybutylene tere
phthalate (Polyactive 70/30 HC Implants, The Netherlands) for reconstitution
of iliac crest bone graft harvest sites. Although this substance had shown very
promising results in multiple animal models, human trials failed to show a favor
able effect on clinical or radiographic outcome. This study serves as a reminder
that there is no substitute for properly designed human clinical trials for the
dramroo
development of bone graft substitutes. Differences in the healing capabilities
of the animal species, influences of patient age, and differences in the graft
environment all play a role in bone healing [25, 26].
Resorbable Fixation Devices
Implants placed during spinal fusion surgery function in various ways.

Screw/rod or screw/plate systems function to provide immediate structural sta
bility, thus allowing early mobilization of the patient, maintaining spinal align
ment and promoting the ultimate arthrodesis. Traditionally, these systems were
manufactured of stainless steel, which has more recently been largely replaced
by titanium alloy. Although both of these metals have performed well in their
primary function, there are drawbacks such as stress shielding due to the exces
sive rigidity and permanence of the constructs that can in turn lead to bone
resorption and osteopenia [27, 28]. Corrosion, wear debris and even rare aller
gic reactions have been seen, more so with stainless steel implants. Metallic
artifact on imaging studies can obscure anatomic detail and cause diagnostic
dilemmas. Once arthrodesis has occurred, the fixation systems no longer have
a purpose [27, 29]. The indications for explantation of spinal fixation hardware
are still controversial, but it is commonly performed.
Bioresorbable polymers have been increasingly explored as replacements
for metal, bone and nonresorbable synthetic materials in recent years. Obviously,
biodegradable implants eliminate the need for explantation. In addition, they
can reduce stress shielding, by having a better match of strength and elasticity
to bone as well as by gradually reducing load sharing as structural integrity is
lost [27, 28, 30]. Lastly, they avoid the problems of metal corrosion, debris, and
imaging artifact. This section will explore the properties and potential uses of
these polymers in spinal surgery.
Bioresorbable materials and plates have been used to repair orbital and
mandibular fractures since 1972 [31, 32]. Polylactic acid (L-isomer) (PLLA)
and polyglycolic acid (PGA) and other polyhydroxy acids have been identified
in the research to be promising polymers for resorbable implantable devices [33].
PLLA alone is resistant to hydrolytic degradation, which makes implants of this
polymer quite strong but takes years for resorption to occur, if it occurs at all.
These implants may also result in a foreign body reaction over time and thus
require removal [34, 35]. In comparison, pure PGA implants absorb much more
quickly when exposed to moisture yet suffer loss of tensile strength and rapid
polymer breakdown [36]. Researchers have experimented with combinations of
both polymers in order to utilize the favorable characteristics of each. LactoSorb
(Biomet, Warsaw, Ind., USA) is a resorbable plate system approved for use in

dramroo
Polyhydroxy acid Monoacid Pyruvate
Krebs TCA cycle
CO2 H2O
Fig. 3. Schematic illustration of the degradation of a prototypical resorbable polymer.

(Certain intermediate products of degradation can also be excreted in the urine.)
craniomaxillofacial surgery. It consists of a combination of synthetic PGA and

PLLA. This PGA/PLLA copolymer is manufactured in an 82% PLLA and 18%
PGA ratio. The increased amount of PLLA in this copolymer gives it degrada
tion characteristics midway between the predominately PGA products and pure
PLLA implants [37]. MacroPore (MacroPore., San Diego, Calif., USA), another
resorbable product with potential spine applications, is produced by the combi
nation of PLLA and a noncrystalline copolymer mixture of poly(L-lactide) and
poly(D,L-lactide).
The implant resorption time is most strongly affected by the chemical
composition of the polymer, but there are many additional variables in this
process [29, 38]. Gogolewski [28] has identified the implants physical struc
ture and mass, polymeric molecular weight, chain orientation, and presence of
additives as some of the many factors influencing the rate of degradation, along
with the stress on the implant and the characteristics of the implantation site.
For example, he notes that an implant placed under significant load in a highly
vascularized site will likely degrade at an accelerated rate [28].
The PGA component of the LactoSorb implant is degraded to glycolic acid
by hydrolysis and further degraded to glycine, which may be used for protein
synthesis or further converted to serine. All other implant (LactoSorb and
MacroPore) components are hydrolyzed to lactic acid, which then enters the
tricarboxylic acid cycle (fig. 3) [39]. The use of these plates avoids several
potential complications associated with metallic fixation devices, such as migra
tion, stress shielding, and interference with imaging techniques [39]. Currently,
the most commonly reported side effect is a sterile inflammatory reaction to the
implants [35, 4043].
Local tissue reaction to polyhydroxy acid implants is governed by the
chemical nature of the polymer, the physical characteristics of the implant, and
by its degradation rate. Towards the latter part of the degradation process, the
implant may rapidly lose structural integrity, and the production rate of
dramroo
polymeric debris may exceed the tissue tolerance and transport potential of the
implantation site [27, 38]. This in turn can stimulate what is said to be a non
specific foreign-body reaction rather than a true inflammatory response [27, 38].
Thus, a large implant made of a fast-degrading polymer is likely to produce a
more pronounced inflammation than a small implant composed of a slow-
degrading polymer. PGA polymers have been associated with a higher rate of
tissue reaction, even including the formation of fibrous capsules, sterile cysts
and sinuses. On the other hand, pure polylactic acid (PLA) polymers have a
low or nonexistent rate of tissue reaction, while once again, copolymers are
intermediate [27, 38, 44].
Histological evaluation of these inflammatory sites has shown typical
nonspecific foreign-body reactions. Similar foreign-body reactions were found
on histological examination of implant sites of patients without clinical evi
dence of an inflammatory response [40]. These specimens were obtained upon
reoperation for fixation failure. No explanation is available for why some patients
have clinical signs of the foreign body reaction while others do not. Most stud
ies have shown the inflammatory response, when clinically present, to occur
several months after implantation. Bostman [40] postulates that this time to
occurrence of the reaction is reflective of the estimated degradation time of
PGA and PLLA/PGA implants which is 90120 days. Patients who have had
plates and screws made solely of PLLA, implanted for fixation of zygomatic
fractures, have shown inflammatory reactions from 3.3 to 5.7 years after surgery
[35]. These findings seem to correlate with degradation rates of PLLA, stated
in some studies to be from 2 to over 4 years [39].
PLA polymers and PLA-PGA copolymers are biocompatible with the dura
[4547]. PLA biocompatibility has also been specifically tested in reference to
neural tissue, including brain and spinal cord tissue as well as peripheral nerves.
No effect on neuronal cells, nonneuronal cells or axonal growth has been noted
[45, 48]. No significant toxicity, carcinogenicity, teratogenicity or mutagenesis
has been associated with either PLA or PGA [27, 38].
LactoSorb plates and screws became commercially available for cranio
facial applications in 1996. Most of the interest in bioresorbable plates for cranial
reconstruction has been in the pediatric patient population because of compli
cations with metallic implants related to bone growth. LactoSorb plates have
been used in neurosurgery for fixation of craniotomy flaps, repair of depressed
skull fractures and for repair of craniosynostosis [33, 4953]. Encouragingly, it
has been found that the dissolution and gradual loss of tensile strength of the
devices minimize growth restrictions as well as the potential for transcranial
migration [54]. In addition, good results have been obtained in both the pedi
atric and adult population when LactoSorb plates and screws were used for
cranial fixation and reconstruction.

dramroo
LactoSorb plates are designed to withstand similar external forces and
approximate the strength of traditional titanium plates. In order for a biore
sorbable plate to perform similarly to a titanium implant, it must retain similar
strength characteristics long enough for osteosynthesis to occur. Because PGA,
PLLA, and copolymers are all significantly less rigid than titanium, the design
of plate-like implants must be adjusted somewhat. In general, resorbable plate
designs incorporate a thicker plate, in order to recapitulate the mechanical
properties of a standard titanium implant. Flexural strength is significantly
improved in the LactoSorb plate, for example, due to its I-beam-like construc
tion (raised rails along the side of the plate). In vitro studies of flexural strength
show the LactoSorb plate has an initial strength comparable to Lorenz titanium
plates and retains 70% of this strength for 68 weeks [55, 56]. The improved
strength does, however, create a much higher profile implant than the traditional
metallic plate.
The most common reported clinical difficulties with the LactoSorb products
are related to implantation technique. LactoSorb screws are not self-tapping
which increases the time necessary to place the screws. Second, unlike metal
plates, which can be shaped by hand or with instruments at room temperature,
polymer plates must be heated before they can be shaped. Heating is achieved
by use of a bag of calcium chloride which, when injected with sterile water, will
liberate heat for approximately 20 min. By placing the bag around the LactoSorb
plate, the implant can be made malleable. Some authors have stated that the
20-min window may be too small, making a second calcium chloride bag
necessary, which involves adding time and expenses. Despite these initial
technical difficulties, most authors agree that the LactoSorb system is a safe,
effective system for cranial fixation and reconstruction, which, after minimal
experience with the system, adds very little time to the operation compared to
the traditional titanium plating systems.
Six studies of the use of bioresorbable plates for cranial fixation, either
alone or in conjunction with facial reconstruction and orthosis, have been
published with a combined total of 115 patients [33, 4953]. No patients were
noted to have significant complications from inflammatory reactions, seizures
or failure of the fixation. In fact, clinical studies of patients who have had
LactoSorb plates implanted and have been followed for greater than 9 months
show no evidence of inflammatory reactions [49, 57]. It is believed that the
intermediate rate of degradation for this predominately PLLA copolymer
allows for efficient removal of its byproducts without overwhelming the bodys
mechanisms of clearance [50, 58].
Because of its relatively high profile, LactoSorb can produce a palpable
and sometimes visible implant. Studies have noted this factor to occur occa
sionally; however, the authors state that in all cases the external evidence of the
dramroo
Fig. 4. Explanted LactoSorb plate. Histological examination of explanted LactoSorb
plate demonstates complete resorption of implant material 1 year following surgery (bone on
right, normal soft tissues on left) without evidence of persistent inflammatory reaction.
Table 1. Strength of MacroPore

implant during degradation Time after implantation %
Implantation 100
6 months 90
9 months 70
12 months 50
18 months 0
implant disappears within 36 months [33, 51, 59]. Most studies of LactoSorb
plates show the entire implant to be 95% resorbed by 9 months and completely
resorbed by 1 year (fig. 4) [60].
MacroPore is characterized by a degradation time of 1836 months. The loss
of strength during degradation has been well characterized, and is predictable [61]
(table1). As noted above, however, the strength and degradation characteristics
of any specific MacroPore implant will be influenced by the manufacturing
processes, implant size and geometry, and characteristics of the implantation
site, among other factors. MacroPore can be formed into a wide variety of shapes,
and it can be stored and sterilized utilizing conventional techniques. In appro
priate shapes, MacroPore can be heated and shaped for conformation to the
actual site of implantation, and it will hold the desired shape once cooled with
out loss of structural integrity [MacroPore: pers. commun.].

dramroo
Fig. 5. Absorbable spacer. This resorbable spacer represents one possible avenue of
exploration for resorbable technology for spine surgery.
MacroPore devices cannot be readily visualized on routine radiographic

studies, although they can be seen distinctly on CT scans prior to significant
degradation [Branch: pers. commun; 62]. PLA implants do not degrade MRI
images, and MRI scanning has been utilized to evaluate the tissue response of
PLA implants [30, 6365]. PLA implants are visible on MRI images as areas
of homogeneous low-signal intensity, which can be distinguished from the high
signal intensity of the adjacent bone [63].
MacroPore sheets have been utilized to reconstruct iliac crest donor site
defects [66]. Iliac crest reconstruction may diminish pain, prevent bowel herni
ation through large defects [67], improve cosmesis, and optimize donor site
regeneration [66]. The benefits of a protected healing space have been recog
nized in promoting optimal bone healing [47, 68]. Specifically, if soft tissue is
prevented from prolapsing into a bone defect, the regrowth of bone may be
better than with graft materials alone. In terms of the iliac crest, once the bone
is harvested, the donor site is backfilled with allograft or bone matrix material
if desired. A MacroPore sheet is then heated to 70C, contoured to the defect
site and allowed to cool. It is then secured with screws or tacks [66]. This is an
example of the potential use of MacroPore as a barrier type of implant.
MacroPore implants are being utilized in pilot studies as load-sharing
implants in spine fusion constructs [62]. The versatile nature of this material
allows it to be formed into cages, dowels and interbody spacer shapes (fig. 5).
This, coupled with the desirable strength and degradation characteristics, lack
of artifact on imaging, low potential for foreign body reaction and the biocom
patibility with the dura and nervous tissue, makes it a very promising material.
dramroo
One such study examines the use of MacroPore devices in instrumented poste
rior lumbar interbody fusion constructs [62]. The devices are placed following
complete discectomy, along with morsellized autograft bone, and maintain the
disc space height during the early phase of bone healing. Preliminary results
show equivalent clinical outcomes to those obtained with the use of allograft
bone spacers. The devices produce no artifact on postoperative imaging, allow
ing better visualization of the maturing bone fusion [62].
Other products are being developed worldwide in order to expand the
applications of bioabsorbable products. Bostman et al. [40] used a biodegrad
able rod for internal fixation of extremity fractures and osteotomies. The com
position of the rods was either polyglycolide or a lactide-glycolide copolymer
(polyglactin). A small percentage of the 516 patients treated required reopera
tion for fixation failure (1.2%). The incidence of bacterial wound infections
was low (1.7%). The most significant complication was a foreign-body reaction
in 7.9% of the patients. The reaction caused a painful, reddened, fluctuant
swelling at the operative site 24 months after surgery. Godard et al. [69] have
investigated a bioresorbable implant for spinal arthrodesis, which is reported to
undergo ossification. Brunon et al. [70] have used a bioresorbable plate devel
oped from PLLA for anterior cervical interbody stabilization in 5 patients. The
development of bioabsorbable plates, screws, and rods for use in the spine is an
area of intense activity at the present time.
Conclusion
Absorbable bone substitutes and fixation devices are being developed for
spinal surgery. Advances in material technology may allow for the avoidance of
autograft harvest morbidity and allograft preparation costs and toxicities.
Resorbable implants may result in decreased morbidity and better radiographic
evaluation of spinal fusion procedures. The pace of new development in this
field is remarkable. As new products are brought to market, thorough and
rigorous testing in well-designed clinical trials will be required to establish their
ultimate worth.
References
1 Delecrin J, Takahashi S, Gouin F, Passuti N: A synthetic porous ceramic as a bone graft substitute
in the surgical management of scoliosis: A prospective, randomized study. Spine 2000;25:
563569.
2 Khan S, Sandhu H, Parvataneni H, Girardi F, Cammisa F: Bone graft substitutes in spine surgery.
Bull Hosp Jt Dis 2000;59:510.

dramroo
3 Bostman M, Yang X, Kennan M, Sandhu H, Dicarlo E, Lane J: An unexpected outcome during
testing of commercially available demineralized bone graft materials. Spine 2001;36:14251428.
4 Wang J, Kanim L, Nagakawa I, Yamane B, Vinters H, Dawson E: Dose-dependent toxicity of a
commercially available demineralized bone matrix material. Spine 2001;26:14291436.
5 Boden S: Bone repair and enhancement clinical trial design. Clin Orthop 1998;355S:S336S346.
6 Shors EC: Coralline bone graft substitutes. Orthop Clin North Am 1999;30:599613.
7 Kelly CM, Wilkins RM, Gitelis S, Hartjen C, Watson JT, Kim PT: The use of a surgical grade cal
cium sulfate as a bone graft substitute: Results of a multicenter trial. Clin Orthop 2001;382:4250.
8 Dreesman H: ber Knochenplombierung. Beitr Klin Chir 1892;9:804810.
9 Peltier L, Jones R: Treatment of unicameral bone cysts by curettage and packing with plaster of
Paris pellets. J Bone Joint Surg Am 1978;60:820822.
10 Coetzee A: Regeneration of bone in the presence of calcium sulfate. Arch Otolaryngol 1980;106:
405409.
11 Hadjipavlou A, Simmons J, Yang J, Nicodemus C, Esch O, Simmons D: Plaster of Paris as an
osteoconductive material for interbody vertebral fusion in mature sheep. Spine 2000;25:1015.
12 White E, Shors E: Biomaterial aspects of interpore-200 porous hydroxyapatite. Dent Clin North
Am 1986;30:4967.
13 Yukna RA, Yukna CN: A 5-year follow-up of 16 patients treated with coralline calcium carbonate
(BIOCORAL) bone replacement grafts in infrabony defects. J Clin Periodontol 1998;25:
10361040.
14 Zdeblick T, Cooke M, Kunz D, Wilson D, McCabe R: Anterior cervical discectomy and fusion
using a porous hydroxyapatite bone graft substitute. Spine 1994;19:23482357.
15 Boden S, Martin G, Morone M, Ugbo J, Titus L, Hutton W: The use of coralline hydroxyapatite
with bone marrow, autogenous bone graft, or osteoinductive bone protein extract for posterolat
eral lumbar spine fusion. Spine 1999;24:320327.
16 Bozic K, Glazer P, Zurakowski D, Simon B, Lipson S, Hayes W: In vivo evaluation of coralline
hydroxyapatite and direct current electrical stimulation in lumbar spinal fusion. Spine 1999;24:
21272133.
17 Baramki H, Steffen T, Lander P, Chang M, Marchesi D: The efficacy of interconnected porous
hydroxyapatite in achieving posterolateral fusion in sheep. Spine 2000;25:10531060.
18 Thalgott J, Fritts K, Giuffre J, Timlin M: Anterior interbody fusion of the cervical spine with
coralline hydroxyapatite. Spine 1999;24:12951300.
19 Iseda T, Nakano S, Suzuki Y, et al: Radiographic and scintigraphic courses of union in cervical
interbody fusion: Hydroxyapatite grafts versus iliac crest bone autografts. J Nucl Med 2000;41:
16421645.
20 Emery S, Fuller D, Stevenson S: Ceramic anterior spinal fusion: Biologic and biomechanical
comparison in a canine model. Spine 1996;23:27132719.
21 Toth J, An H, Lim T, et al: Evaluation of porous biphasic calcium phosphate ceramics for anterior
cervical interbody fusion in a caprine model. Spine 1995;20:22032210.
22 Gunzburg R, Szpalski M: A highly porous beta tri-calcium phosphate/bone marrow graft
promotes osteoinductivity in rat subcutaneous model. EuroSpine, Gothenburg, 2001.
23 Madawi A, Powell M, Crockard H: Biocompatible osteoinductive polymer versus iliac graft:
A prospective comparative study for the evaluation of fusion pattern after anterior cervical
discectomy. Spine 1996;21:21232129.
24 Schulte M, Schultheiss M, Hartig E, et al: Vertebral body replacement with a bioglass-polyurethane
composite in spine metastases Clinical, radiological, and biomechanical results. Eur Spine J
2000;9:437444.
25 Roessler M, Wilke A, Griss P, Kienapfel H: Missing osteoconductive effect of a resorbable
PEO/PBT copolymer in human bone defects: A clinically relevant pilot study with contrary results
to previous animal studies. J Biomed Mater Res 2000;53:167173.
26 Delecrin J, Aguado E, Nguyen J, Pyre D, Royer J, Passuti N: Influence of local microenvironment
on incorporation of ceramic for lumbar fusion: Comparison of laminar and intertransverse sites in
a canine model. Spine 1997;22:16831689.
27 Bostman O, Pihlajamaki H: Clinical biocompatibility of biodegradable orthopaedic implants for
internal fixation: A review. Biomaterials 2000;21:26152621.
dramroo
28 Gogolewski S: Bioresorbable polymers in trauma and bone surgery. Injury 2000;4:2832.
29 An YH, Woolf SK, Friedman RJ: Pre-clinical in vivo evaluation of orthopaedic bioabsorbable
devices. Biomaterials 2000;21:26352652.
30 Viljanen J, Kinnunen J, Bondestam S, Majola A, Rokkanen P, Tormala P: Bone changes after
experimental osteotomies fixed with absorbable self-reinforced poly-L-lactide screws or metallic
screws studied by plain radiographs, quantitative computed tomography, and magnetic resonance
imaging. Biomaterials 1995;16:13531358.
31 Cutright D, Hunsuck E: The repair of fractures of the orbital floor using biodegradable polylactic
acid. Oral Surg 1972;33:28.
32 Getter L, Cutright D, Bhaskar S, Augsburg J: A biodegradable intraosseus appliance in the treat
ment of mandibular fractures. J Oral Surg 1972;30:344.
33 Montag ME, Morales L Jr, Daane S: Bioabsorbables: Their use in pediatric craniofacial surgery.
J Craniofac Surg 1997;8:100102.
34 Bergsma E, Rozema F, Bos R, de Bruijn W: Foreign body reactions to resorbable poly(L-lactide)
bone plates and screws used for the fixation of unstable zygomatic fractures. J Oral Maxillofac
Surg 1993;51:666670.
35 Bergsma J, de Bruijn W, Rozema F, Bos R, Boering G: Late degradation tissue response to
poly(L-lactide) bone plates and screws. Biomaterials 1995;16:2531.
36 Barrows T: Degradable implant materials: A review of synthetic absorbable polymers and their
applications. Clin Mater 1986;1:233.
37 Pietrzak W, Sarver D, Verstynen M: Bioabsorbable polymer science for the practicing surgeon.
J Craniofac Surg 1997;8:8791.
38 Rokkanen PU, Bostman O, Hirvensalo E, et al: Bioabsorbable fixation in orthopaedic surgery and
traumatology. Biomaterials 2000;21:26072613.
39 Vainionpaa S: Surgical applications of biodegradable polymers in human tissue. Prog Polym Sci
1989;14:679716.
40 Bostman O, Hirvensalo E, Makinen J, Rokkanen P: Foreign body reactions to fracture fixation
implants of biodegradable synthetic polymers. J Bone Joint Surg Br 1990;72:592596.
41 Bostman O: Osteolytic changes accompanying degradation of absorbable fracture implants.
J Bone Joint Surg Br 1991;73:679682.
42 Bostman O, Parito E, Hirvensalo E, Rokkanen P: Foreign body reactions to polyglycolide screws:
Observations in 24/26 malleolar fracture cases. Acta Orthop Scand 1992;63:173176.
43 Svensson P, Janarv P, Hirsch G: Internal fixation with biodegradable rods in pediatric fractures:
One year follow-up of fifty patients. J Pediatr Orthop 1994;14:220224.
44 Van der Elst M, Klein C, Bliek-Hogervorst J, Patka P: Bone tissue response to biodegradable poly
mers used for intramedullary fracture fixation: A long-term in vivo study in sheep. Biomaterials
1999;20:121128.
45 Gautier S, Oudega M, Fraguso M, et al: Poly(alpha-hydroxyacids) for application in the spinal
cord: Resorbability and biocompatibility with adult rat Schwann cells and spinal cord. J Biomed
Mater Res 1998;42:642654.
46 Levy F, Hollinger J, Szachowicz E: Effect of a bioresorbable film on regeneration of cranial bone.
Plast Reconstr Surg 1994;93:307311.
47 Lundgren D, Nyman S, Mathisen T, Isaksson S, Klinge B: Guided bone regeneration of cranial
defects using biodegradable barriers: An experimental pilot study in the rabbit. J Craniomaxillofac
Surg 1992;20:257260.
48 Medinaceli L, Khoury R, Merle M: Large amounts of polylactic acid in contact with divided nerve
sheaths have no adverse effects on regeneration. J Reconstr Microsurg 1995;11:4349.
49 Eppley B, Sadove A: Surgical correction of metopic suture synostosis; in Kolk C (ed): Clinics in
Plastic Surgery. Philadelphia, Saunders, 1994, pp 555562.
50 Eppley B, Sadove A: Resorbable coupling fixation in craniosynostosis surgery: Experimental and
clinical applications. J Craniofac Surg 1995;6:477482.
51 Pensler J: Role of resorbable plates and screws in craniofacial surgery. J Craniofac Surg 1997;8:
129134.
52 Kumar A, Staffenberg D, Petronio J, Wood R: Bioabsorbable plates and screws in pediatric cranio
facial surgery: A review of 22 cases. J Craniofac Surg 1997;8:9799.

dramroo
53 Goldstein J, Quereshy F, Cohen A: Early experience with biodegradable fixation for congenital
pediatric craniofacial surgery. J Craniofac Surg 1997;8:110115.
54 Eppley B, Sadove A: Effects of resorbable fixation on craniofacial skeletal growth. J Craniofac
Surg 1994;2:110115.
55 Pietrzak W, Sarver D, Verstynen M: Bioresorbable impants: Practical considerations. Bone
1996;19:109S119S.
56 Pietrzak W, Verstynen M, Sarver D: Bioabsorbable fixation device: Status for the craniomaxillo
facial surgeon. J Craniofac Surg 1997;8:9296.
57 Eppley B, Prevel C, Sadove A, Sarver D: Resorbable bone fixation: Its potential role in cranio
maxillofacial trauma. J Craniomaxillofac Trauma 1996;2:5660.
58 Eppley B, Sadove A: A comparison of resorbable and metallic fixation in healing of calvarial bone
grafts. Plast Reconstr Surg 1995;96:316322.
59 Eppley B, Prevel C: Nonmetallic fixation in traumatic midfacial fractures. J Craniofac Surg
1997;8:103109.
60 Eppley B, Reilly M: Degradation characteristics of PLLA-PGA bone fixation devices. J Craniofac
Surg 1997;8:116120.
61 Claes L, Ignatius A, Rehm K, Scholz C: New bioresorbable pin for the reduction of bony fragments:
Design, mechanical properties, and in vitro degradation. Biomaterials 1996;17:16211626.
62 Subach B, Branch C, Alexander J, Haid R, Rodts G: Posterior lumbar interbody fusion using an
impacted bioabsorbable device. AANS/CNS Joint Section on Disorders of the Spine and
Peripheral Nerves, Orlando, 2002.
63 Pihlajamaki H, Kinnunen J, Bostman O: In vivo monitoring of the degradation process of biore
sorbable polymeric implant using magnetic resonance imaging. Biomaterials 1997;18:13111315.
64 Stroman P, Dorvil J, Marois Y, Poddevin N, Guidoin R: In vivo time course studies of the tissue
responses to resorbable polylactic acid implants by means of MRI. Magn Reson Med 1999;42:
210214.
65 Warden W, Friedman R, Teresi L, Jackson D: Magnetic resonance imaging of bioabsorbable poly-
lactic acid screws during the first two years after anterior cruciate ligament reconstruction.
Arthroscopy 1999;15:474480.
66 Taylor W, Ames C, Cataletto M: MacroPore resorbable sheeting: Early experience in immediate
iliac crest contour reconstruction. Personal communication.
67 Cowley S, Anderson L: Hernias through donor sites for iliac bone grafts. J Bone Joint Surg Am
1983;65:10231025.
68 Holmes R, Lemperle S, Calhoun C: Protected bone regeneration. Unpubl. data.
69 Godard J, Jacquet G, Farhat O, Steimle R: Implant intervertbral au biocopoymre (BOP) pour
arthrodse: tude de 45 cas. Chirugie 1991;117:398404.
70 Brunon J, Duthel R, Fotso M, Tudor C: Anterior osteosynthesis of the cervical spine by phusiline
bioresorbable screws and plates: Initial results apropos of 5 cases. Neurochirurgie 1994;40:
196202.
Daniel K. Resnick, MD MS
Assistant Professor, Department of Neurological Surgery
University of Wisconsin, K4/834 Clinical Science Center
600 Highland Ave, Madison, WI 53792 (USA)
dramroo
Lumbar Interbody Fusion Using Bone

Morphogenetic Protein: Results and
Fusion Assessment
J. Kenneth Burkusa, Kevin T. Foleyb, Regis W. Haid, Jr.c
a
The Hughston Clinic, PC, Columbus, Ga., bSemmes-murphy Clinic, Memphis,
Tenn., and cDepartment of Neurosurgery, Emory University, Atlanta, Ga., USA
No uniform criteria for determining fusion have been used in the reports
of radiographic outcomes of anterior lumbar interbody fusion (ALIF) surgery
[1, 2]. The most commonly reported standard for establishing the presence of a
successful lumbar interbody fusion is evidence of trabecular bone formation
linking the adjacent vertebral bodies. A wide range of fusion rates for the ALIF
procedure has been reported [37]. The variability in fusion rates is, in part,
determined by the surgical technique, the type and quantity of bone graft used,
and the presence of an interbody fusion device. With these variables, the deter
mination of a successful arthrodesis is largely determined by the investigators
definition of fusion.
The use of metallic interbody fusion devices creates new challenges in
establishing fusion criteria. These devices obscure vertebral landmarks used in
the assessment of fusion and create artifacts that degrade some imaging tech
niques [810]. The intradiscal radiographic patterns of fusion after surgery with
interbody fusion cages are determined by the extent of the discectomy, the end
plate preparation, the reaming techniques, the interface between the cage and
the host bone, and the characteristics of the bone graft materials used. Direct
surgical exploration is not a viable option for determining the status of an inter-
body fusion [1113]. Indirect assessment at surgery of an anterior fusion through
the manipulation of the posterior spinous processes is subjective and can iden
tify only gross patterns of instability and cannot reliably identify micromotion
across an interspace.
Other than by histologic biopsy, interbody fusion can only be assessed
objectively by various radiographic imaging techniques. Although the presence
dramroo
of a pseudarthrosis can be determined by a single finding on an isolated radi
ographic study, the presence of a solid fusion cannot be determined by a single
radiographic finding. Failure of fusion is established by the absence of bridging
trabecular bone and the presence of a radiolucent area that extends through the
entire fusion mass. Pseudarthrosis can also be identified by marginal radiolucency
around the implant, progressive subsidence of implants, and angular changes in
the spinal motion segment.
At present, no single study or technique is definitive for establishing the
presence of a fusion after anterior interbody surgery [14]. A successful arthrodesis
within a spinal motion segment can be determined by using radiographic eval
uation to assess a stable spinal alignment on sequential examinations, a reduc
tion in angular and translational changes on dynamic motion studies, an
absence of fibrous tissue reaction at the device-host interface, and the presence
of new bone formation and bone remodeling [15].
The most comprehensive and accurate means of radiographically assessing
fusion of the lumbar spine after ALIF with intradiscal implants are plain radi
ographs, dynamic motion radiographs, and thin-cut computerized tomography
(CT) scans. Technetium bone scans and MR imaging are not effective in assess
ing interbody fusion [9, 16]. Plain radiographs are effective for determining
changes in spinal alignment over time. Dynamic plain radiographs can accu
rately assess changes in implant-host bone interface and instability patterns
within the spinal motion segment. CT imaging can identify new bone formation
and bone remodeling within and around the spinal implants [8, 17]. By using
all three imaging technologies, the physician can accurately determine a suc
cessful and clinically relevant interbody fusion [15].
Lumbar Spine Imaging Studies
Plain Radiographs
The plain radiographic studies that are used most often to assess fusion
include the following: (1) standing or weight-bearing radiographs, (2) supine
radiographs, (3) dynamic flexion-extension radiographs, and (4) dynamic side-
bending radiographs. Standing or weight-bearing films are more valuable than
supine films in assessing sagittal plane balance and alignment; a weight-bearing
radiographic technique stresses the interbody fusion can help identify instabil
ity patterns. Supine anteroposterior radiographs show lucencies around metal
lic implants (fig. 1).
Metallic implants create artifacts that make interpretation of plain radi
ographs alone inaccurate in the assessment of fusion [8, 10]. Bone growth within
the implants cannot be assessed on plain radiographs. The thread patterns of
Burkus/Foley/Haid 56
dramroo
a b
Fig. 1. a Supine anteroposterior radiograph shows lucencies around both implants.

b Thin-cut CT scan sagittal reconstruction confirms the lucencies surrounding the cages.
interbody fusion cages create varying amounts of artifact at the implant-host bone
interface. Radiographic lucencies in this area can be misinterpreted. The following
findings on plain radiographs help to identify a fusion: (1) incorporation of grafts
to vertebral end plates, (2) bridging trabecular bone across the interspace,
(3) absence of lucencies at the graft-host interface, (4) absence of subsidence, and
(5) absence of graft or implant migration. The absence of subsidence and implant
migration can only be established by a review of serial radiographs.
Dynamic Plain Radiographs

Dynamic motion studies of the lumbar spine are done to identify subtle
changes within the spinal motion segment after ALIF surgery. Intra- and inter-
observer measurement error is always a factor in the assessment of lumbar
intervertebral segmental angulation and translation on serial radiographs. In their
clinical studies, examiners have considered spinal motion segments to be fused
despite measured differences in both angular and sagittal translation [1, 5].
Reports of measurement error range from 1 to 5. Brantigan [18] proposed that
only 1 of motion was consistent with a fusion. Zdeblick [19] proposed 2, and
Ray [7] proposed 3 of motion as support for the presence of a solid fusion.
These studies contrast with the findings of Kuslich et al. [5] who reported that
angular measurements of less than 5 are not accurate. Commonly, the presence
of 35 of motion within an instrumented spinal segment and 5 mm of transla
tion is considered fused. These differences on dynamic studies are accepted
because of measurement errors. Clinicians often find these results difficult to
interpret because motion should not occur within a fused spinal segment.
Amplifying the problem of intra- and interobserver measurement error on
dynamic radiographs is the fact that no standard method for obtaining flexion-
extension radiographs has been established. Biplanar studies are among the
Lumbar Interbody Fusion Using Bone Morphogenetic Protein 57

dramroo
most reliable plain radiographic methods. Dynamic radiographs taken with the
patient in the standing position are not reliable because of variable patient com
pliance and the technical difficulty of properly centering the x-ray beam paral
lel to the appropriate interspace. With standing lateral dynamic studies, the
pelvis is not locked and motion may be occurring at the hip joints rather than
within the lumbar spine. Supine lateral dynamic studies obtained in the lateral
decubitus position are also unreliable for the same reasons the pelvis is not
locked or properly supported. Lateral radiographic images are frequently rotated
unless the lumbar spine has been supported. It is technically difficult to main
tain the x-ray beam parallel to the spinal motion segment during these dynamic
studies.
Flexion-extension radiographs should be obtained with the pelvis fixed so
that motion occurs within the lumbar spine. Dynamic studies with the patient
in the seated position restrict pelvic and hip motion and enable the technician
to consistently obtain radiographs centered at the appropriate disc space with
minimal rotational distortion at the interspace. Restricting motion to the lum
bar spine by eliminating hip and pelvic motion and consistently obtaining
radiographs that are parallel to the end plate of the instrumented spinal
segment sufficiently improves the accuracy of dynamic radiographic studies
and demonstrates subtle changes within the spinal motion segment (fig. 1).
Biplanar radiographic techniques have been introduced to reduce measure
ment error [20].
Thin-Cut 1-mm CT Scans

Axial CT has also been used to establish fusion [8, 13, 2123]. This cross-
sectional imaging technique eliminates overlapping and rotational errors pre
sent on plain radiographs and allows direct visualization of the fusion mass.
Thin-cut CT scans with sagittal and coronal reconstructions have a greater abil
ity to detect radiolucent areas within a developing fusion mass. Using thin-cut
CT scans, clinicians have identified and classified complex forms of spinal
pseudarthrosis into distinct morphologic categories [1, 20]. In animal studies,
thin-cut CT scans using 1-mm-thick axial sections through interbody fusion
cages have been used to reliably identify new bone formation within metallic
fusion cages and to identify pseudarthroses after interbody surgery [12, 24, 25].
These findings were correlated with histologic confirmation to establish the
presence or absence of a fusion. In these animal studies, CT scans were used to
show trabecular bone formation patterns within the disc space and identify
bridging bone formation that crossed the interspace. CT scanning was also used
to identify lucencies at the implant-bone interface. In our clinical experience,
thin-cut CT scanning has been the most precise and accurate method of evalu
ating interbody fusion [26, 27].
dramroo
A single CT scan may not be able to distinguish between unincorporated
or necrotic bone grafts from bridging trabecular bone within a metallic fusion
cage. Cunningham et al. [28] analyzed fusion through a histomorphometric
assessment. They could not differentiate between residual autograft and new
bone formation. However, serial CT scans can be used to identify maturation of
corticocancellous grafts within fusion cages and can demonstrate incorporation
of these grafts through the openings in the second-generation fusion cages [26].
CT scans can be used to accurately identify new bone formation within the disc
space but outside of the interbody fusion devices [17, 26, 27].
Interbody Fusion Assessment
The assessment of fusion in a patient who has an interbody fusion device

includes four key elements: (1) spinal alignment, (2) segmental spinal stabiliza
tion, (3) device-host bone interface, and (4) new bone formation and bone remod
eling. Spinal alignment must be maintained over time. Similarly, with an intact
fusion, no significant angular or translational change should occur on dynamic
motion studies. The contact points between the device and the host cortical bone
and cancellous bone must also be assessed. For an intervertebral body fusion to
be considered intact and complete, there should be no radiolucent areas sur
rounding the devices at the interface with the host bone. Identification of new tra
becular bone formation within the disc space and remodeling of the grafts within
and around the interbody devices must also be assessed and is the most important
aspect of the fusion criteria [17, 29], but the sentinel sign [30, 31] of the progres
sive anterior bone formation alone is not helpful in determining fusion.
Spinal Alignment
ALIF using stand-alone implants often improve the frontal and sagittal
plain contours of the lumbar spine. Immediate postoperative improvements in
frontal and sagittal plane contours are not maintained over time in all patients
[3234]. Stand-alone implants are susceptible to subsidence into the vertebral
end plates. Subsidence of the implants, which occurs over the course of several
years after surgery, often leads to segmental spinal instability, loss of lordosis,
angular frontal plain deformities, and sagittal plain translation. It is evidence of
a delayed fusion or frank pseudarthrosis (fig. 2). The ability of an implant to
resist subsidence is, in part, related to its design [35]. Subsidence, loss of disc
space height, and angular deformity are also related to the position of the
implants within the disc space [17, 31].
Interbody fusion can only be determined to be complete if there is no change
in the alignment of the spine at the instrumented fusion site for a minimum of

dramroo
a b
Fig. 2. a Standing lateral radiograph at 6 weeks after surgery. Segmental lordosis at

L4L5 measures 14 and the implant rests on the cortical margin of the adjacent vertebral end
plate. b At 18 months after surgery, the segmental lordosis is reduced to 9 and the implant
has subsided through the vertebral end plate. Anterior radial osteophytes have formed.
a b
Fig. 3. a Standing anteroposterior radiographs at 3 months after surgery show no

significant frontal plane deformity. b Standing anteroposterior radiograph at 1 year after
surgery shows a 7 angular deformity across the instrumented interspace.
6 months. Standing anteroposterior and lateral radiographic views must show no

significant change in segmental lordosis (3), sagittal translation (5 mm), or
frontal plane angulation (3) on sequential radiographs taken at least 6 months
apart. Interbody fusion cannot be considered intact if there are progressive changes
in any frontal or sagittal plane angular or translational measurements (fig. 3).
dramroo
Segmental Spinal Stabilization
Subtle changes in the lumbar contours can be identified on dynamic
lateral radiographs; changes in segmental lordosis and sagittal plane translation
can be identified on these studies. Incorporating known measurement error,
criteria for fusion on dynamic radiographic studies include angular motion
of 3 or less and reduction in sagittal plane or frontal plane translation of
5 mm or less. The documentation of persistent motion across a fused motion
segment has led some clinicians and researchers to conclude that the error in
measuring dynamic plane radiographs often precludes an accurate determi
nation of fusion.
Device-Bone Interface
The host bone reaction to an interbody fusion device helps to ascertain
fusion. Although the composition and shape of the implant influence the
region around the device, the host bone reaction to the implant remains an
important aspect of determining fusion. The presence of sclerosis or cystic
radiolucencies on the margins of the implant within the subchondral bone can
result from bony reabsorption and fibrosis tissue reaction secondary micro
motion in the presence of a pseudarthrosis. Radiolucency surrounding the
implants represents the interposition of fibrous tissue at the host bone-implant
interface. It is commonly agreed that this is also a sign of micromotion and
instability. End plate sclerosis that extends through the subchondral bone and
alters the trabecular pattern of the vertebral body is also consistent with micro
motion and pseudarthrosis. Plain radiographs, dynamic extension radiographs,
and thin-cut CT scans are helpful in assessing the device-bone interface. These
changes can be seen on plain radiographs but are best detailed on thin-cut
CT scans.
Progressive collapse of the interspace and migration of implants are gross
radiographic signs of instability, delayed union, and pseudarthrosis. Plain
radiographs can also identify migration and subsidence of the implants within
the disc space. Dynamic lateral extension radiographs help to identify subtle
patterns of motion at the disc space and can help to identify interface lucen
cies. If a fusion is not present, hyperextension lateral radiographs can increase
the gap between the implant and host bone. This gap appears as increased
radiolucency surrounding the implant or may appear as a gap in the anterior
fusion mass.
Thin-cut CT scans are best at detailing cystic changes within the vertebral
end plates, sclerosis, and interface lucency. Heithoff et al. [10] found CT scans
of little value in assessing fusion for the first-generation BAK cages. It is diffi
cult to assess for fusion in patients with the thick-walled and square-threaded
BAK cage because of the radiographic scatter inherently associated with its

dramroo
Fig. 4. Axial CT scan at the device-bone interface shows lucent lines surrounding the
contact points of the implant and the host bone.
design. However, end plate sclerosis and the presence of cyst formation within
the end plate adjacent to the implants can be readily determined even on these
first-generation implants (fig. 4). The interface between the host bone and tita
nium implants can be more easily assessed with second-generation cages such
as the INTER FIX (Medtronic Sofamor Danek, Memphis, Tenn., USA) and
LT-Cage (Medtronic Sofamor Danek). The thread patterns on these cages are
self-tapping and thinner than on earlier devices. The radiographic scatter and
artifact are reduced. With the second-generation cages, it is possible to assess
the interface between the implant and the host bone for the development of
fibrous lucency.
The assessment of interbody fusion with cortical allografts must include
incorporation of the graft materials in addition to the morcellized autogenous
grafts. Complete fusion of an allograft-autograft montage must include incor
poration of the allograft into both vertebral end plates and trabecular bone for
mation across the interspace [22, 36]. With threaded cylindrical bone dowels
and femoral rings, it is possible to determine incorporation of the allograft to
end plates of the host vertebra. On plain radiographs and CT scans, it is common
to find early trabecular bone formation crossing the interspace. In the presence
of spanning trabecular bone formation around the implant, there is often incor
poration of the allograft to only one vertebral end plate. The lucencies surround
ing the contact points of the allograft to one end plate often resolve over time
[34, 37, 38]. They are commonly present at 1 year after surgery and do not
dramroo
a b
Fig. 5. a Coronal CT scan reconstruction immediately after interbody surgery with

LT-Cage devices and autogenous grafts shows the cages are well seated within the L5S1
disc space. There are autogenous bone grafts within the cages; no grafts were placed lateral
to the cages. b Coronal CT scan reconstruction at 1 year after surgery shows maturation of
grafts with the cages and new bone formation outside of the cages and within the confines
of the disc space in the lateral fusion zones.
resolve until 23 years after surgery. These unilateral lucencies on dynamic

plain radiographs are not associated with poor clinical outcomes, subsidence,
or instability. However, they do represent incomplete incorporation and fusion
of the allograft.
New Bone Formation and Bone Remodeling

New bone formation and bone remodeling in and around interbody fusion
cages can be assessed radiographically. Carbon fiber implants and cortical allo
grafts are readily assessed radiographically. The ability to assess bone forma
tion around titanium implants depends, in part, on the size of the implant, the
configuration of the implant, and the porosity of the implant. The first-generation
BAK implant is thick-walled and square-threaded. It also has two small open
ings that are bordered by an internal strut for driving the implant. The configu
ration of this thick-walled titanium implant is not conducive to radiographic
visualization of bone graft within or immediately adjacent to the cage. Second
generation titanium implants produce significantly less scatter and artifact on
plain radiographs and CT scans. The LT-Cage, INTER FIX and Ray TFC
(Surgical Dynamics, Norwalk, Conn., USA) fusion cages [10] are hollow, fen
estrated cylinders with no internal driving device. These cages are significantly
more porous and their thread patterns are also not square and do not produce as
much scatter (fig. 5).
The appearance of bone within an interbody fusion cage is not always
indicative of a fusion after surgery using autograft. CT scanning cannot be used
to distinguish between unincorporated necrotic bone present in the cage and
new trabecular bone formation. Identification of new bone formation outside

dramroo
a b
Fig. 6. a Coronal CT scan reconstruction shows two LT-Cage devices centrally placed
in the L5S1 disc space 48 h after surgery. The cages were filled with InFUSE bone graft
substitute (rhBMP-2 and an absorbable collagen sponge). No autogenous grafts were placed
within the disc space or cage. b At 1 year after surgery, there is new formation within both
cages. There is also new bone formation outside the cages in the lateral fusion zones. All new
bone formation remains confined to the disc space.
the cages in an area where no bone graft was placed is the most important
radiographic sign indicating fusion. Similarly, identification of annular ossifi
cation and bridging osteophytes crossing a disc space are secondary signs of
new bone formation after a successful interbody fusion.
Recombinant Human Bone Morphogenetic Protein-2

Assessing new bone formation after interbody surgery with InFUSE
Bone Graft (Medtronic Sofamor Danek) substitute does not have the inherent
limitations of differentiating between de novo new bone formation and residual
necrotic bone. InFUSE is recombinant human bone morphogenetic protein
(rhBMP-2) applied to an absorbable collagen sponge. Its use replaces the need
for autogenous bone grafts and eliminates the complications associated with
iliac crest graft harvesting. In animal experimental models and in all human
studies, InFUSE has been used without any autogenous or autologous grafts
[2427, 39].
In a prospective human study, osteoinduction after lumbar interbody
surgery was shown to occur with the use of InFUSE [25, 26]. These early find
ings were confirmed in a larger study involving 143 treated with InFUSE [27].
New bone formation occurred in all patients treated with the LT-Cage and
InFUSE (rhBMP-2). The overall fusion rate at 24 months was greater than
94% in these patients. In a smaller sample of these rhBMP-2-treated patients,
bone formation, as evidenced by progressive density on thin-cut CT scans,
almost doubled within 6 months of surgery and increased almost 2.5 times by
24 months [26] (fig. 6). Similar to the bone formation within the cages, new
dramroo
Posterior
zone
Anterior
zone
Fig. 7. Anterior and posterior fusion

zones.
bone formation had occurred outside of the cages in the rhBMP-2 group by 6
months after surgery, and it occurred in all patients by 24 months [26, 27].
New bone formation within the cages occurred most markedly during the first
612 months after surgery; rates of new bone formation exceeded those of the
autograft control group. All new bone formation outside of the cages occurred
within the confines of the disc space. All CT scan slices and all reconstructed
images were studied to evaluate new bone formation. No new bone formation
extended outside of the annulus fibrosus; no bone growth was observed
extending posteriorly into the spinal canal or posterolaterally into the neuro
foramina.
Fusion Zones
New bone formation in a region or zone of the interspace that is free of
autogenous graft or growth factors represents osteoinduction within the soft
tissue elements of the spinal motion segment. Identifying osteoinduction
within the disc space is the most accurate means of determining fusion after
an ALIF procedure. New bone formation only occurs in a spinal motion
segment that is adequately stabilized and, therefore, represents a fused motion
segment. Five fusion zones have been established for interbody fusion
devices. The zones of fusion can be assessed on plain radiographs and CT
scans.
The anterior zone is an area of bone formation in front of the cages along
the anterior margins of the disc space (fig. 7). Bone formation in this zone is
the least reliable indication of interbody fusion. The formation of radial
osteophytes, which is indicative of instability, often masquerades as an early
sentinel sign. On the basis of anterior bone formation alone, it is impossible
to tell if it is a good sign or a bad sign. For a fusion to be present, trabecular bone

dramroo
Fig. 8. Lateral plain radiograph shows
bone anterior to the cage within the L4L5
disc space. There has been significant subsi
dence of the device through the L5 end
plate. This sentinel sign does not represent a
fusion.
formation in the anterior zone must be complete from end plate to end plate.
Bone formation that extends past the confines of the disc space can be an early
indication of a developing pseudarthrosis (fig. 8). Frequently the sentinel sign
represents radial bone spur formation, not interbody fusion. The isolated
sentinel sign may indicate progressive instability instead of progressive
fusion.
The posterior zone is the posterior margin of the interspace (see fig. 7).
Trabecular bone formation in this zone is most likely the best radiographic
indication of interbody fusion. Bone formation in the posterior zone is the most
reliable indication of fusion.
The lateral zone or the lateral margins of the disc space are divided into left
and right regions (fig. 9). Bone formation between the lateral borders of the
implants and the annulus is difficult to visualize on plain radiographs. On
anteroposterior and Ferguson views at the L45 and L5S1 interspace, the
posterior facet joints overlie the lateral fusion zones and early bone formation
in the lateral zone on these plain radiographs. CT scans are essential in visualiz
ing early trabecular bone formation in the lateral zones (fig. 10). Only the final
stages of ossification of the annulus fibrosus are apparent on plain anteropos
terior radiographs. Bone formation in these zones is also a very good predictor
of fusion and typically occurs here before it does in the posterior zone. Bone
formation with the two lateral zones is often asymmetric; this may be related to
asymmetric cage placement.
The between zone is the area of bone formation between the implants
(fig. 9, 11). Bone formation in this zone is best visualized with thin-cut CT
scans and in those patients where the implants have been adequately spaced
away from each other.
dramroo
Bone graft
Bone graft
in lateral
in lateral
zone
zone Posterior
zone
Lateral Lateral
zone zone
Bone graft
between cages Anterior
zone
9 10
Fig. 9. Anteroposterior view of the lateral zones and the between zone. The interbody
fusion cages are placed equidistant from the midline of the disc space. However, there is space
between the cage walls and the annulus fibrosus that compromises new bone formation in the
lateral zone.
Fig. 10. Drawing of an axial CT scan shows the position of the anterior, posterior, and
lateral zones.
Fig. 11. Drawing of an axial CT scan

highlights the within (W) and between (B)
zones.
The within zone is the area of bone formation within the interbody fusion
device (fig. 9, 11). It is very difficult to differentiate between living and dead
bone within the cages. The size, configuration, and material of the cages also sig
nificantly influence the ability to accurately assess bone formation in this zone.
Assessment of bone formation is not practical with a single CT scan. It is best
assessed over time with serial CT scans.

dramroo
Conclusion
Radiographic criteria have been established to reliably assess fusion after

ALIF with threaded and impacted implants and for titanium, carbon fiber, and
allograft devices. Determination of fusion involves the radiographic evaluation
of spinal alignment, stabilization of the spinal motion segment dynamic studies,
assessment of the device-host interface, and identification of new bone forma
tion and bone remodeling. Each of the fusion criteria must be met to ensure that
an arthrodesis is complete.
Changes in the sagittal or frontal plane contours over time indicate a
delayed fusion. Progressive subsidence or any change in sagittal or frontal plane
contours also represents a failed fusion. A fusion can be considered intact if
there is no change in spinal alignment within the spinal motion segment over a
6-month period.
Dynamic radiographic studies must be obtained in a manner that applies
stress to the instrumented spinal motion segment and in a manner that can be
successfully replicated on serial radiographs. The pelvis must be stabilized
and radiographs must be taken parallel to the vertebral end plates. A fusion
is considered intact only if there is no significant motion on dynamic
studies.
Changes in or the appearance of lucencies at the implant-end plate inter
face are indicative of a failed fusion. Second-generation cages permit a closer
evaluation of the device-bone interface. The development of cystic or sclerotic
changes within the subchondral bone of the vertebral end plates is suggestive
of a fusion failure. If a pseudarthrosis is present, hyperextension lateral radi
ographs frequently highlight radiolucencies at the device-bone interface and
can identify a gap within an anterior sentinel fusion.
The formation of new bone adjacent to or within the intradiscal implants
is the most reliable finding for establishing fusion. New bone formation occurs
outside the intradiscal implants when fusion has occurred. New bone formation
within the lateral and posterior zone is the most reliable radiographic indication
of a fusion. Remodeling of autogenous grafts or allografts is also consistent
with an intact fusion.
References
1 Frymoyer JW, Hanley EN Jr, Howe J, Kuhlmann D, Matteri RE: A comparison of radiographic
findings in fusion and nonfusion patients ten or more years following lumbar disc surgery. Spine
1979;4:435440.
2 Weiner BK, Fraser RD: Spine update: Lumbar interbody cages. Spine 1998;23:634640.
3 Crock HV: Anterior lumbar interbody fusion: Indications for its use and notes on surgical tech
nique. Clin Orthop 1982;165:157163.
dramroo
4 Harmon PH: Anterior excision and vertebral body fusion operation for intervertebral disc
syndromes of the lower lumbar spine. Clin Orthop 1963;26:107127.
5 Kuslich SD, Ulstrom CL, Griffith SL, Ahern JW, Dowdle JD: The Bagby and Kuslich method of
lumbar interbody fusion. History, techniques, and 2-year follow-up results of a United States
prospective, multicenter trial. Spine 1998;23:12671278.
6 Lane JD, Moore ES: Transperitoneal approach to the intervertebral disc in the lumbar area. Ann
Surg 1948;127:537551.
7 Ray CD: Threaded titanium cages for lumbar interbody fusions. Spine 1997;22:667679.
8 Cizek GR, Boyd LM: Imaging pitfalls of interbody spinal implants. Spine 2000;25:26332636.
9 Djukic S, Lang P, Hoaglund F, Genant HK: The postoperative spine. Magnetic resonance imaging.
Orthop Clin North Am 1990;21:603624.
10 Heithoff KB, Mullin JW, Holte D, Renfrew DL, Gilbert TJ: Failure of radiographic detection of
pseudarthrosis in patients with titanium lumbar interbody fusion cages. Proceedings of the 14th
Annual Meeting North American Spine Society, Chicago, 1999, pp 1417.
11 Kant AP, Daum WJ, Dean SM, Uchida T: Evaluation of lumbar spine fusion. Plain radiographs
versus direct surgical exploration and observation. Spine 1995;20:23132317.
12 Sandhu HS, Toth JM, Diwan AD, Seim HB 3rd, Kanim LE, Kabo JM, Turner AS: Histologic eval
uation of efficacy of rhBMP-2 compared with autograft bone in sheep spinal anterior interbody
fusion. Spine 2002;27:567575.
13 Steinmann JC, Herkowitz HN: Pseudarthrosis of the spine. Clin Orthop 1992;284:8090.
14 Watkins R: Anterior lumbar interbody fusion surgical complications. Clin Orthop 1992;284:4753.
15 Burkus JK, Foley K, Haid R, LeHuec JC: Radiographic assessment of interbody fusion devices:
Fusion criteria for anterior lumbar interbody surgery. Surgical Interbody Research Group. Neurosurg
Focus 2001;10/4:19.
16 Hannon KM, Wetta WJ: Failure of technetium bone scanning to detect pseudarthroses in spinal
fusion for scoliosis. Clin Orthop 1977;123:4244.
17 Gilbert TJ, Heithoff KB, Mullin WJ: Radiographic assessment of cage-assisted interbody fusions
in the lumbar spine. Semin Spine Surg 2001;13/4:311315.
18 Brantigan JW: Carbon fiber interbody fusion cages: Indications, results and radiographic inter
pretation of fusion. Spine State Art Rev 1997;11:287306.
19 Zdeblick TA: A prospective randomized study of lumbar fusion. Preliminary results. Spine 1993;
18:983991.
20 Pearcy M, Burrough S: Assessment of bony union after interbody fusion of the lumbar spine using
a biplanar radiographic technique. J Bone Joint Surg Br 1982;64:228232.
21 Heggeness MH, Esses SI: Classification of pseudarthroses of the lumbar spine. Spine 1991;
16(suppl):S449S454.
22 Rothman SL, Glenn WV Jr: CT evaluation of interbody fusion. Clin Orthop 1985;193:4756.
23 Siambanes D, Mather S: Comparison of plain radiographs and CT scans in instrumented posterior
lumbar interbody fusion. Orthopedics 1998;21:165167.
24 Boden SD, Martin GJ Jr, Horton WC, Truss TL, Sandhu HS: Laproscopic anterior spinal arthrode
sis with rhBMP-2 in titanium interbody threaded cage. J Spinal Disord 1998;11:95101.
25 Boden SD, Zdeblick TA, Sandhu HS, Heim SE: The use of rhBMP-2 in interbody fusion cages.
Definitive evidence of osteoinduction in humans: A preliminary report. Spine 2000;25:376381.
26 Burkus JK, Dorchak JD, Sanders DL: Prospective randomized study of radiographic assessment
of interbody fusion using rhBMP-2. Proceedings of the 16th Annual Meeting North American
Spine Society, Seattle, 2001, p 98.
27 Gornet M, Burkus JK, Dickman C, Zdeblick TA: RhBMP-2 with tapered cages: A prospective,
randomized lumbar fusion study. Proceedings of the 16th Annual Meeting North American Spine
Society, Seattle, 2001, pp 2122.
28 Cunningham BW, Kanayama M, Parker LM, Weis JC, Sefter JC, Fedder IL, McAfee PC: Osteogenic
protein versus autologous interbody arthrodesis in the sheep thoracic spine: A comparative endo
scopic study using the Bagby and Kuslich interbody fusion device. Spine 1999;24:509518.
29 McAfee PC: Interbody fusion cages in reconstructive operations on the spine. J Bone Joint Surg
Am 1999;81:859880.

dramroo
30 McAfee PC, Boden SD, Brantigan JW, Fraser RD, Kuslich SD, Oxland TR, Panjabi MM, Ray CD,
Zdeblick TA: Symposium: A critical discrepancy: A criteria of successful arthrodesis following
interbody fusions. Spine 2001;26:320334.
31 McAfee PC, Cunningham, Lee GA, Orbegoso CM, Haggerty CJ, Fedder IL, Griffith SL: Revision
strategies for salvaging or improving failed cylindrical cages. Spine 1999;24:21472153.
32 Dennis S, Watkins R, Landaker S, Dillin W, Springer D: Comparison of disc space heights after
anterior lumbar interbody fusion. Spine 1989;14:876878.
33 Dorchak JD, Burkus JK, Foor BD, Sanders DL: Dual paired proximity and combined BAK/
proximity interbody fusion cages: Radiographic results. Proceedings of the 15th Annual Meeting
North American Spine Society, New Orleans, 2000, pp 8385.
34 Kumar A, Kozak JA, Doherty BJ, Dickson JH: Interspace distraction and graft subsidence after
anterior lumbar fusion with femoral strut allograft. Spine 1993;18:23932400.
35 Burkus JK, Dorchak JD, Estes BT, et al: Subsidence evaluation of reduced lateral profile threaded
constructs. 7th International Meeting on Advanced Spine Techniques, Barcelona, 2000.
36 Wimmer C, Krismer M, Gluch H, Ogon M, Stockl B: Autogenic versus allogenic bone grafts in
anterior lumbar interbody fusion. Clin Orthop 1999;360:122126.
37 Flynn JC, Hoque MA: Anterior fusion of the lumbar spine. End-result study with long-term follow-
up. J Bone Joint Surg Am 1979;61:11431150.
38 Sacks S: Anterior interbody fusion of the lumbar spine. Indications and results in 200 cases. Clin
Orthop 1966;44:163170.
39 Zdeblick TA Ghanayem AJ, Rapoff AJ, Swain C, Bassett T, Cooke ME, Markel M: Cervical inter-
body fusion cages: An animal model with and without bone morphogenetic protein. Spine 1998;
23:758765.
J. K. Burkus, MD
The Hughston Clinic
PC, 6262 Veterans Parkway, Columbus, GA 31909 (USA)
dramroo
Neuronavigational Advances
Overview of Spinal Navigation

Mark R. McLaughlin, Juan Bartolomei
Neuro-Group, PA, Lawrenceville, N.J., USA
An important advance in spinal surgery has been the development and

application of image-guided techniques for spinal navigation and fixation.
Image-guided technology includes both spinal stereotaxis as well as fluoroscopy-
based image guidance systems. Both of these techniques offer significant advan
tages over commonly used plain radiography and fluoroscopy for complex spine
procedures. Image-guided spine surgery has been utilized for cervical, thoracic,
and lumbar fixation [1, 35, 7, 911, 13, 1517, 1922, 24, 28, 31, 34, 35]. This
chapter will highlight the newest techniques in image-guided spine surgery and
discuss their advantages and nuances compared to standard open techniques.
Unreliability of Intraoperative Radiography
Traditionally, intraoperative image guidance in spine surgery was directed

primarily by plain radiography and fluoroscopy. Although these conventional
imaging techniques offer surgeons better visualization compared with simple
open exposure and recognition of the anatomy, they still have a limited accu
racy [6, 7, 14, 23, 27, 29, 35]. Intraoperative plain films and fluoroscopy lack a
three-dimensional (3-D) perspective of the anatomy. In addition, imaging in the
thoracic region with these two techniques is problematic related to the rib cage
and difficulty in localizing the appropriate level. Berlemann et al. [2] found that
only 41% of 119 thoracolumbar pedicle screws were accurately placed with
plain radiography at the time of surgery. Weinstein et al. [33] found a 21% fail
ure rate in pedicle screw placement in a cadaver study. He found that success
was independent of experience or approach and that 92% of the failures were
cortical perforations within the spinal canal. Odgers et al. [23] reviewed a series
of 72 patients undergoing placement of pedicle screws utilizing plain radiography
dramroo
for image guidance. Out of a total of 238 pedicle screw placements, 24 pene
trated the pedicle wall and 2 resulted in neurological injury. Other studies have
also documented a significant incidence of pedicle violation with standard
techniques [6, 18, 29]. Anatomical variations have been described within the
cervical spine that suggest image guidance may improve accuracy of screw
placement [13, 17, 18, 26, 30, 34].
Numerous techniques have been described attempting to increase the accu
racy of pedicle screw placement. These include open laminar techniques, evoked
and spontaneous EMG potentials to record pedicle wall breakthrough, and
intrapedicular or epidural endoscopy [8, 12]. Each of these techniques has
significant limitations. Although intuitively open laminar techniques seem to
the safest method of placing pedicle screws, there are various anatomical anom
alies that sometimes can mislead a surgeon. In addition, optimal screw trajec
tory is difficult to predict based solely on anatomy. Evoked and spontaneous
EMG potentials have been utilized to determine pedicle wall breakthrough;
however, this technique presents information to the surgeon after potential neu
rological injury has occurred. Intrapedicular or epidural endoscopy is problem
atic primarily in the thoracic spine region given the small epidural space and
presence of thoracic spine cord.
Image-Guided Technology
There are multiple advantages of image guidance in complex spine surgery.

This technology can determine preoperatively the feasibility of performing
certain difficult spinal instrumentation procedures and can assist the surgeon in
navigating surgical instruments intraoperatively in real time. This technology
can improve accuracy in the placement of spinal instrumentation minimizing
the risk of neurological and vascular injury. In addition to reducing potential
complications, there is an added advantage of achieving optimal bone purchase
at each instrumented level. Once the surgeon and operating room team are
familiar with the use of image-guided techniques, operating time can be reduced
compared with other conventional technologies. Lastly, intraoperative exposure
of patient and surgeon to ionizing radiation can be reduced or eliminated
depending on the technique used [25].
Spinal Stereotaxis
The steps involved in utilizing spinal stereotaxis include both preopera

tive data acquisition and correlation of this data with the intraoperative field.
First, a preoperative CT or MRI scan is obtained with 1-mm slices at a single
McLaughlin/Bartolomei 72
dramroo
Fig. 1. The Stealthstation (Medtronic Sofamor Danek) includes a high-powered
computer and high-resolution monitor. The display generates triplanar images and a 3-D
model of the patients spine. After registration, the digitized instruments are passed through
the operative field and are displayed on the monitor.
gantry angle. Although the patient is in a prone position during surgery, a

preoperative data set is obtained while the patient is in the supine position.
Although there are slight differences in position, these are not quantifiably sig
nificant with the exception of cases of severe instability. Respiratory variation
during scanning or surgery is not relevant to the navigation accuracy either
[10, 11, 26]. Once the data set has been acquired, a 3-D model is built on the
computer (fig. 1).
The patient is positioned in a standard fashion for surgery, adequate expo
sure is obtained and the reference arc is then firmly attached to a spinous
Overview of Spinal Navigation 73

dramroo
Fig. 2. The reference arc is attached firmly to one of the spinous processes exposed in
the operative field. Theoretically the arc should be attached at each vertebral level being
registered and instrumented. We have found for short segment fixation that only one level
registration yields clinically relevant accuracy.
process (fig. 2). It has been our experience that the spinous process is the most
secure and reliable landmark for anchoring the reference arch. The arc should
be placed at each vertebral level being registered and instrumented. At L5, S1
it is useful to place the reference arc on the sacral spinous process with the
angled arch toward the feet. This allows for free movement of the surgeons
hands and instruments without interference of the arc. For upper lumbar and
thoracic instrumentation, the arc should be placed pointing cephalad.
Once the reference arc has been secured, it must be visible to the optical
camera (fig. 3). The surgeon can then register analogous anatomical points on
the computer and on the patients spine. For registration in the lumbar and tho
racic regions specific landmarks include the middle posterior lateral aspect of
the transverse process tips bilaterally, the inferior and superior aspect of the
spinous process tips, the mamillary tubercles as the base of the transverse
processes (when present) (fig. 4).
These points are easily identifiable both on the computer model as well as
the patients spine. Once the reference arc is placed and registration is complete
the computer will then verify the location of the patients spinal anatomy in
relation to the operating room with submillimetric precision. Accuracy is con
firmed by correlating the patients anatomy with the computer model. Digitized
instruments then can be moved within the operative field and using both
passive and active light-admitting diodes (LED). Frameless navigation can be
performed. Standard lumbar instrumentation equipment is available and digitized
dramroo
Fig. 3. The optical camera must be positioned approximately 6 ft away from the refer
ence arc. It must have unobstructed views of the LEDs on the reference arc and the digitized
instruments.
Fig. 4. Registration is carried out by

locating homologous points of anatomy
between the exposed osseous anatomy and the
model generated on the computer worksta
tion. Typically the spinous processes, trans
verse processes, and pars intra-articularis
can be identified and correlated with the
computer-generated model.
including a probe-all drill guide and tap that can contact the spine. With each
movement, the surgeon can look on the computer screen and visualize the
instruments in relation to the patients spine. Images are available in the axial,
sagittal and coronal planes as well as a 3-D image. This allows the surgeon to
see the anatomy in any of the three necessary planes. As an instrument is passed

dramroo
through the surgical field, its position is projected onto the computer screen
with this triplanar view.
The advantages of frameless stereotaxis were discussed previously.
Instrumentation can be placed more accurately decreasing the risk of vascular
and neurological injury. Operating time can be decreased as a result of
increased speed of hardware placement. The exposure of surgeon and patient to
ionizing radiation is also decreased because there is less need for reliance on
traditional imaging techniques such as fluoroscopy or plain films.
Limitations of frameless spinal stereotaxis include the additional cost of
obtaining a preoperative CT or MRI scan. Also any significant unexpected
movement related to instability at the time of positioning or surgery renders this
type of spinal navigation inaccurate. This is because the anatomical relation
ships will differ from the preoperative data set. If the patient has had prior surgery,
potential anatomical landmarks used for registration may not be available.
Lastly, any disruption of the reference arc will require re-registration. Complicated
multilevel fusions can pose a difficult problem in avoiding any unintended
manipulation of the reference arc.
Virtual Fluoroscopy
The introduction of virtual fluoroscopy has offered surgeons several

advantages over frameless spinal stereotaxis. Virtual fluoroscopy combines the
optical tracking technology of spinal stereotaxis with fluoroscopic images. This
technique does not require any preoperative data set acquisition and allows for
instantaneous registration of the patients anatomy. Virtual instruments are then
superimposed on standard fluoroscopic images and data sets can be updated
instantaneously by simply obtaining a new fluoroscopic image. Imaging studies
utilized as anatomic data are obtained from lateral or AP fluoroscopic
intraoperative views. Only one or two images are required to obtain a complete
data set. The C-arm can then be removed from the field. The FluoroNav system
(Medtronic Sofamor Danek, Memphis, Tenn., USA) allows for four images to
be displayed simultaneously in navigated computer-generated representations
of the digitized instruments (fig. 5). The advantages of utilizing virtual fluo
roscopy include minimizing the surgeons exposure to radiation and eliminating
the need for protective aprons. Also, there is an ergonomic benefit being able to
stand flush against the table when placing instrumentation rather than uncom
fortably leaning against the C-arm. Equipment needed for virtual fluoroscopy
includes a C-arm with LED attachments and fiducial array, a computer work sta
tion compatible with most stereotactic systems, an optical camera, a reference
arc, and digitized instruments (fig. 6).
dramroo
5 6
Fig. 5. The Fluoronav system (Medtronic Sofamor Danek) includes a workstation with
a high-powered computer and a high-resolution monitor, an optical camera, and an attach
ment with LEDs and fiducial display for the C-arm.
Fig. 6. The C-arm is modified slightly in that a fiducial display is attached to the
receiver. This metallic attachment contains LEDs that the optical camera sees to begin the
registration process.
Similar to the use of frameless stereotaxis, the patient is positioned in a

standard fashion and the relevant anatomy is exposed. Again, the reference arc
is attached to the spine, and lateral, AT or oblique fluoroscopic images can be
obtained depending on the surgeons preference. The initial image is demon
strated with an overlaid array of spherical fiducials, which are called the cali
bration target. The anatomy is then instantaneously registered and the computer
combines the known geometry of the calibration target with the fluoroscopic
image. Signals are sent from LEDs on the reference arc, instruments, and
C-arm to the optical camera, which then transmits this data to the computer.
The computer calculates the anatomy of the spine relative to the C-arm. No
point for point registration is required.

dramroo
Fig. 7. Fluoro to CT registration. Through computer deformation algorithms, the software
merges the selected vertebral body with the fluoroscopic images. This minimizes extensive
tissue dissection and it becomes convenient when there are no bony anatomical landmarks
(i.e. postlaminectomy) for registration.
The surgeon can choose any instrument needed for pedicle screw place
ment and a single universal instrument handle is attached. The software system
projects to the computer screen a virtual instrument in relation to fluoroscopic
image of the spine.
Various instruments can be projected onto the screen. As the instruments
move in real time with surgeon manipulation, the changes are replicated instan
taneously on the computer screen. The surgeon is thus aware of the depth and
specific location of the instrument at all times. Distances such as pedicle length
can be calculated and screw angulation for optimal screw trajectory can be
determined with simple movements of the hand.
Because virtual fluoroscopy relies on fluoroscopic images, it is most useful
in navigating bony anatomy. Tumor resections and soft tissue abnormalities are
not suited to virtual fluoroscopy but are best suited for standard frameless
stereotaxis. Patients with 3-D or coronal deformities are also unsuitable candi
dates for virtual fluoroscopy and would be better suited for frameless stereo
tactic techniques as well.
Another spinal navigational system has recently gained popularity with the
advancement in fluoro to CT registration technology in which the two fluoro
scopic images are used to noninvasively merge into the preoperative CT image
data set (Vector Vision, Brainlab, Germany) (fig. 7). A preoperative CT scan
with the standard image-guided protocol is necessary prior to using the fluoro
to CT registration technology. This protocol consists of 1-mm axial cuts incor
porating one or two levels above and below the area of interest or the surgical
field where instrumentation is going to be placed. The images are then trans
ported via ether net or zip drive into the working station where a 3-D image is
created. With the 3-D image operative planning can be performed such as iden
tification of entry and target points for pedicle, transarticular, and lateral mass
screw insertion or biopsies.
dramroo
Fig. 8. Computer screen displaying (clockwise) 3-D image, lateral fluoro image, axial
CT, and Auto Pilot view while inserting a pedicle screw.
Once the images are transformed into a 3-D rendition, surgical exposure
is obtained in a standard fashion and the reference array is clamped to the spin
ous process. A new device is also available that allows the reference array to be
placed anteriorly into the vertebral body if an anterior surgical approach is
desired. This has been described previously with other systems [32]. By adjust
ing the flexible joints, the reference clamp can be positioned without obstruct
ing the surgical field. Once the array is secured AP and lateral fluoroscopic
images over the area to be instrumented are obtained. There are three methods
for registration. Like other systems there is the paired point registration in
which anatomical landmarks are identified and registered on both the spine and
on the CT images, and there is the surface matching algorithm in which ten or
more points are rapidly collected on the surface of the spine and registered
to the CT images. As mentioned above, unique to the Brain Lab system is the
fluoro to CT registration.
Fluoro to CT registration is performed by matching on the computer
screen the vertebral body of interest in a spine model to the AP and lateral films
obtained. Once the vertebral bodies are identified and matched, it takes approx
imately 5 min for the software to register and have a navigational 3-D image.
Once registered, the software allows the surgeon to navigate in 3-D, CT, and
fluoro images or a combination of all by adjusting the computer screen (fig. 8).
When targeting a specific end point such as a lesion biopsy or placement of
a pedicle screw, the Brain Lab provides a software called Auto Pilot view that

dramroo
Fig. 9. Auto Pilot view. Concentric circles form a tunnel which provides immediate
directional feedback for pedicle screw insertion or biopsy guidance.
enables safe navigation and instant feedback. The tool consists of a series of
aligned concentric circles forming a tunnel on a direct path to the target. The first
concentric circle is the entry point and the target point is the smallest black circle
at the end of the tunnel. If during navigation there are any deviations from the
trajectory, the tunnel will curve as the last and first circle become noncongruent.
The software provides immediate feedback providing distance to end point or
any compensatory changes in direction (i.e. angle) necessary to safely reach the
desired target. This tool can be used on difficult cases were the anatomy might
be distorted such as in scoliotic patients or when performing percutaneous
pedicle screw insertion where visible landmarks are not available (fig. 9).
dramroo
The Brain Lab system allows the surgeon to adapt any desired instrument
into the image-guided field. This is performed by clamping a passive marker
array to any instrument (i.e. penfield, pedicle finder, tap, or awl) and inserting
the tip of the instrument into the calibration matrix box. This allows any
instrument to be utilized during the surgery.
Conclusions
Image-guided spine surgery enhances a surgeons ability to navigate

instruments and spinal instrumentation throughout the entire spine. These tech
niques not only minimize morbidity related to instrumentation, but also optimize
screw purchase, decrease operating time, and minimize patient and surgeon
exposure to ionizing radiation. Frameless stereotactic techniques offer certain
advantages in the 3-D navigation of the spine. This technology is particularly
helpful in the placement of C1,2 transarticular screw fixation and in patients
with tumors infiltrating the spinal column. Virtual fluoroscopy, which is a
combination of computer-assisted stereotaxis with C-arm fluoroscopic guid
ance, allows a surgeon to manipulate instruments within a surgical field while
viewing a virtual image of the instrument in real time on a computer screen.
Both of these techniques represent significant advances in spinal stabilization
and will most definitely advance future minimally invasive percutaneous
techniques. One remaining challenge in spinal navigation is improving regis
tration accuracy when the spinal alignment is deformed during surgery. Also,
severe spinal deformities tend to pose significant obstacles when trying to reg
ister the images using fluoroscopy. Innovative advances using fluoro to CT
merge registration technology and the 3-D fluoroscopy units appear to
provide new horizons for the improvement of real time registration and safe
navigation.
References
1 Abdel-Malek K, et al: Bone registration method for robot assisted surgery: Pedicle screw inser
tion. Proc Inst Mech Eng 1997;211/3:221233.
2 Berlemann U, et al: Reliability of pedicle screw assessment utilizing plain radiographs versus CT
reconstruction. Eur Spine J 1997;6:406410.
3 Bolger C, Wigfield C: Image-guided surgery: Applications to the cervical and thoracic spine and
a review of the first 120 procedures. J Neurosurg 2000;92:175180.
4 Brodwater BK, Roberts DW, Nakajima T, Friets EM, Strohbehn JW: Extracranial application of
the frameless stereotactic operating microscope: Experience with lumbar spine. Neurosurgery
1993;32/2:209213.
5 Carl AL, et al: In vitro simulation. Early results of stereotaxy for pedicle screw placement. Spine
1997;22:11601164.

dramroo
6 Castro WH, Halm H, Jerosch J, Malms J, Steinbeck J, Blasius S: Accuracy of pedicle screw place
ment in lumbar vertebrae. Spine 1996;21:13201324.
7 Cinotti G, et al: Pedicle instrumentation in the thoracic spine. A morphometric and cadaveric
study for placement of screws. Spine 1999;24/2:114119.
8 Clements DH, et al: Evoked and spontaneous electromyography to evaluate lumbosacral pedicle
screw placement. Spine 1996;21:600604.
9 Foley KT, Rampersaud YR: Stereotactic image guidance of cervical spine lateral mass fixation.
Tech Neurosurg 1999;5/2:95100.
10 Foley KT, Smith KR, Bucholz RD: Stereotactic applications in spine surgery. Contemporary
Update on Disorders of the Spine, Snowbird, 1994.
11 Foley KT, Smith MM: Image-guided spine surgery. Neursurg Clin N Am 1996;7/2:171186.
12 Frank EH, Chamberland DL: An endoscopic pedicle probe: Preliminary development. Neurol Res
1997;6:657661.
13 Glossop ND, Hu RW, Randle JA: Computer-aided pedicle screw placement using frameless stereo
taxis. Spine 1996;21:20262034.
14 Heller JG, Carlson GD, Abitol JJ, Garfin SR: Anatomic comparison of the Roy-Camille and
Magerl techniques for screw placement in the lower cervical spine. Spine 1991;16: S552S557.
15 Kalfas J, Kormos D, Murphy M, et al: Application of frameless stereotaxy to pedicle screw
fixation of the spine. J Neurosurg 1995;83:641647.
16 Kim KD, et al: Universal calibration of surgical instruments for spinal stereotaxy. Neurosurgery
1999;44:173177.
17 Laine T, et al: Improved accuracy of pedicle screw insertion with computer-assisted surgery.
A prospective clinical trial of 30 patients. Spine 1997;22:12541258.
18 Laine T, Lund T, Ylikoski M, Lohikoski J, Schlenzka D: Accuracy of pedicle screw insertion with
and without computer assistance: A randomized controlled clinical study in 100 consecutive
patients. Eur Spine J 2000;9:235240.
19 Ludwig SC, et al: Transpedicle screw fixation of the cervical spine. Clin Orthop 1999;359:
7788.
20 Merloz P, Tonetti J, Pillet L: Pedicle screw placement using image-guided techniques. Clin Orthop
1998;354:3948.
21 Nolte L, et al: Image-guided computer-assisted spine surgery: A pilot study on pedicle screw
fixation. Stereotact Funct Neurosurg 1996;66/12:108117.
22 Nolte LP, Zamorano LJ, Jiang Z, Wang Q, Langlotz F, Berlemann U: Image-guided insertion of
transpedicular screws: A laboratory set-up. Spine 1995;20:497500.
23 Odgers CJ, et al: Accuracy of pedicle screw placement with the assistance of lateral plain radio
graphy. J Spinal Disord 1996;9:334338.
24 Patel N, et al: Interactive image-guided surgery of the spine: Use of the ISG/Elekta Viewing Wand
to aid intraoperative localization. Br J Neurosurg 1997;11/1:6064.
25 Rampersaud YR, Foley KT, Shen AC: Radiation exposure to the spine surgeon during fluoro
scopically assisted pedicle screw insertion. Spine 2000;25:26372645.
26 Rampersaud YR, Simon DA, Foley KT: Accuracy requirements for image-guided spinal pedicle
27 Robertson PA, et al: Reliability of axial landmarks for pedicle screw placement in the lower
lumbar spine. Spine 1998;23/1:6066.
28 Roessler K, Ungersboeck K, Dietrich W, Aichholzer M: Frameless stereotactic guided neuro
surgery: Clinical experience with an infrared based pointer device navigation system. Acta Neurochir
1997;139:551559.
29 Schulze CJ, Munzinger E, Weber U: Clinical relevance of accuracy of pedicle screw placement:
A computed tomographic-supported analysis. Spine 1998;23:22152220.
30 Schwarzenbach O, Berlemann U, Jost B, et al: Accuracy of computer-assisted pedicle screw place
ment: An in vivo computed tomography analysis. Spine 1997;22:452458.
31 Tessman C: Exploring frameless stereotactic image guided surgery. AORN J 1999;69:498512.
32 Vinas FC, Holdener H, Zamorano L, et al: Use of interactive-intraoperative guidance during
vertebrectomy and anterior spinal fusion with instrumental fixation. Minim Invasive Neurosurg
1998;41:166171.
dramroo
33 Weinstein JN, Spratt KF, Spengler D, et al: Spinal pedicle fixation: reliability and validity of
roentgenogram-based assessment and surgical factors on successful screw placement. Spine
1988;13:10121018.
34 Welch WC, Subach BR, Pollack IF, Jacobs GB: Frameless stereotactic guidance for surgery of
the upper cervical spine. Neurosurgery 1997;40:958963.
35 Xu R, et al: Anatomic considerations of pedicle screw placement in the thoracic spine. Roy-Camille
technique versus open-lamina technique. Spine 1998;23:10651068.
Mark R. McLaughlin, MD
Neuro-Group, PA, 123 Franklin Corner Road
Lawrenceville, NJ 08648 (USA)

dramroo
Virtual Fluoroscopy: Overview
and Future Implications
Y. Raja Rampersaud a, Kevin T. Foley b, c

a
Divisions of Orthopaedic and Neurosurgery, University of Toronto,
Toronto, Canada;
b
Division of Neurosurgery, University of Tennessee and
c
Image-Guided Surgery Center, Memphis, Tenn., USA
The utilization of neuronavigation in spinal surgery continues to grow.

Numerous advances in this field have facilitated more practical clinical appli
cations and have led to greater acceptance of this technology. The successful
combination of neuronavigational technology with a practical intraoperative
imaging modality (i.e. fluoroscopy) has significantly increased the clinical util
ity of computer-assisted surgery. Consequently, the introduction of computer-
assisted fluoroscopy (virtual fluoroscopy) has had a significant impact across
multiple surgical specialties.
The complex anatomy of the spine necessitates reliable preoperative and
intraoperative imaging. In spinal surgery, a significant number of surgical compli
cations are a direct result of poor intraoperative anatomical localization. However,
due to technical limitations and cost considerations, detailed intraoperative imag
ing modalities such as computed tomography (CT) or magnetic resonance imag
ing (MRI) are not readily available. Consequently, the need to improve
intraoperative visualization has led to the development of image-guided naviga
tional systems for spinal surgery. The first such system was three-dimensional
(3D) and relied on preoperative imaging (CT based) [1, 2]. Although conventional
3D image guidance technology has gained significant recognition and acceptance,
it is by no means widely utilized by spine surgeons. This technology is still con
sidered to be early in its evolution and is not without limitations. From a clinical
perspective, the greatest limitation of current 3D systems is the time-consuming
and often frustrating process of registration. Another limitation is the current
inability to update the preoperative imaging to reflect the intraoperative position of
the spine. At present, intersegmental changes in the position of the spine either due
dramroo
to a change in the patients overall position compared to their preoperative position
or following a reduction maneuver cannot be updated without the presence of an
intraoperative CT or MRI. This limitation mandates that each segment be individ
ually registered to provide maximal navigational accuracy at that respective
segment. As previously stated, this can often be a rate-limiting step to the novice
user of a conventional 3D image guidance system. Furthermore, the accuracy pro
vided by these systems is often felt to be unnecessary for the majority of spinal
procedures. These limitations continue to drive the need to develop more user-
friendly and practical computer-assisted techniques, such as virtual fluoroscopy.
C-arm fluoroscopy is an intraoperative imaging technique that is familiar
to all spine surgeons. It is routinely employed for real-time intraoperative local
ization of patient anatomy and surgical instruments in a variety of spinal
procedures. Fluoroscopic localization facilitates improved accuracy and in
many instances reduces surgical exposure in a variety of spinal procedures such
as pedicle screw insertion, interbody cage placement, odontoid screw insertion,
and atlantoaxial transarticular screw fixation. This imaging technology has also
enabled the development of percutaneous spinal procedures, such as vertebro
plasty. Despite the advantages of intraoperative fluoroscopy, there are several
limitations. Without a second fluoroscope, only a single projection can be visu
alized at one time. This limitation makes it necessary to reposition the C-arm
during procedures that require multiple planes of visualization. Frequent
repositioning of the C-arm is tedious, time-consuming, and frustrating. In addi
tion, maintaining ideal sterility is a challenge. Furthermore, the position(s) of
the C-arm is often ergonomically challenging to the surgeon. Finally, the
potential deleterious effect of repeated radiation exposure to the surgeons
hands is always an important consideration [3].
Combining neuronavigational technology with a standard C-arm fluoroscope
has enabled the optimization of this versatile intraoperative imaging modality.
A computer-assisted image-guided fluoroscopy (virtual fluoroscopy) system
provides the user with real-time multiplanar anatomical localization of a tracked
surgical instrument in relation to patient-specific fluoroscopic-based images
[48]. These systems eliminate the need for special preoperative imaging stud
ies, manual image to patient registration (see above), C-arm repositioning, or the
associated ergonomic challenges of using a C-arm fluoroscope. Furthermore,
patient and occupational radiation exposure is significantly reduced, as is the need
for wearing lead protection. Virtual fluoroscopy is currently being used for a wide
variety of spinal applications. It is also being used for assistance in osseous cranial
navigation (e.g. transsphenoidal access to the parasellar region) [9]. In addition,
these systems are being extensively used for a variety of orthopedic trauma proce
dures and more recently total joint replacement. A novel variant of this technology
is also being utilized in interventional radiology.
Virtual Fluoroscopy 85
dramroo
System Overview
A typical virtual fluoroscopy system consists of a surgical navigational

computer system, a commercially available digital C-arm fluoroscope, a calibra
tion target that attaches to the C-arm, and a variety of modified surgical instru
ments that are capable of being tracked by the image guidance system. The
process of virtual fluoroscopy can be divided into four basic steps: (1) acquisi
tion of one or more fluoroscopic images, (2) capturing the position of the
patient and C-arm at the time of image acquisition, (3) mathematically recreat
ing the fluoroscopic image formation process of the C-arm for each image
acquired, and (4) superimposing the relative position of a tracked surgical
instrument onto the virtual fluoroscopic image(s) [6, 8]. In step 1, conventional
fluoroscopic images acquired in any plane are automatically transferred to the
computer for image processing. In step 2, information about the relative posi
tion of the patient and the C-arm at the time of image acquisition (i.e. genera
tion of the fluoroscopic image) is captured by a position-sensing camera array.
The most commonly used tracking device is an electro-optical camera that
detects the position of light-emitting diodes or passive reflectors attached to the
object(s) of interest. Other means of position sensing (e.g. electromagnetic) can
also be used. The patient position is defined by attachment of a dynamic refer
ence array (DRA), which rigidly attaches to the portion of the patients anatomy
that is to be imaged. The C-arm position is defined by position markers or sig
nal emitters built into the calibration target that is affixed to the image intensi
fier. In step 3, utilizing a mathematical algorithm, the computer recreates
the fluoroscopic image formation process of the C-arm (i.e. virtual fluoroscopy).
The algorithm calibrates the acquired fluoroscopic image by taking into
account the positional data acquired in the second step. The calibration process
compensates for factors such as gravity-dependent changes in the C-arm image
center, the effect of external electromagnetic fields generated by electrical
equipment in the operating room, and the effect of changes in the C-arms posi
tion with respect to the earths magnetic field. These factors are unique to every
C-arm position; therefore, it is necessary that every acquired image be inde
pendently calibrated. Essentially the computer creates multiple live virtual flu
oroscopic beams that are maintained in their original spatial orientation relative
to the patient. In step 4, the computer determines the position of one or more
trackable surgical instruments using a position-measuring camera and then super
imposes an image of the instrument(s) in the virtual fluoroscopic display.
Dedicated image-guided instruments such as awls, probes, taps, and screwdrivers
are commercially available. Alternatively, any rigid surgical instrument may be
customized with a tracking device or tracked with the assistance of a universal
tool array. The system can display the position of the surgical instrument(s)
Rampersaud/Foley 86
dramroo
a b
Fig. 1. AP and lateral virtual fluoroscopic views of the lumbar spine. Note the high-
grade spondylolisthesis. a The pedicle probe (dark line) has been positioned at the pedicle
entry point (white arrowhead). Its trajectory (light line) has been virtually extended 15 mm
to the base of the pedicle (black arrow). b Pedicle probe advancement is displayed in both
the AP and lateral images simultaneously.
in any of the previously acquired fluoroscopic images, in multiple planes,

simultaneously. The system also allows the actual projection of a surgical
instrument (in one color) and the simultaneous projection of the linear exten
sion of that instruments proposed trajectory (in a second color) (fig. 1).
Accuracy Validation
Bench top testing of these systems has shown excellent (submillimetric)
correlation of the live fluoroscopic position of a probe and its virtual fluoro
scopic projection. In a simulated operating room setting, Foley et al. [4] inves
tigated the in vitro accuracy of the FluoroNav (Medtronic Surgical Navigation
Systems, Louisville, Colo., USA) system. In this study, the difference in posi
tioning of an implanted pedicle probe (tip and trajectory angle) was measured
comparing live and virtual fluoroscopic images. The mean error in probe tip
localization was 0.97 0.40 mm (99% confidence interval 2.2 mm, maximum
probe tip error 3 mm). The mean trajectory angle difference between the vir
tual and actual probe images was 2.7 0.6 (99% confidence interval 4.6,
maximum trajectory angle difference 5).
Clinical Spinal Applications
As this form of 2D neuronavigation is based on fluoroscopy, the use of a

good fluoroscopic technique is necessary. Specifically, the anatomic region of
interest should be placed in the center of the fluoroscopic image to minimize
dramroo
parallax. Furthermore, a true image should always be obtained if possible (e.g.
in a true anterior-posterior (AP) view, the spinous process should bisect the
pedicle). Prior to clinical use, it must be clearly understood that virtual fluo
roscopy systems typically cannot improve the image quality generated by a
given C-arm and it cannot compensate for a surgeons misinterpretation of
fluoroscopic (2D) osseous anatomic data. Therefore, a sound knowledge of
fluoroscopic spinal anatomy is essential for the successful use of a virtual
fluoroscopy system.
Virtual fluoroscopy is presently being used for the following spinal proce
dures: open pedicle screw placement (C7S1), percutaneous pedicle screw
placement (T10S1), lumbar interbody cage insertion, C1C2 transarticular
screw placement, lateral mass screw placement, odontoid screw fixation and
transoral decompression.
Posterior Spinal Procedures

For posterior spinal procedures, patient setup and surgical exposure are
identical to the surgeons normal technique. The DRA is rigidly affixed to the
spinous process of the vertebra. For a typical degenerative case (without gross
instability), the end vertebrae can be utilized for this purpose (e.g., attach the
DRA to L4 for screw placement at L4, L5, and S1). The fluoroscopic views
preferred by the surgeon are then acquired. These may include lateral views,
AP views, or oblique (owls eye) views down the length of the pedicle. Once
processed (see System Overview), the desired images are simultaneously
displayed and anatomical correlation of the tracked instrument compared to
its virtual position is carried out. Utilizing virtual multiplanar fluoroscopic
information, the starting point and trajectory for a given task are chosen.
Clinical judgment and tactile feedback are used to refine the trajectory as
necessary. These systems can provide the virtual projection of any desired
length and diameter onto the chosen trajectory. The use of tracked awls,
probes, and taps permits continuous visualization of the instruments along
their course (fig. 2).
Percutaneous Pedicle Screws

Virtual fluoroscopy systems do not require direct exposure of the spine for
registration, thus percutaneous navigation may be performed as follows.
Through a small incision, the DRA is rigidly affixed to the desired spinous
process. The appropriate fluoroscopic projections (AP, lateral, and oblique) are
obtained and calibrated. From the skin surface, the trajectory of the tracked
probe is virtually extended through the pedicle to visualize the anticipated
course of the pedicle screw. Preoperative CT and MRI are used to decide the
optimal entry point and trajectory for the pedicle screw. Using a cannulated
Rampersaud/Foley 88
dramroo
Fig. 2. AP and lateral virtual fluoroscopic views of the lumbar spine (same patient as
in fig. 1). The ability to follow the course of the pedicle probe simultaneously in the AP and
lateral planes enables reliable placement of the probe along an intended trajectory.
Advancement of image-guided pedicle probe trough the S1 pedicle to the sacral promontory
(white arrow) is demonstrated.
screw system the pedicle is probed and tapped under virtual fluoroscopic assis
tance (fig. 3) [10].
Anterior-Cervical Spine
The surface anatomy of the anterior cervical spine is typically prohibitive
to anatomical registration using conventional 3D neuronavigation. Furthermore,
the anatomy of the anterior cervical spine and the nature of anterior cervical
spinal procedures typically do not allow the practical or safe attachment of a
DRA directly to the spine. At the craniocervical junction, the anatomical rela
tionship of the occiput-C1-C2 can often be maintained by the use of a Mayfield
apparatus. As a result of the rigid relationship between the Mayfield apparatus
and the occiput, a DRA directly attached to the Mayfield combined with auto
matic registration enables virtual fluoroscopic navigation during procedures at
the craniocervical junction (e.g. odontoid screw insertion or transoral decom
pression). For anterior procedures of the subaxial cervical spine, the DRA can
be directly attached to one or more vertebral bodies using threaded distraction
pins. As noted for posterior procedures, the multiplanar position and trajectory
of tracked surgical tools are provided in real time. Specific to anterior cervical
procedures, information with respect to the anatomical midline and depth is
also utilized.
dramroo
3a
Pedicle probe
Projected trajectory
of pedicle probe Dorsal
Dorsal
Medial Medial
Lateral
Spinal cord
Ventral Ventral
3D AP Lateral
3b
3e
3c
3f
3d 3g
Rampersaud/Foley 90
dramroo
Clinical Outcomes
The authors have significant clinical experience utilizing the FluoroNav

system (Medtronic Surgical Navigation Technologies, Louisville, Colo., USA).
In a current prospective study the clinical accuracy of virtual fluoroscopy for
the placement of thoracic and lumbosacral pedicle screws is being evaluated
(unpubl. data). In this study, postoperative CT of 360 pedicle screws [45 patients;
lumbosacral (L1S1) 279; thoracic T (T3T12) 81] were independently
reviewed by a spinal fellow and neuroradiologist. All screws were placed at
an academic teaching center using the FluoroNav system. Residents and fel
lows placed over 50% of the screws. The relative position of the screw to the
pedicle was assessed and graded as follows: A completely in; B 2 mm;
C 24 mm; D 46 mm. Any borderline position was automatically
downgraded. If an osseous breach occurred, the direction of the breach was fur
ther classified. To date, 49 out of 360 screws (13.6%) breached the pedicle wall
(fig. 4). Overall pedicle breaches were grade B in 11.9%, grade C in 1.4% and
grade D in 0.3% of screws. The majority (87.5%) of breaches were minor
(grade B). Overall medial and lateral breaches occurred equally. Due to
anatomic constraints, breaches were 3 times more likely to occur in the thoracic
spine. Thirty-five percent of breaches were secondary to a pedicle screw that
was larger than the size of the pedicle. There were no clinically significant
screw misplacements and no screws required revision. In this series, the poten
tial for any neurological injury (medial pedicle breach, greater than 2 mm) was
0.6%. This study represents a worse case scenario assessment (CT analysis). The
overall misplacement rate in this ongoing study is less than or comparable to
reported misplacement rates using other techniques [1113].
Fig. 3. a Percutaneous pedicle screw insertion can be accomplished utilizing multipla

nar virtual fluoroscopy and an image-guided pedicle probe such as a bone biopsy needle.
b As depicted, the pedicle can be pictured as a cylinder. c A percutaneously placed pedicle
probe is positioned at the pedicle entry point (the top of the pedicle cylinder on the lateral
view and the lateral edge of the pedicle cylinder on the AP view). Its trajectory is then virtu
ally extended to the bottom of the pedicle cylinder (the junction of the pedicle and the
vertebral body). By maintaining a probe trajectory that is lateral to the medial edge of the
pedicle (arrow AP view) at the bottom of the pedicle, the pedicle can be percutaneously
probed with minimal risk of a medial pedicle breach. When using this technique, it is imper
ative that true AP and lateral views are utilized. d After the pedicle is probed, a guide wire is
introduced and a cannulated pedicle screw can be placed. Postoperative AP (e) and lateral
x-ray ( f ) following a minimally invasive posterior lumbar interbody fusion augmented with
the Sextant percutaneous pedicle screw system (Medtronic Sofamor Danek, Memphis,
Tenn., USA). g Postoperative CT image (see e, f ) demonstrating ideal placement of percuta
neously placed L5 pedicle screws utilizing virtual fluoroscopic assistance.
dramroo
a b
Fig. 4. Postoperative CT images demonstrating pedicle screw placement using virtual

fluoroscopic assistance. a At T6 both screws (5.5 mm) are in excellent position within the
pedicles. b At T4, the left screw is in good position and the right screw demonstrates a minor
breach (2 mm) of the medial pedicle wall (arrowhead). In this case the screw diameter
(5.5 mm) was greater than that of the pedicle.
Using a novel instrumentation technique, Foley et al. [10] placed percuta

neous lumbar pedicle screws utilizing the FluoroNav system. Twelve patients
were successfully treated using this technique. The versatility of the imaging
system allowed registration of unexposed spine elements for the percutaneous
procedure. Screw placement was without complication in all cases.
Although the authors only have clinical experience using the FluoroNav
system, other commercially available virtual fluoroscopy systems seem to have
similar accuracy and likely will demonstrate similar clinical results. In an
in vitro study comparing a virtual fluoroscopy system to a conventional 3D
image-guided technique, Choi et al. [7] demonstrated comparable accuracy.
The overall screw misplacement rate (17.9%) demonstrated in this in vitro
study is comparable to the clinical findings of Rampersaud et al. (see above,
unpubl. data). Based on extensive in vitro data and in vivo experience the
authors have concluded that FluoroNav appears to be a safe adjunct for the
placement of thoracic and lumbosacral pedicle screws.
Disadvantages
Although virtual fluoroscopy has many advantages over conventional

fluoroscopy, its primary limitation is 2D image generation. Also, the quality of
Rampersaud/Foley 92
dramroo
the fluoroscopically generated images and the surgeons ability to meaningfully
interpret these views are unchanged compared to conventional fluoroscopy.
For all current image guidance systems, the navigational accuracy is greatest
at the spinal segment to which the DRA is attached. The rigid relationship between
the DRA and the spinal segment to which it is attached allows for accurate track
ing and compensation for any motion occurring at that specific spinal segment. As
a result of intersegmental motion and other technical reasons, overall navigational
accuracy decreases as one moves further away from the DRA. Consequently, to
maintain maximal navigational accuracy, the DRA should be moved to each spinal
segment of interest with new fluoroscopic images acquired at each segment. This,
however, is clinically impractical for multilevel cases. Unless gross segmental
spinal instability is present, acceptable clinical accuracy is typically maintained
for 25 spinal segments for one DRA position (1 segment: above, 1: at and 13:
below). However, clinical judgment must be used to determine whether the system
is adequately correlating to the segment of interest before navigation is carried out.
For spinal procedures such as C12 transarticular screws where a low margin of
error exists the DRA must be attached to the vertebrae of interest. Furthermore, by
obtaining a live image to confirm the position of a tracked instrument the fluoro
scope can serve as its own internal validation system.
Although virtual fluoroscopy provides the advantages of intraoperative imag
ing, it is still limited by the inability to easily track multiple spinal segments.
Therefore, it is still necessary (for the reasons mentioned above) to move the DRA
and acquire new fluoroscopic images several times during long segment proce
dures. The overall number of C-arm positional changes and images required,
however, is still much less when compared to conventional C-arm fluoroscopy.
Under ideal circumstances surgical navigational systems can be very accu
rate. However, there are numerous potential sources of error that can significantly
affect the clinical accuracy of any image guidance system. These include errors
related to imaging, tracking, registration algorithm(s), and the tracked surgical
tools. These errors, although typically submillimetric, are cumulative and can
result in a clinically significant overall navigational error. One of the potentially
largest, but often understated sources of errors is the surgeon. Like any other
tool, a virtual fluoroscopic navigational system is only as good as its user.
Sound knowledge of the limitations associated with these systems is paramount
for their optimal use.
Future Implications
The future of virtual fluoroscopy is one of versatility. The practical imple

mentation of an easy to use intraoperative imaging modality combined with
dramroo
surgical navigational technology has enabled a robust platform from which only
improvement is possible. Virtual fluoroscopy has the potential to bring practi
cal 3D intraoperative image guidance to the operating room. This is currently
evolving in two pathways. The first is a direct method utilizing a novel isocen
tric C-arm (Siemens). This C-arm is capable of generating a 3D image data set,
therefore allowing 3D neuronavigation without the hassles of 3D registration.
Furthermore, because it is an intraoperative imaging device certain limitations
such as positional (patient) intersegmental motion and intraoperative image
update of conventional 3D neuronavigation are eliminated. The second pathway
involves technology that allows automated registration of intraoperative 2D
fluoroscopic images to a preacquired 3D image data set [14]. The goal of both
techniques is to eliminate the time-consuming step of tactile anatomic registra
tion, while providing unparalleled 3D anatomic information and practical
intraoperative imaging capabilities. 2D to 3D image fusion also allows incor
poration of multiple imaging modalities (e.g. CT, MRI, ultrasound or angio
graphic). The added anatomical detail of a 3D image data set without the need
for tactile registration is ideally suited for the refinement and development of
percutaneous or minimal access spinal surgery.
Virtual fluoroscopy is also facilitating another exciting area of research
that involves practical solutions for tracking multiple segments of the spine.
Combined with segmentation of an image data set this will allow the user to
maintain optimal navigational accuracy over numerous segments and update
the relative position of individual segments in real time. Finally, it is only
logical that this technology will likely be integrated directly into the C-arm
fluoroscope, thus providing a versatile intraoperative imaging-navigational
system.
Conclusion
Virtual fluoroscopy is a safe and beneficial adjunct to complex spinal

surgery. The use of this technology has shown improved or comparable pedicle
screw misplacement rates when compared to published results utilizing
anatomical landmarks with or without fluoroscopy and conventional 3D image
guidance, respectively. As with all neuronavigation devices, advantages and
disadvantages specific to virtual fluoroscopy exist and must be considered in
the decision to utilize this technology. The real-time multiplanar information
provided by these systems can be a powerful adjunct to the surgeons ability to
perform complex tasks during spinal surgery. However, it must never be for
gotten that this technology is intended to serve as an adjunct to the surgeons
clinical judgment and technical skills, not a replacement.
Rampersaud/Foley 94
dramroo
References
1 Foley KT, Smith MM: Image-guided spine surgery. Neurosurg Clin N Am 1996;7/2:171186.
2 Rampersaud YR, Foley KF: Image-guided spinal surgery. Operative Tech Orthopaed 2000;10/1:
6468.
3 Rampersaud YR, Foley KT, Shen AC, Williams S, Solomito M: Radiation exposure to the spine
surgeon during fluoroscopically assisted pedicle screw insertion. Spine 2000;25:26372645.
4 Foley KT, Rampersaud YR, Simon DA: Virtual fluoroscopy: Multi-planar x-ray guidance with
minimal radiation exposure. Eur Spine J 1999;8(suppl 1):S36.

5 Nolte LP, Slomczykowski MA, Berlemann U, Strauss MJ, Hofstetter R, Schlenzka D, Laine T,
Lund T: A new approach to computer-aided spine surgery: Fluoroscopy-based surgical navigation.
Eur Spine J 2000;9(suppl 1):S78S88.
6 Foley KT, Simon DA, Rampersaud YR: Virtual fluoroscopy. Operative Tech Orthopaed 2000;10/1:
7781.
7 Choi WW, Green BA, Levi ADO: Computer-assisted fluoroscopic targeting system for pedicle
screw insertion. Neurosurgery 2000;47:872878.
8 Foley KT, Simon DA, Rampersaud YR: Virtual fluoroscopy: Computer-assisted fluoroscopic nav
igation. Spine 2001;26:347351.
9 Jane JA, Thapar K, Alden TD, Laws ER Jr: Fluoroscopic frameless stereotaxy for transsphenoidal
surgery. Neurosurgery 2001;48:13021308.
10 Foley KT, Gupta SK, Justis JR, Sherman MC: Percutaneous pedicle screw fixation of the lumbar
spine. Neurosurg Focus 2001;10/4:18.
11 Merloz P, Tonetti J, Pittet L, Coulomb M, Lavallee S, Sautot P: Pedicle screw placement using
image guided techniques. Clin Orthop 1998;354:3948.
12 Laine T, Lund T, Ylikoski M, Lohikoski J, Schlenzka D: Accuracy of pedicle screw insertion with
and without computer assistance: A randomised controlled clinical study in 100 consecutive
patients. Eur Spine J 2000;9:235240.
13 Youkilis AS, Quint DJ, McGillicuddy JE, Papadopoulos SM: Stereotactic navigation for place
ment of pedicle screws in the thoracic spine. Neurosurgery 2001;48:771778.
14 Theodore N, Sonntag VKH: Spinal surgery: The past century and the next. Neurosurgery 2000;46:
767776.
Y. Raja Rampersaud MD, FRCSC
Toronto Western Hospital, University Health Network
Edith Cavell Wing, 1-039, 399 Bathurst St., Toronto, ON M5T 2S8 (Canada)
dramroo
Placement of Thoracic Pedicle
Screws
David W. Polly, Jr., Timothy R. Kuklo

Department of Orthopaedic Surgery and Rehabilitation,
Walter Reed Army Medical Center, Washington, D.C., USA
Thoracic pedicle screws are experiencing a significant increase in utiliza

tion in North America, as well as in the rest of the developed world [2, 5, 20,
21, 3034]. Consequently, there is a heightened interest in factors leading to
successful use.
When placing pedicle screws in the thoracic spine, the surgeon primar
ily uses his or her knowledge of general thoracic pedicle anatomy along with
a preoperative plan founded on sound biomechanical principles for the initial
approach. Preoperative patient-specific imaging studies aid the surgeon in
adapting general anatomy knowledge to patient-specific anatomy. These
include a detailed study of the preoperative radiographs for pedicle position
and size, in addition to possible computerized tomography (CT) and magnetic
resonance (MR) imaging, which can further detail the pertinent anatomy. In
concert with a thorough understanding of this anatomy, tactile feedback,
however, remains the primary means of confirmation for successful screw
placement.
More recent screw placement guidance tools include conventional fluo
roscopy (fig. 1), two-dimensional (2-D) fluoroscopy (fig. 2), computer-aided
image guidance (fig. 3), intraoperative advanced imaging (such as CT or MR),
physiological monitoring guidance and emerging three-dimensional (3-D)
The opinions or assertions contained herein are the private views of the authors and are
not to be construed as official or as reflecting the views of the United States Army or the
Department of Defense. The authors are employees of the United States government. This
work was prepared as part of their official duties and as such, there is no copyright to be
transferred.
dramroo
Fluoro guidance
Fig. 1. This is conventional image

guidance with use of intraoperative fluo
roscopy. This image demonstrates the typi
cal 22 sagittal inclination of the thoracic
pedicle when utilizing the anatomic axis.
2-D guidance
Fig. 2. This example of 2-D image guidance demonstrates two planar images allowing
the surgeon to navigate simultaneously in these planes. Advantages of this include familiar
ity of the display (fluoroscopy) and the ability to update the images intraoperatively as
needed.
Placement of Thoracic Pedicle Screws 97

dramroo
3-D image guidance
Fig. 3. This example of 3-D image guidance demonstrates the utility of multiplanar
navigation. It allows for simultaneous sagittal, coronal and axial planar navigation. This
allows for true optimization of pedicle screw fit, fill and trajectory.
fluoroscopic guidance. Screw tract and screw placement confirmation is ulti

mately achieved by palpation, imaging, physiological monitoring or a combina
tion of all of these techniques. Successful screw placement can be defined in a
number of ways [2, 11]. With successful surgical outcome as the primary goal,
successful screw placement may be a necessary, but not sufficient requirement.
Pedicle Anatomy
Applied thoracic pedicle anatomy has been well studied, and normative
tables and graphs have been developed [7, 8, 14, 16, 22, 24, 3538, 40, 41].
Currently, there are primarily two accepted screw trajectories the straight-
ahead trajectory (initially popularized by Roy-Camille, then Suk and Lenke),
and the anatomic trajectory (utilized by Harms and others, the term initially
suggested by Polly) [31].
Advantages of the straight-ahead trajectory include permitting the use of a
fixed head (monoaxial) screw, as well as offering increased insertional torque
Polly/Kuklo 98
dramroo
and pull-out strength when compared to the anatomic trajectory [17, 19]. The
major disadvantage is that this trajectory may require the screw to traverse a
narrower portion of the pedicle isthmus to remain fully contained. The
anatomic trajectory (directed along the sagittal pedicle axis a 22 inclination
from dorsal rostral to ventral caudal) permits the surgeon to navigate a larger
portion of the pedicle isthmus and place a longer screw within the bone. It may
also provide a toenail effect, which typically provides the advantage of not
sustaining direct in-line pull-out forces to the screw. Any potential benefit in
construct performance, as opposed to individual screw performance, is
unknown.
Pedicle Screw Tract Placement
A number of techniques have been used to successfully navigate the pedicle.

Initial techniques utilized tactile feedback along with knowledge of the pedicle
anatomy to place screws. The cancellous feel, familiar to surgeons experi
enced in lumbar pedicle screw placement, can also be used for thoracic screw
placement. Since cortical bone has a distinctly different feel, a progressive
cancellous resistance to the advancing pedicle probe is highly reassuring.
Likewise, a sudden change in this feel is a cause for reassessment.
The neurocentral junction, however, is a predictable point where a distinct
change in the feel occurs. This is because the physeal growth plate is distinctly
more dense than the cancellous bone of the pedicle and the vertebral body.
Breeching this physeal scar has a similar feel to violating a cortical margin and
requires a certain level of experience (or anatomic confirmation) to differenti
ate acceptable from unacceptable pedicle tract navigation.
Surgical Pedicle Navigation Techniques
Proposed techniques for pedicle navigation include the freehand tech

nique as well described by Lenke et al. [20]. Similarly, drilling has been utilized
by many surgeons while the use of a small diameter drill followed by placement
of k-wires with conventional radiographic confirmation has been popularized
by Suk et al. [32]. The funnel technique of breeching the dorsal cortex with a
burr and then navigating the pedicle with a small size curette (such as a 2-0
Cloward curette) has been popularized by Gaines [10], and espoused by others
for cervical pedicle navigation (find the hole, do not make the hole).
Nonetheless, each of these techniques requires tactile feedback as the primary
means of confirmation.

dramroo
This tactile feedback can be supplemented with a number of adjunctive
technologies [1]. Conventional fluoroscopy has been widely used, with a high
level of accuracy [2]. It has the advantage of being a real time evaluation in a
familiar format. Disadvantages include the presence of the fluoroscopic unit at
the operative field, the conventional difficulties associated with planar radiog
raphy (visualizing the upper thoracic spine, penetrating large body mass
patients, surgeon impairment from lead gowns) and exposure to both the patient
and the operative team. A skilled technologist is also required.
2-D image guidance (such as FluoroNav, Medtronic Sofamor Danek,
Memphis, Tenn., USA) acquires intraoperative images through a tracking arc
attached to the patient and displays the images on a computer screen to assist
the surgeon to navigate the pedicle [27, 28] (fig. 2). The major advantage of this
system is that it permits the surgeon to navigate in multiple conventional fluo
roscopic planes simultaneously. This concept, the use of conventional imaging
linked in multiple planes, is a comfortable concept for many surgeons due to
the familiarity of the radiographic images. The disadvantages include the
requirement for a sophisticated unit that requires an additional user familiar
with the technical operation of the system.
3-D image guidance, or frameless stereotactic image guidance, utilizes
preoperative axial imaging studies (most typically CT) with intraoperative
registration to correlate segmental anatomy and display multiplanar images for
navigation [4] (fig. 3). The most significant advantage of this technique is the
display of axial images during the navigation process. This is perhaps the most
critical anatomic information required by the surgeon during pedicle naviga
tion. Disadvantages include the need for the sophisticated unit (and subsequent
cost), obtaining the preoperative axial imaging studies (and additional cost),
transferring the information to the unit and then obtaining appropriate intraop
erative segmental registration of acceptable precision. There is a significant
time investment necessary in the early use of the technology.
Physiological intraoperative guidance is based on stimulating a pedicle
probe during screw tract navigation, thus looking for a neural response from
either a nerve root or the spinal cord. This technique has been extensively
researched in the lumbar spine where the nerve root level myotomes provide
good discrimination. In the thoracic spine, this has proven to be more difficult.
Monitoring from the rectus abdominis appears to give reasonable evoked
potentials for T6T12 [29]. To date, this technique has proven to be less reliable
rostral to T6. Newer technologies may allow real time monitoring of nerve root
proximity, as well as directional information.
Future technological developments may include sophisticated haptic feed
back that can discriminate between bone, nerve and blood vessels. Ideally, this
technology should permit differentiation between cortical and cancellous bone,
Polly/Kuklo 100
dramroo
thus providing an excellent adjunct in pedicle navigation. The merging of
multiple modalities could further enhance the margin of safety provided for
navigating these narrow access corridors.
Evaluation of the Pedicle Tract
Classically, evaluation of the pedicle tract is done by direct bony palpation.

This includes five-wall confirmation, or palpation of the medial, lateral, superior,
inferior and anterior walls. However, the predictive value of this maneuver
varies between surgeons, and is not foolproof [18]. If a breech is detected,
there is a high positive predictive value; however, if no breech is detected, the
negative predictive value is not as good. An experience factor is also present
and appears to play a significant role.
Confirmation of a bony pedicle breach requires further analysis by the
surgeon to determine whether or not a screw can still be placed safely. One
technique, described by Dvorak et al. [6], places a pedicle screw along a
trajectory lateral to the pedicle, but within the confines of the pedicle-rib
complex, and ultimately into the vertebral body. This has subsequently been
termed the in-out-in technique or the pedicle-rib technique. OBrien et al. [23]
have demonstrated the anatomy of this corridor. So simply having a small
lateral cortical breach is of no particular clinical consequence. It also appears
that a small medial breach can be tolerated [2, 11]. Further, Polly et al. [26]
analyzed volumetric spinal canal intrusion from medially positioned screws in
a computer model and found that a screw must have over a 2-mm medial breach
to have the same volumetric spinal canal intrusion as a perfectly positioned
pediatric laminar hook. However, in spinal deformity, a medial breach on the
concave apex would not be tolerated, and any change in neurological monitoring
is always a cause for significant concern [25].
Following insertion, screw purchase is routinely evaluated. If the screw is
loose (poor purchase), then it is not acceptably placed, whereas good purchase
does not necessarily confirm adequate placement. Again, screw removal and
direct palpation may assist the surgeon in identifying a breach. Similarly, elec
trophysiological monitoring can suggest a medial cortical violation, especially
from T6T12, but it is less reliable at detecting lateral or anterior breaches.
Intraoperative imaging (fluoro or plain film) can also be used to assess place
ment, yet interpretation of the images can be challenging. Several checks have
been found to be helpful. First, with multiple level fusions, one can appreciate
the progressive orientation of the screws on the PA view, and a break in the
progression will alert the surgeon to a potentially misplaced screw. Further, the
pedicle can usually be identified, and the screw should appear to be tracking

dramroo
Case example
4a
T5 T6
4b 4c
T7 T8
4d 4e
T9 T10
4f 4g
Polly/Kuklo 102
dramroo
T11 T12
4h 4i
L1 L2
4j 4k
L3
4l
Fig. 4. al This is an example of a right thoracic left lumbar progressive scoliotic

deformity in a skeletally immature individual (curve pattern Lenke 3CN, King curve type
double major). Because of the curve magnitude (50), and her immaturity, she met opera
tive indications. The lumbar modifier (Lenke C) and the thoracolumbar junctional kyphosis
mandated fusion of both curves. The preoperative coronal and sagittal radiographs demon
strate the deformity, the postoperative radiographs demonstrate the excellent correction in
the coronal and sagittal planes while preserving the motion of the intervertebral disks from
L3 to the sacrum. The sequential CT images demonstrate the axial views of the screw
position at each level T5 through L3. The physeal scar (or neurocentral growth plate) is well
visualized on the right at T10 and T11.

dramroo
through the pedicle projection. Multiple views, especially in spinal deformity,
can be of further help. To date, the accuracy of planar radiography to determine
pedicular screw placement in the thoracic spine has not been validated.
Currently, CT scanning is considered the gold standard, although it is more real
istically a worst case analysis since there is some magnification artifact even
with titanium implants [3, 9, 12, 13, 39]. Intraoperative CT scanners have been
utilized at some institutions. Obviously, this requires a significant infrastructure
and operating room modifications. For particularly high-risk screw placement,
one strategy might involve a planned intraoperative transport to a scanner
within the institution. Again, this requires significant coordination and
increased operative time. We have found that the routine use of postoperative
CT scans in patients who have undergone placement of thoracic pedicle screws
has added an extra margin of safety and an invaluable educational opportunity
for the operating surgeons to critically assess their technique (fig. 4al).
Successful Screw Placement
Successful screw placement can be defined in a number of ways. The

most obvious criteria would be the presence or absence of neurological or
visceral compromise as a result of the screw placement [15]. By all defini
tions, presence of these conditions would be considered an unsuccessful
screw placement. The second criterion may be screw purchase. In other
words, does the screw have enough purchase to hold against the applied
loads? However, even if placed with anatomic precision, placement can still
be unsuccessful if there is inadequate bone stock. Third, did the screws suc
cessfully achieve the goal of instrumentation, i.e. was the deformity corrected
or was the spine stabilized sufficiently to achieve the long-term biological
solution of fusion?
Conclusions
Individual decisions and techniques, including the use of neuronaviga

tional aids, are premised to successfully and safely achieve the goals of surgical
treatment. Placement of thoracic pedicle screws is only one component of the
surgical process. However, for the surgery to be a success, the screws must be
placed safely and effectively. This can be achieved through a variety of tech
niques using one of several accepted strategies. The surgeon must have enough
resources available to accomplish the task. Given the variety of experience, skill
and patient-specific anatomy, different resources will be required.
Polly/Kuklo 104
dramroo
References
1 Amiot LP, Labelle H, DeGuise JA, et al: Computer-assisted pedicle screw fixation: A feasibility
study. Spine 1995;20:12081212.
2 Belmont PJ, Klemme WR, Dhawan A, Polly DW: In vivo accuracy of thoracic pedicle screws.
Spine 2001;26:23402346.
3 Berlemann U, Heini P, Mller U, Stoupis C, Schwarzenbach O: Reliability of pedicle screw assess
ment utilizing plain radiographs versus CT reconstruction. Eur Spine J 1997;6:406411.
4 Carl AL, Khanuja HS, Sachs BL, et al: In vitro simulation: Early results of stereotaxy for pedicle
5 Cinotti G, Gumina S, Ripani M, et al: Pedicle instrumentation in the thoracic spine. Spine
1999;24/21:114119.
6 Dvorak M, MacDonald S, Gurr KR, Bailey SI, Haddad RG: An anatomic, radiographic, and
biomechanical assessment of extrapedicular screw fixation in the thoracic spine. Spine
1993;18:16891694.
7 Ebraheim NA, Jabaly G, Xu R, et al: Anatomic relations of the thoracic pedicle to the adjacent
neural structures. Spine 1997;22:15531557.
8 Ebraheim NA, Xu R, Ahmad M, et al: Projection of the thoracic pedicle and its morphometric
analysis. Spine 1997;22/3:233238.
9 Farber GI, Place HM, Mazur RA, Jones CDE, Damiano TR: Accuracy of pedicle screw placement
in lumbar fusions by plain radiographs and computed tomography. Spine 1995;13:14941499.
10 Gaines RW: The use of pedicle screw internal fixation for the operative treatment of spinal disorders.
J Bone Joint Surg Am 2000;82:14581476.
11 Gertzbein SD, Robbins SE: Accuracy of pedicular screw placement in vivo. Spine 1990;15:1114.
12 Kim KD, Johnson JP, Bloch O, Masciopinto JE: Computer-assisted thoracic pedicle screw place
ment: An in vitro feasibility study. Spine 2001;26:360364.
13 Klemme WR, Belmont PJ, Polly DW: Transpedicular thoracic screws: General concepts and inser
tion technique using fluoroscopic guidance. Semin Spine Surg 2002;14/1:4347.
14 Kothe R, Panjabi MM, Liu W: Multidirectional instability of the thoracic spine due to iatrogenic
pedicle injuries during transpedicular fixation. A biomechanical investigation. Spine
1997;22:18361842.
15 Kuklo TR, Lehman RA, Schroeder T: Perils and pitfalls of thoracic pedicle screw placement.
Semin Spine Surg 2002;14/1:97102.
16 Kuklo TR, Polly DW: Anatomy of the thoracic pedicle. Semin Spine Surg 2002;14/1:37.
17 Lehman RA, Kuklo TR: Pedicle screw salvage using anatomic trajectory. Scoliosis Research
Society Annual Meeting, Seattle, 2002.
18 Lehman RA, Kuklo TR, Orchowski JR, Potter BK, Polly DW: Probing for detection of thoracic
pedicle breach: Is it valid? Scoliosis Research Society Annual Meeting, Seattle, 2002.
19 Lehman RA, Polly DW, Kuklo TR, Belmont PJ, Cunningham B, Kirk KL: Advantage of straight
forward versus anatomic trajectory for placement of thoracic pedicle screws: A biomechanical
study. Scoliosis Research Society Annual Meeting, Seattle, 2002.
20 Lenke LG, Rinella A, Kim Y-J: Free hand thoracic pedicle screw placement. Semin Spine Surg
2002;14/1:4857.
21 Liljenqvist UR, Halm HF, Link TM: Pedicle screw instrumentation of the thoracic spine in idio
pathic scoliosis. Spine 1997;22:22392245.
22 McCormack BM, Benzel EC, Adams MS, et al: Anatomy of the thoracic pedicle. Neurosurgery
1995;37:303308.
23 OBrien MF, Lenke LG, Mardjetko S, et al: Pedicle morphology in thoracic adolescent idiopathic
scoliosis: Is pedicle fixation an anatomically viable technique? Spine 2000;25:22852293.
24 Panjabi MM, OHolleran JD, Crisco JJ, et al: Complexity of the thoracic spine pedicle anatomy.
Eur Spine J 1997;6:1924.
25 Polly DW, Belmont PJ, Klemme WR, Robinson M: In vivo accuracy of transpedicular thoracic
screws in patients with and without coronal-plane spinal deformities. Scoliosis Research Society
Annual Meeting, Cleveland, 2001.

dramroo
26 Polly DW, Young JS, Johnson C, Kuklo TR: Comparison of volumetric intrusion of pedicle screws
versus hooks. Scoliosis Research Society 36th Annual Meeting, Cleveland, 2001.
28 Rampersaud YR: Advanced image guidance for thoracic pedicle screws. Semin Spine Surg
2002;14/1:5865.
29 Raynor BA, Lenke LG, Kim YJ, et al: Can triggered EMG thresholds predict safe thoracic pedicle
screw placement? Scoliosis Research Society Annual Meeting, Cleveland, 2001.
30 Roy-Camille R, Saillant G, Mazel C: Internal fixation of the lumbar spine with pedicle screw
plating. Clin Orthop 1986;203:717.
31 Roy-Camille R, Saillant G, Mazel C: Plating of thoracic, thoracolumbar, and lumbar injuries with
pedicle screw plates. Orthop Clin North Am 1986;17:147159.
32 Suk SI, Kim WJ, Lee SM, et al: Thoracic pedicle screw fixation in spinal deformities: Are they
really safe? Spine 2001;26:20492057.
33 Suk SI, Lee CK, Kim WJ, et al: Segmental pedicle screw fixation in the treatment of thoracic
idiopathic scoliosis. Spine 195;20:13991405.
34 Suk SI, Lee CK, Min HJ, et al: Comparison of Cotrel-Dubousset pedicle screws and hooks in the
treatment of idiopathic scoliosis. Int Orthop 1994;18:341346.
35 Ugur HC, Attar A, Uz A, et al: Thoracic pedicle: Surgical anatomic evaluation and relations.
J Spinal Disord 2001;14/1:3945.
36 Vaccaro AR, Rizzolo SJ, Allardyce TJ, et al: Placement of pedicle screws in the thoracic spine.
I. Morphometric analysis of the thoracic vertebrae. J Bone Joint Surg Am 1995;77:19931999.
37 Vaccaro AR, Rizzolo SJ, Balderston RA, et al: Placement of pedicle screws in the thoracic spine.
II. An anatomic and radiographic assessment. J Bone Joint Surg Am 1995;77:12001206.
38 Xu R, Ebraheim NA, Ou Y, Yeasting RA: Anatomic considerations of pedicle screw placement in
the thoracic spine. Roy-Camille technique versus open-lamina technique. Spine 1998;
23:10651068.
39 Yoo JU, Ghanayem A, Petersilge C, Lewin J: Accuracy of using computed tomography to identify
pedicle screw placement in cadaveric human lumbar spine. Spine 1997;22:26682671.
40 Zindrick MR, Knight GW, Sartori MJ, et al: Pedicle morphology of the immature thoracolumbar
spine. Spine 2000;25:27262735.
41 Zindrick MR, Wiltse LL, Doornik A, et al: Analysis of the morphometric characteristics of the
thoracic and lumbar pedicles. Spine 1987;12:160166.
David W. Polly, Jr., MD, COL, MC, USA

Department of Orthopaedic Surgery and Rehabilitation
Walter Reed Army Medical Center, Washington, DC 20307 (USA)
Polly/Kuklo 106
dramroo
Cervical Techniques with
Image-Guided Spinal Navigation
Iain H. Kalfas
Section of Spinal Surgery, Department of Neurosurgery,
Cleveland Clinic Foundation, Cleveland, Ohio, USA
Image-guided spinal navigation is a computer-based surgical technology that

was developed to improve intraoperative orientation to the unexposed anatomy
during complex spinal procedures [12, 17]. It evolved from the principles of
stereotaxy which have been used by neurosurgeons for several decades to help
localize intracranial lesions. Stereotaxy is defined as the localization of a specific
point in space using three-dimensional coordinates. The application of stereotaxy
to intracranial surgery initially involved the use of an external frame attached to
the patients head. However, the evolution of computer-based technologies has
eliminated the need for this frame and has allowed for the expansion of stereo
tactic technology into other surgical fields, in particular spinal surgery.
The management of complex spinal disorders has been greatly influenced
by the increased acceptance and use of spinal instrumentation devices as well
as the development of more complex operative exposures. Many of these tech
niques place a greater demand on the spinal surgeon by requiring a precise
orientation to that part of the spinal anatomy that is not exposed in the surgical
field. In particular, the various fixation techniques that require placing bone
screws into the pedicles of the thoracic, lumbar and sacral spine, into the lateral
masses of the cervical spine and across joint spaces in the upper cervical spine
require visualization of the unexposed spinal anatomy. Although conventional
intraoperative imaging techniques, such as fluoroscopy, have proven useful,
they are limited in that they provide only two-dimensional imaging of a complex
three-dimensional structure. Consequently, the surgeon is required to extrapo
late the third dimension based on an interpretation of the images and a knowl
edge of the pertinent anatomy. This so-called dead reckoning of the anatomy
can result in varying degrees of inaccuracy when placing screws into the
unexposed spinal column.
dramroo
Several studies have shown the unreliability of routine radiography in
assessing pedicle screw placement in the lumbosacral spine. The rate of pene
tration of the pedicle cortex by an inserted screw ranges from 21 to 31% in these
studies [6, 9, 20]. The disadvantage of these conventional radiographic tech
niques in orienting the spinal surgeon to the unexposed spinal anatomy is that
they display, at most, only two planar images. While the lateral view can be
relatively easy to assess, the anteroposterior or oblique view can be difficult to
interpret. For most screw fixation procedures, it is the position of the screw in
the axial plane that is most important. This plane best demonstrates the position
of the screw relative to the neural canal. Conventional intraoperative imaging
cannot provide this view. To assess the potential advantage of axial imaging for
screw placement, Steinmann et al. [19] used an image-based technique for pedi
cle screw placement that combined computed tomography (CT) axial images of
cadaver spine specimens with fluoroscopy. This study demonstrated an improve
ment in pedicle screw insertion accuracy with an error rate of only 5.5%.
Image-guided spinal navigation minimizes much of the guesswork associ
ated with complex spinal surgery. It allows for the intraoperative manipulation
of multiplanar CT images that can be oriented to any selected point in the
surgical field. Although it is not an intraoperative imaging device, it provides
the spinal surgeon with superior image data compared to conventional intraop
erative imaging technology (i.e. fluoroscopy). It improves the speed, accuracy
and precision of complex spinal surgery while, in most cases, eliminates the
need for cumbersome intraoperative fluoroscopy. This chapter will focus on its
use in the cervical spine.
Principles of Image-Guided Spinal Navigation
The use of an image-guide navigational system for localizing intracranial

lesions has been previously described [1, 2]. Image-guided navigation estab
lishes a spatial relationship between a preoperative CT image data and its
corresponding intraoperative anatomy. Both the CT image data and the anatomy
can each be viewed as a three-dimensional coordinate system with each point
in that system having a specific x, y and z Cartesian coordinate. Using defined
mathematical algorithms, a specific point in the image data set can be matched
to its corresponding point in the surgical field. This process is called registra
tion and represents the critical step of image-guided navigation. A minimum of
three points needs to be matched or registered, to allow for accurate navigation.
A variety of navigational systems have evolved over the past decade. The
common components of most of these systems include an image-processing
computer workstation interfaced with a two-camera optical localizer (fig. 1).
Kalfas 108
dramroo
Fig. 1. Image-guided navigational workstation with infrared camera localizer system.
When positioned during surgery, the optical localizer emits infrared light
towards the operative field. A hand-held navigational probe mounted with a
fixed array of passive reflective spheres serves as the link between the surgeon
and the computer workstation (fig. 2). Alternatively, passive reflectors may be
attached to standard surgical instruments. The spacing and positioning of the
passive reflectors on each navigational probe or customized trackable surgical
instrument are known by the computer workstation. The infrared light that is
transmitted towards the operative field is reflected back to the optical localizer
by the passive reflectors. This information is relayed to the computer worksta
tion which can then calculate the precise location of the instrument tip in the
surgical field as well as the location of the anatomic point on which the instru
ment tip is resting.
The initial application of navigational principles to spinal surgery was not
intuitive. Early navigational technology applied to intracranial surgery used an
external frame mounted to the patients head to provide a point of reference to
link preoperative image data to intracranial anatomy. This was not practical for
spinal surgery. The current generation of intracranial navigational technology
Cervical Techniques with Image-Guided Spinal Navigation 109

dramroo
Fig. 2. Navigation probe and drill
guide for spinal surgery.
Fig. 3. Navigational workstation screen

demonstrating a paired point registration plan
for the insertion of C2 pedicle screws. Three
discreet bony landmarks are selected at the
C2 level. In this case the lateral margins of the
two C23 facets and the spinous process tip
of C2 have been selected.
uses reference markers or fiducials that are glued to the patients scalp prior to
imaging. However, the use of these surface-mounted fiducials for spinal navi
gation is not practical because of accuracy issues related to a greater degree of
skin movement over the spinal column [4, 5]. This is less of a problem with
intracranial applications because of the relatively fixed position of the overly
ing scalp to the attached fiducials.
The application of navigational technology to spinal surgery involves
using the rigid spinal anatomy itself as a frame of reference. Bone landmarks
on the exposed surface of the spinal column provide the points of reference
necessary for image-guided navigation. Specifically, any anatomic landmark
that can be identified intraoperatively as well as in the preoperative image data
set can be used as a reference point. The tip of a spinous or transverse process,
a facet joint or a prominent osteophyte can all serve as potential reference
points (fig. 3). Since each vertebra is a fixed, rigid body, the spatial relationship
of the selected registration points to the vertebral anatomy at a single spinal
level is not affected by changes in body position.
Two different registration techniques can be used for spinal navigation,
paired point registration and surface matching. Paired point registration involves
Kalfas 110
dramroo
selecting a series of corresponding points in a CT or magnetic resonance imag
ing data set and in the exposed spinal anatomy. The registration process is
performed immediately after surgical exposure and prior to any planned
decompressive procedure. This allows for the use of the spinous processes as
registration points.
A specific registration point in the CT image data set is selected by high
lighting it with the computer cursor. The tip of the probe is then placed on the
corresponding point in the surgical field and the reflective spheres on the probe
handle are aimed towards the camera. Infrared light from the camera is
reflected back allowing the spatial position of the probes tip to be identified.
This initial step of the registration process effectively links the point selected in
the image data with the point selected in the surgical field. When a minimum
of three such points are registered, the probe can be placed on any other point
in the surgical field and the corresponding point in the image data set will be
identified on the computer workstation.
Alternatively, a second registration technique called surface matching can be
used. This technique involves selecting multiple nondiscreet points only on the
exposed and debrided surface of the spine in the surgical field. This technique
does not require the preselection of points in the image set although several dis
creet points in both the image data set and in the surgical field are frequently
required to improve the accuracy of surface mapping. The positional information
of these points is transferred to the workstation and a topographic map of the
selected anatomy is created and matched to the patients image set [18].
Typically, paired point registration can be done more quickly than surface
mapping. The average time needed for paired point registration is 1015 s. The
time needed for surface mapping is much longer with difficult cases requiring
as much as 1015 min. With the need to perform several registration processes
during each surgery, this time difference can significantly impact the length of
the navigational procedure and the surgery itself [16].
The purpose of the registration process is to establish a precise spatial rela
tionship between the image space of the data with the physical space of the
patients corresponding surgical anatomy. If the patient is moved after registra
tion, this spatial relationship is distorted making the navigational information
inaccurate. This problem can be minimized by the optional use of a spinal
tracking device which consists of a separate set of passive reflectors mounted
on an instrument that can be attached to the exposed spinal anatomy (fig. 4).
The position of the reference frame can be tracked by the camera system.
Movement of the frame alerts the navigational system to any inadvertent move
ment of the spine. The system can then make correctional steps to keep the
registration process accurate and eliminate the need to repeat the registration
process. The disadvantage of using a tracking device is the added time needed

dramroo
Fig. 4. Reference frame attached to a
spinous process in the surgical field. The
reference frame monitors inadvertent move
ment of the spinal anatomy that may affect
navigational accuracy.
for its attachment to the spine, the need to maintain a line of sight between it
and the camera and the inconvenience of having to perform the procedure with
the device placed in the surgical field. It is particularly cumbersome when
image-guided navigation is used during cervical procedures.
Alternatively, image-guided spinal navigation can be performed without a
tracking device [12, 16]. This involves acknowledging the effect of patient
movement on the accuracy of image-guided navigation and maintaining a
reasonably stable patient position during the relatively short amount of time
needed (i.e. 1020 s) for the selection of each appropriate screw trajectory.
Patient movement can potentially occur with respiration, from the surgical team
leaning on the table or from a change of table position. Movement associated
with patient respiration is negligible and does not require any tracking even in
the thoracic spine. Although movement associated with leaning on the table or
repositioning the table or the patient will affect registration accuracy, it can be
easily avoided during the short navigational procedure. If inadvertent patient
movement does occur, the registration process can be repeated. Repeating the
registration process is easiest when using the shorter paired point technique as
opposed to the more time-consuming surface mapping technique.
When the registration process has been completed, the probe can be posi
tioned on any surface point in the surgical field and three separate reformatted
CT images centered on the corresponding point in the image data set are imme
diately presented on the workstation monitor. Each reformatted image is refer
enced to the long axis of the probe. If the probe is placed on the spinal anatomy
directly perpendicular to its long axis, the three images will be in the sagittal,
coronal and axial planes. A trajectory line representing the orientation of the
Kalfas 112
dramroo
long axis of the probe will overlay the sagittal and axial planes. A cursor repre
senting a cross section through the selected trajectory will overlay the coronal
plane. The insertional depth of the trajectory can be adjusted to correspond to
selected screw lengths. As the depth is adjusted, the specific coronal plane will
also adjust accordingly with the position of the cursor demonstrating the final
position of the tip of a screw placed at that depth along the selected trajectory.
As the probe is moved to another point in the surgical field, the reformatted
images as well as the position of the cursor and trajectory line will also change.
The planar orientation of the three reformatted images will also change as the
probes angle relative to the spinal axis changes. When the probes orientation
is not perpendicular to the long axis of the spine, the images displayed will be
in oblique or orthogonal planes. Regardless of the probes orientation, the
navigational workstation will provide the surgeon with a greater degree of
anatomic information than can be provided by any intraoperative imaging
technique.
The application of image-guided navigation to spinal surgery is directed by
the complexity of the procedure and, specifically, by the need to visualize the
unexposed spinal anatomy. Image-guided navigation can be used with or with
out standard intraoperative imaging techniques (i.e. fluoroscopy). In either case,
image-guided navigation provides the surgeon with an improved orientation to
the pertinent spinal anatomy, which subsequently facilitates the accuracy and
effectiveness of the procedure.
Clinical Applications
Prior to using image-guided navigation for spinal surgery, testing of the

system was carried out in cadaver spine specimens. Image guidance was used
to direct screws into the thoracic and lumbosacral pedicles of these specimens.
The accuracy of screw insertion was assessed with plain film radiography and
thin section CT imaging of the instrumented levels. All inserted pedicle screws
were determined to be satisfactorily positioned [17].
The initial clinical application of image-guided spinal navigation was for
lumbosacral pedicle fixation [8, 10, 12]. It proved to be an efficient and effec
tive replacement for intraoperative fluoroscopy and was gradually applied to
other spinal procedures such as thoracic pedicle fixation and anterior thora
columbar decompression and screw fixation. The use of image-guided technol
ogy in the cervical spine has also evolved. It is now used for such procedures
as C12 transarticular screw fixation, lateral mass screws at C1, pedicle screws
at C2 and C7 and anterior procedures such as transoral surgery and cervical
corpectomy [3, 16, 21].

dramroo
Fig. 5. Positioning of the image-guided system for a C12 transarticular screw fixation
procedure. The camera is positioned at the head of the table in a vertical orientation.
This minimizes any potential visual obstruction between the camera and the surgical field.
The workstation is positioned across the table from the surgeon (left arrow). The fluoro
scopic monitor sits next to the navigational workstation (right arrow).
Navigational Technique
Image-guided navigation requires the acquisition of a preoperative CT scan
through the appropriate spinal segments to be instrumented. The image data is
then transferred to the computer workstation via optical disc or a high-speed data
link. If paired point registration is to be used, three to five reference points for
each spinal segment to be instrumented are selected and stored in the image data
set. For most cervical procedures, the camera can be positioned at the head of
the table with the navigation workstation positioned across from the surgeon.
If fluoroscopy is also used, it can be positioned next to the workstation (fig. 5).
Following a standard surgical exposure, either the paired point or surface
matching registration technique is performed. When the registration process
has been completed, most navigational workstations will calculate a registration
error (expressed in millimeters) that is directly dependent on the surgeons
registration technique. The error presented does not represent a linear error but
rather a volumetric calculation comparing the spacing of registration points in
the surgical field to the spacing of the corresponding points in the image data
set. This figure is, at best, a relative indicator of accuracy.
A more practical method of assuring registration accuracy is the verifica
tion step. This step is typically performed immediately after completing either
Kalfas 114
dramroo
registration process. The surgeon places the navigational probe on a discreet
landmark in the surgical field. With the navigational system now tracking the
movement and position of the probe, the trajectory line and cursor on the work
station screen will, if accurate registration has been achieved, move to the cor
responding point in the image data set (fig. 6a). If registration accuracy has not
been achieved, the cursor and trajectory line may rest on something other than
the point selected in the surgical field (fig. 6b). If this occurs to a significant
degree, the registration process needs to be repeated. This step is more of an
absolute indicator of registration accuracy and is a necessary step to perform
prior to proceeding with navigation.
C12 Transarticular Screw Fixation

This procedure involves the passage of a screw through the pars interartic
ularis of C2, across the facet joint and into the lateral mass of C1. The risks of
screw insertion include injury to the vertebral artery if the screw is placed too
laterally or ventrally, injury to the spinal cord if the screw is placed too medi
ally, and failure to engage the lateral mass of C1 if the screw trajectory is too
ventral. The insertion of a screw on either side may be contraindicated if the
pars interarticularis of C2 is too narrow. The procedure is typically performed
bilaterally using fluoroscopic guidance.
The selection of the appropriate screw entry site and trajectory requires a
thorough understanding of the atlantoaxial anatomy. Although fluoroscopy pro
vides real-time imaging of the relevant spinal anatomy, the views generated rep
resent only two-dimensional images of a complex three-dimensional anatomic
region. Manipulation of the fluoroscopic unit can reduce this problem but these
maneuvers can be cumbersome and time-consuming. Other disadvantages
include the radiation exposure and the need to wear lead aprons during the pro
cedure. Fluoroscopy cannot provide a view of the spinal anatomy in the axial
plane. It is this axial view provided by image-guided navigation that makes it
superior to fluoroscopy for spinal screw fixation procedures. The application of
image-guided navigation to this procedure adds a significant layer of accuracy
for proper screw placement.
The technique for applying image-guided navigation to posterior C12
screw fixation involves acquiring a preoperative CT scan that extends from the
lower occipital region to C3. The image data is transferred to the computer
workstation and can be used to create a preoperative screw trajectory plan.
A proposed entry point and target can be selected at the C2 and C1 levels,
respectively. The image data set can then be manipulated in multiple planes
between these two points to demonstrate the position of a screw placed along
the selected trajectory. In addition to a sagittal image that demonstrates the
same information provided by lateral fluoroscopy, two other images are

dramroo
6a
6b
Kalfas 116
dramroo
presented. One of the images lies perpendicular to the sagittal image along the
selected trajectory. It represents an orthogonal view that lies approximately
midway between the coronal and axial planes through the spine. It demonstrates
a second view of the selected trajectory.
An additional view demonstrates an image oriented perpendicular to the
long axis of the probe and, therefore, the selected trajectory. A cursor superim
posed on this image can show the position of the screw tip along the selected
trajectory at millimetric increments. By scrolling through this image, the
proposed position of the screw along the selected trajectory can be assessed
along its entire path. While this planning technique does not assure safe screw
placement intraoperatively, it can preoperatively alert the surgeon to avoid
screw placement in patients with insufficient anatomy and to select an alternate
approach.
Intraoperatively, the patient is positioned and the posterior C12 complex
is exposed. A wire (cable) and bone graft stabilization procedure at the C12
level is performed prior to navigation and screw insertion. Performing this step
first minimizes any independent motion between C1 and C2 during navigation
and makes tap and screw insertion easier. If a reference frame is used, it is
typically attached to the spinous process of C2.
Following placement of the graft and cable, three to five registration points
are selected at the C2 level. It is not necessary to include registration points at
C1. Although the spatial relationship of C1 and C2 may change between the
preoperative scanned position and the intraoperative position, the ability of
image-guided navigation to facilitate accurate screw placement is not signifi
cantly affected. The technical difficulty of this procedure is the accurate pas
sage of the screw through the narrow pars interarticularis of C2. The lateral
mass of C1 is a relatively large target that can be easily reached provided there
is a reasonably acceptable realignment of C1 and C2 as well as an optimal posi
tioning of the screw within the appropriate C2 anatomy. While the relative posi
tion of C1 and C2 in both the preoperative image set and in the surgical field is
important, it is not critical enough to interfere with the process of image-guided
navigation.
Fig. 6. a Navigational workstation screen demonstrating satisfactory verification of reg

istration accuracy. While the navigational probe is positioned on the C2 spinous process in the
surgical field, the workstation screen should show the cursor and trajectory line in a correlative
position in the CT image set. b Navigational workstation screen demonstrating an unsatisfactory
verification of registration accuracy. If the navigational probe is positioned on the C2 spinous
process in the surgical field but the workstation screen shows the cursor and trajectory line in a
noncorrelative position (i.e. not on the C2 spinous process) accurate registration has not been
achieved and the registration process needs to be repeated before proceeding with navigation.

dramroo
Fig. 7. Workstation screen demonstrating a trajectory for insertion of a C12 transar
ticular screw. The lower right screen shows the trajectory in the sagittal plane. The lower left
screen represents an orthogonal plane lying between the axial and coronal planes. It conveys
the medial-lateral trajectory. The upper left screen represents a plane that is perpendicular to
the two other images. It demonstrates the location of the screw tip inserted along the selected
trajectory at the indicated depth.
Two separate stab incisions are made on either side of the midline at the
C7T1 level. A drill guide is placed through one of the stab incisions, passed
through the paravertebral musculature and into the operative field. A small
divot is drilled at the proposed entry site in order to provide for secure place
ment of the drill guide. The registration process is performed at the C2 level and
its accuracy confirmed using the verification step. The probe is passed through
the drill guide and, as its position is adjusted in the surgical field, the images
on the workstation screen will adjust accordingly to show the corresponding
trajectory in two separate planes and the projected location of the screw tip in
the third plane. Orientation to the correct screw position can be assessed rapidly
and accurately (fig. 7). Any errors in trajectory or entry point selection can be
determined and corrected by adjusting the position of the probe and the drill
guide through which it passes. When the correct screw insertion parameters
have been selected, the probe is removed from the drill guide and a drill
inserted. A hole is drilled along the selected trajectory, tapped and the appro
priate length screw inserted. The process is repeated on the opposite side.
Kalfas 118
dramroo
The purpose of the drill guide is to preserve the physical trajectory and
entry point information just acquired through the navigation of that pedicle. If a
drill guide is not used, it may be difficult to precisely position a drill or pedicle
probe on the same point and with the same trajectory previously conveyed by
the navigational probe after probe removal.
While image-guided navigation does not guarantee accurate screw place
ment, it does provide the surgeon with a greater degree of anatomical informa
tion than fluoroscopy alone. The addition of fluoroscopy to this navigational
technique provides the greatest degree of precision to the procedure. In this case
however, navigational technology significantly reduces the time of intraoperative
fluoroscopic usage as it is typically used only to help position the patient preop
eratively and as a final check of the selected trajectory in the sagittal plane imme
diately following the navigational step.
Segmental C12 Screw Fixation

As an alternative to transarticular screw fixation, segmental fixation of
C12 can be used for managing atlantoaxial instability [11]. The procedure
involves placing a screw into each of the two lateral masses of C1 and two
screws down the pedicles of C2. The polyaxial screw heads on each side are
then connected with rods. Although this approach potentially reduces the risk
of injury to the vertebral artery during screw insertion, it does not eliminate the
risk altogether. As with the transarticular technique, precise anatomic orienta
tion is required to avoid arterial or neural injury. Image guidance can supple
ment intraoperative fluoroscopy in order to provide the necessary orientation
for accurate screw insertion.
As with the transarticular screw fixation technique, a preoperative CT is
obtained. The posterior C12 spine is exposed and a wire and cable fixation pro
cedure is carried out. Registration is first performed at C1 for placement of the
C1 lateral mass screws. The three registration points typically used at C1 include
the midline posterior tubercle and the bilateral points marked by the junction of
the small pedicle of C1 with its lateral mass (immediately above the two exiting
C2 nerve roots). Once registered, the correct trajectory into the lateral mass can
be displayed on the workstation screen and the screws inserted (fig. 8).
To use image guidance for inserting C2 pedicle screws, the same registra
tion points are used at C2 as those used for transarticular fixation (the C2
spinous process and the two lateral margins of the C23 facet). The entry point
for the screw is more lateral and the trajectory more medially oriented than for
a transarticular screw. The navigation probe is placed through a drill guide onto
this entry point and the selected trajectory is displayed on the workstation
screen. When the correct entry point and trajectory have been selected, the
probe is removed, a drill is inserted and the pilot hole is drilled (fig. 9).

dramroo
Fig. 8. Workstation screen demonstrating navigational information for placement of a
screw into the lateral mass of C1.
Fig. 9. Workstation screen demonstrating navigational information for placement of a

screw into the pedicle of C2.
Kalfas 120
dramroo
The process is repeated for the other side. The heads of the screws are then
connected with two short rods.
Transoral Surgery
Transoral decompression of the upper cervical spine typically requires
intraoperative fluoroscopy to help maintain proper anatomic orientation during
the procedure. Although orientation in the sagittal plane is easy to obtain with
fluoroscopy, depth and medial-lateral orientation are more difficult to assess.
Image-guided technology can be used to orient the surgeon in multiple planes
during transoral surgery [16, 21].
Unlike other spinal applications of image guidance, discreet registration
points are not readily available during transoral surgery. In this setting,
surface-mounted markers (fiducials) are applied to the patient prior to
obtaining the preoperative CT. Typically, two fiducials are applied to the
mastoid processes and two are applied to the lateral orbital margins or to
both malar eminences. The nasal septum can also be used as an inherent
registration point.
The patient is positioned in a three-point head holder. The registration
process is performed prior to draping the patient using the surface-mounted
fiducials. Because the registration points will not be accessible during the
procedure, a reference frame is used for transoral navigation. This allows for
changes in patient positioning during surgery without the need to re-register.
The reference frame can be attached to the three-point head holder.
During the procedure, the probe can be placed into the site of the decom
pression. Reformatted sagittal, axial and coronal CT images are immediately
generated providing the surgeon with a precise orientation to the pertinent
surgical anatomy. In particular, orientation in the axial plane minimizes the risk
of lateral deviation towards the vertebral artery during the decompression
(fig. 10). If a posterior fixation is indicated following transoral decompression,
the same CT image data set can be used for C12 screw placement.
Other Cervical Applications
There are several other applications for image-guided navigation in the

cervical spine. Any procedure in which intraoperative imaging is required to
improve a surgeons orientation to nonexposed spinal anatomy can benefit from
image guidance. The other cervical procedures to which image guidance has
been applied include navigation during the removal of cervical neoplasms
(fig. 11), the placement of anterior fixation screws for the management of
nondisplaced odontoid fractures, lateral mass screw fixation in the subaxial

dramroo
Fig. 10. Workstation screen demonstrating navigational information during transoral
decompression (probe tip location and trajectory highlighted by arrows).
Fig. 11. Workstation screen demonstrating the use of image guidance to help localize
an osteoid osteoma within the lamina and articular pillar of C7.
Kalfas 122
dramroo
Fig. 12. Workstation screen demonstrating navigational information for the placement
of a screw into the pedicle of C7.
spine, cervical corpectomy and the placement of pedicle screws into C7

(fig. 12) [3, 1316]. While anterior cervical applications of image guidance
present unique registration difficulties because of the relative lack of discreet
registration points on the anterior surface of the cervical spine, sufficient
navigational information can be obtained in order to improve the precision of
these procedures.
Pitfalls of Image-Guided Spinal Navigation
Like any other computer-based technology, image-guided navigation is

highly dependent on the quality of the information imported into the system.
While obtaining the properly formatted CT images and having them correctly
transferred to the navigational workstation is important, the critical step of
image guidance is the registration process. If the surgeon takes too casual an
approach to registration, inaccurate information will be displayed during intra
operative navigation.
Another important principle of image guidance is the need to correlate the
navigational information with the surgeons own knowledge of the surgical

dramroo
anatomy and the appropriate screw trajectories through that anatomy. Image-
guided navigation is not a replacement for the surgeon knowing the pertinent
spinal anatomy and surgical technique. It merely serves to help confirm a sur
geons estimation of the nonexposed anatomy by providing image information
that typically exceeds that of intraoperative fluoroscopy.
Image-guided technology also has varying degrees of intraoperative func
tionality depending on the features of the navigational system used. This trans
lates into an ease of use factor that can either simplify or complicate the overall
procedure. Typically, the use of the surface mapping registration technique
and a reference frame add time to the navigational procedure, frequently making
it longer and more complicated than using fluoroscopy alone. The use of the
paired point registration technique without a reference frame further simplifies
the spinal navigation process. The option of using a reference frame is depen
dent on the particular navigational system used.
Fluoroscopic Navigation
Fluoroscopic navigation is the combination of standard fluoroscopy with

image-guided navigational technology. It was developed to counter the user dif
ficulties of some earlier image-guided systems that typically took much longer
to use than standard fluoroscopy. Its advantage is that it allows for a reduction
in fluoroscopic time during the procedure. With the patient in position prior to
surgery, an anteroposterior and lateral fluoroscopic view of the pertinent spinal
anatomy is obtained. This is done with a customized reference frame attached
to the C-arm or to the patient. This frame serves to superimpose a specific grid
on the two images obtained. The navigational workstation can then take the two
images and relate the spatial position of the imaged anatomy to a navigational
probe. A navigational trajectory line and cursor can be superimposed on the
lateral and anteroposterior images, respectively. As the probe is moved over the
exposed spinal anatomy during surgery the trajectory line and cursor will adjust
their position on the stationary fluoroscopic images [7].
The disadvantage of fluoroscopic navigation is that it is still only fluoroscopy.
The same difficulties experienced with standard fluoroscopy are present
with fluoroscopic navigation. Specifically, only an anteroposterior and lateral
images are provided. The critical plane for most spinal screw fixation proce
dures is the axial plane. This is the only plane that can definitively demonstrate
violation of the spinal canal by a medially displaced screw. Only CT-based image-
guided navigation can demonstrate this view although current developments
with standard fluoroscopy will eventually allow for axial reconstructions.
Furthermore, any region of the spinal column that is difficult to image with
Kalfas 124
dramroo
standard fluoroscopy (i.e. upper thoracic) will be difficult to image with
fluoroscopic navigation.
The early goals of image-guided spinal navigation were to improve the
surgeons orientation to the intraoperative spinal anatomy in a time- and cost-
efficient manner and to ultimately replace fluoroscopy. While earlier image-
guided systems were difficult to use intraoperatively, several advances have
made some systems much easier to use. Several years of clinical experience
have helped modify and improve navigational techniques. The use of paired
point registration and the optional use of a reference frame have both been
found to significantly reduce the difficulties of using image-guided technology
for spinal procedures. Advances in computer and localizer technology have also
contributed to an improved functionality of these systems. This improved ease
of use of the advanced image-guided systems coupled with superior accuracy,
image manipulation and orientation capabilities provides image-guided
technology with a clear advantage over any fluoroscopic-based technology.
Conclusion
Image-guided navigational technology has been successfully applied to

spinal surgery. By linking digitized image data to spinal surface anatomy, image-
guided spinal navigation facilitates the surgeons orientation to unexposed
spinal structures improving the precision and accuracy of the surgery. It is typ
ically used to optimize the placement of spinal fixation screws and to monitor
the extent of complex decompressive procedures. It can also be used as a
preoperative planning tool.
While image-guided spinal navigation is a versatile and effective technol
ogy, it is not a replacement for the surgeon having a thorough knowledge of the
pertinent spinal anatomy as well as correct surgical techniques. It merely serves
as an additional source of information used by the surgeon to make selected
intraoperative decisions. In this way, it is similar to more conventional intraop
erative imaging techniques (i.e. fluoroscopy) except that it provides a much
greater degree of information to the surgeon.
Despite the advantages of image guidance, the surgeon must ultimately assess
the information provided by these systems and determine if it correlates with his
or her estimation of the nonexposed anatomy and the proposed surgical plan.
If good correlation exists between the two, the surgical step can be carried out.
However, if sufficient correlation is not present, the surgeon needs to reassess both
the spinal anatomy and the image-guided registration accuracy before proceeding.
Ideally, the clinical application of this technology to spinal surgery should
facilitate a reduction in operative time, morbidity, and costs. It should be capable

dramroo
of minimizing or eliminating the need for conventional intraoperative imaging.
It should be fast, easy to use, reliable and capable of being used briefly to
provide accurate intraoperative information while minimizing any disruption to
the standard routine of each surgical procedure. Ultimately, beyond each indi
vidual surgical application, image-guided navigation technology needs to be
clinically versatile. It is the routine use of this technology by multiple surgical
specialties that will drive its continued evolution and development as well as
establishing it as a cost-effective surgical tool.
References
1 Barnett GH, Kormos DW, Steiner CP, Weisenberger J: Use of a frameless, armless stereotactic
wand for brain tumor localization with two-dimensional and three-dimensional neuroimaging.
Neurosurgery 1993;33:674678.
2 Barnett GH, Kormos DW, Steiner CP, Weisenberger J: Intraoperative localization using an arm
less, frameless stereotactic wand. Technical note. J Neurosurg 1993;78:510514.
3 Bolger C, Wigfield C, Melkent T, Smith K: Frameless sterotaxy and anterior cervical surgery.
Comput Aided Surg 1999;4:322327.
4 Brodwater BK, Roberts DW, Nakajima T, Friets EM, Strohbehn JW: Extracranial application of
the frameless stereotactic operating microscope: Experience with lumbar spine. Neurosurgery
1993;32:209213.
5 Bryant JT, Reid JG, Smith BL, Stevenson JM: A method for determining vertebral body positions
in the sagittal plane using skin markers. Spine 1989;14:258265.
6 George DC, Krag MH, Johnson CC, Van Hal ME, Haugh LD, Grobler LJ: Hole preparation tech
nique for transpedicle screws: Effect on pull-out strength from human cadaveric vertebrae. Spine
1991;16:181184.
7 Foley KT, Simon DA, Rampersaud YR: Virtual fluoroscopy: Computer-assisted fluoroscopic
navigation. Spine 2001;26:347351.
8 Foley KT, Smith MM: Image-guided spine surgery. Neurosurg Clin N Am 1996;7/2:171186.
9 Gertzbein SD, Robbins SE: Accuracy of pedicle screw placement in vivo. Spine 1990;15:
1114.
10 Glossop ND, Hu RW, Randle JA: Computer-aided pedicle screw placement using frameless
sterotaxis. Spine 1996;21:20262034.
11 Harms J, Melcher R: Posterior C1C2 fusion with polyaxial screw and rod fixation. Spine 2001;
26:24672471.
12 Kalfas IH, Kormos DW, Murphy MA, McKenzie RL, Barnett GH, Bell GR, Steiner CP, Trimble
MB, Weisenberger JP: Application of frameless stereotaxy to pedicle screw fixation of the spine.
J Neurosurg 1995;83:641647.
13 Kalfas IH: Spinal surgery; in Barnett GH, Roberts DW, Maciunas RJ (eds): Image-Guided
Neurosurgery: Clinical Applications of Surgical Navigation. St Louis, Quality Medical Publishing,
1998, pp 117134.
14 Kalfas IH: Image-guided spinal navigation: Application to spinal metastasis; in Maciunas RJ (ed):
Advanced Techniques in Central Nervous System Metastasis. Lebanon, AANS Publications, 1998,
pp 245254.
15 Kalfas IH: Frameless stereotaxy assisted spinal surgery; in Renganchary SS (ed): Neurosurgery
Operative Color Atlas. Lebanon, AANS Publications, 2000, pp 123134.
16 Kalfas IH: Image-guided spinal navigation. Clin Neurosurg 1999;46:7088.
17 Murphy MA, McKenzie RL, Kormos DW, Kalfas IH: Frameless stereotaxis for the insertion of
lumbar pedicle screws: A technical note. J Clin Neurosci 1994;1/4:257260.
Kalfas 126
dramroo
18 Pellizzari CA, Levin DN, Chen GTY, Chen CT: Image registration based on anatomic surface
matching; in Maciunas RJ (ed): Interactive Image-Guided Neurosurgery. Park Ridge, American
Association of Neurological Surgeons, 1993, pp 4762.
19 Steinmann JC, Herkowitz HO, El-Kommos H, Wesolowski DP: Spinal pedicle fixation:
Confirmation of an image-based technique for screw placement. Spine 1993;18:18561861.
20 Weinstein JN, Spratt KF, Spengler D, Brick C, Reid S: Spinal pedicle fixation: Reliability and
validity of roentgenogram-based assessment and surgical factors on successful screw placement.
Spine 1988;13:10121018.
21 Welch WC, Subach BR, Pollack IF, Jacobs GB: Frameless sterotactic guidance for surgery of the
upper cervical spine. Neurosurgery 1997;40:958964.
Iain H. Kalfas, MD, FACS
Department of Neurosurgery (S80)
Cleveland Clinic Foundation
9500 Euclid Ave., Cleveland, OH 44195 (USA)
Tel. 1 216 444 9064, Fax 1 216 445 6801

dramroo
Instrumentation and Technique Update
C1 Lateral Mass Fixation

Amory J. Fiorea, Barry D. Birchb, Regis W. Haid, Jr.a
a
Department of Neurosurgery, The Emory Clinic, Atlanta, Ga., and
b
Department of Neurosurgery, The Mayo Clinic, Scottsdale, Ariz., USA
A variety of techniques exist for fixation of the upper cervical spine.

Recently, several authors have published case series reporting the use of C1
lateral mass screws for posterior cervical fixation [13]. In this chapter we
describe our surgical technique for placement of C1 lateral mass screws, includ
ing indications and results from our experience.
C1 lateral mass screws may be used to provide additional fixation points
in occipitocervical constructs, possibly increasing resistance to construct
failure in the cervical spine without increasing the number of cervical levels
fused. Additionally, C1 lateral mass screws may be used as a supplement to or
substitute for other forms of atlantocervical fixation. Techniques for achieving
atlantocervical fusion include posterior interspinous fusion with sublaminar
cables and iliac crest bone graft [4, 5], C1-C2 transarticular screw fixation
[57], and interlaminar clamp fixation [8]. While each of these methods has
been successfully employed to achieve atlantocervical fusion, anatomic factors
may exist in certain situations that preclude their use. Interspinous fusion at
C1-C2 with sublaminar cables or interlaminar clamps cannot be performed if
the posterior elements of C1 or C2 are absent or disrupted. C1-C2
transarticular screws cannot be placed successfully in the presence of a
medially located vertebral artery, irreducible subluxation, severe cervicotho
racic kyphosis, or destruction of the C2 pars interarticularis. In these cases
constructs employing C1 lateral mass screws may be used to achieve fixation.
We present a small case series in which C1 lateral mass screws were used to
achieve atlantocervical fixation when anatomic characteristics precluded the
use of traditional fixation methods. Also included in this series is one case
where C1 lateral mass screws were used to provide additional fixation points
for occipitocervical fusion.
dramroo
Methods
Ten patients with cervicomedullary compression or atlantoaxial instability were treated

surgically between February 1998 and February 2002. Preoperative diagnoses included C2
metastasis in 2 patients, irreducible odontoid fracture in 3 patients, atlantoaxial subluxation
in 2 patients, and transverse ligament synovial cyst in 3 patients. Posterior atlantocervical
fixation was planned in 9 patients, one of whom had a transoral resection of the odontoid
prior to posterior fusion. In 1 patient with a C2 metastasis and pathologic fracture, an occip
itocervical fusion was planned due to the high degree of instability of the atlantoaxial
complex. Atlantoaxial screw fixation was chosen as the initial fixation procedure when pre
operative imaging studies did not reveal anatomic factors precluding screw placement. All
procedures were performed with intraoperative lateral fluoroscopy or CT-guided frameless
stereotaxic navigation (Stealth Station, Sofamor Danek, Memphis, Tenn., USA). Autogenous
iliac crest bone graft was harvested via a separate posterior iliac crest incision and used for
arthrodesis.
Surgical Technique
The patient is positioned prone using a Mayfield head holder (OMI, Cincinnati, Ohio,
USA). The neck is kept neutral with the head in the military tuck position. The arms are
tucked at the sides and the shoulders retracted caudally using tape. A midline incision is
made extending from the inion to the spine of C3 if atlantoaxial fixation is planned. The
incision is extended inferiorly as indicated by the planned procedure. A bilateral sub
periosteal dissection of the paraspinal musculature is performed to expose the lateral margins
of the facet joints at the C2-C3 level. Dissection is continued laterally over the dorsal arch
of C1, exposing the vertebral artery in the vertebral groove on the C1 arch. Bipolar cautery
and hemostatic agents such as gelfoam and fibrillar collagen are used to control bleeding
from the perivertebral venous plexus. The C2 nerve root is identified and mobilized inferiorly.
The lateral mass of C1 inferior to the C1 arch is exposed. The medial wall of the lateral mass
is identified using a forward angle curette to identify the medial limit of screw placement.
The medial aspect of the transverse foramen can also be identified and serves as the lateral
limit for screw placement. The entry point for screw placement is identified 35 mm lateral to
the medial wall of the lateral mass, at the junction of the lateral mass and inferior aspect of
the C1 arch (fig. 1). The entry point may be varied depending on the distance between the
medial wall of the lateral mass and the C1 transverse foramen. A high speed drill with
a 3-mm round burr is used to remove a small portion of the inferior aspect of the C1 arch
overlying the entry point, to create a recess for the screw head and plate or rod
(fig. 2a). An assistant retracts the C2 nerve inferiorly and protects the vertebral artery with
Penfield dissectors or similar instruments during drilling and screw placement. Using fluo
roscopy or image guidance a 3-mm drill bit and guide are used to drill a hole with 1015
of medial angulation to penetrate the anterior cortex of C1 (fig. 2b, 3). On lateral fluoros
copic imaging the drill is aimed toward the anterior tubercle of C1, so that the drill pene
trates the ventral cortex of the lateral mass midway between the superior and inferior facets
of C1 (fig. 2c, 4). The hole is tapped with a 3.5-mm tap. If lateral mass plates are used, an
appropriate-sized plate is selected and contoured, after which a 3.5-mm screw is placed
through the plate (fig. 5). Caudal fixation points are then finalized. If a polyaxial screw-rod
C1 Lateral Mass Fixation 129

dramroo
Fig. 1. Intraoperative photograph dem
onstrating entry point for C1 lateral mass
screw, vertebral artery in vertebral groove of
C1, inferior aspect of C1 transverse foramen,
and C2 nerve root retracted inferiorly.
system is used, all screws are placed after which an appropriate sized and contoured rod is
secured (fig. 2d). In both cases an appropriate screw length is selected to achieve bicortical
fixation.
Once instrumentation placement is complete, decompression is performed if necessary.
Finally, arthrodesis is performed. Posterior arthrodesis with sublaminar cable and inter
spinous bicortical autograft is preferred if the laminae of C1 and C2 are preserved.
Otherwise, lateral arthrodesis is performed by carefully decorticating the exposed surfaces
of the C1-C2 joints with a high-speed drill, and then packing cancellous iliac crest autograft
over these joints. A Hemovac drain is placed prior to wound closure.
Results
Seventeen C1 lateral mass screws were placed in 8 patients (table 1). These
screws were incorporated into several different constructs using lateral mass
plates (Axis, Sofamor Danek) or a polyaxial screw-rod system (Vertex, Sofamor
Danek) to achieve atlantocervical fixation in 9 patients and occipitocervical
fixation in 1 patient. There were no intraoperative complications and no vertebral
artery injuries. One patient died on postoperative day 9 from complications of
aspiration pneumonia. The remaining patients were immediately mobilized post
operatively in hard cervical collars worn for 3 months. Immediate rigid fixation
was achieved in all patients. Follow-up ranged from 9 days to 18 months (mean
6.6 months). Osseous fusion was documented in 2 patients on 9- and 18-month
Fiore/Birch/Haid 130
dramroo
a b
c d
Fig. 2. a Recess for the screw head is created on the inferior aspect of the C1 arch. b Axial
drawing shows medial screw angulation and relationship to transverse foramina. c Lateral view
of C1-C2 construct with C1 lateral mass screw, C2 pedicle screw, and plate, demonstrating screw
trajectory. d Posterior view of completed C1-C2 construct using polyaxial screw and rod system.
Fig. 3. Axial CT images showing medial angulation of C1 screws and relationship to

transverse foramina.

dramroo
Fig. 4. Intraoperative fluoroscopic
image demonstrating C1 screw trajectory
toward midpoint of anterior C1 tubercle.
Fig. 5. Completed C1-C2 construct

using lateral mass plates and screws.
follow-up radiographs, and 6 patients had delayed postoperative flexion/exten

sion radiographs demonstrating construct stability. In 1 patient a unilateral C1
lateral mass screw was placed after the contralateral lateral mass fractured dur
ing drilling. In this patient the polyaxial screw was connected with a rod to a C2
sublaminar hook for atlantoaxial fixation. On follow-up radiographs the screw
was noted to be disconnected from the rod, but there was no instability on
flexion-extension films and the patient was clinically improved.
dramroo
Table 1. Case representations
C1 Lateral Mass Fixation
Patient Age/sex Presentation Diagnosis Treatment Follow-up Outcome
1 16/F Neck pain, Irreducible type II odontoid Bilateral C1 LMS, C2 pars 6 months F/E films stable,
neurologically fracture screws, polyaxial screw/rod neurologically
intact construct; interspinous intact
cable/autograft
2 68/F Neck pain Old type II odontoid fracture, Bilateral C1 LMS, C2 pars 4 months F/E films stable,
nonunion, mobile on F/E films screws, polyaxial screw/rod clinically improved
construct; interspinous
cable/autograft
3 81/M Hand numbness, Transverse ligament cyst, C1 laminectomy, unilateral C1 3 months Rod
dramroo
myelopathy irreducible C1-C2 subluxation LMS, C2 laminar hook, disconnected

polyaxial screw/rod construct; from screw, F/E
lateral autograft arthrodesis films stable,
clinically improved
4 68/M Hand numbness, Transverse ligament cyst, Bilateral C1 LMS, C2 pars 9 months F/E films stable,
neck pain, irreducible C1-C2 subluxation screws, plate construct; symptoms
myelopathy interspinous cable/autograft improved
5 74/F Neck pain, hand Reducible C1-C2 subluxation, L C1-C2 TAS; R C1 LMS, C2 18 months Osseous fusion,
numbness, medial R VA pars screw, plate construct; F/E films stable,
myelopathy interspinous cable/autograft symptoms improved
6 46/F Hand numbness, Transverse ligament cyst, Transoral odontoidectomy; 3 months F/E films stable,
myelopathy cervicomedullary compression, posterior atlantoaxial fixation clinically stable
medial L VA, cervicothoracic with R Cl LMS, C2 pars screw,
kyphosis plate; interspinous
cable/autograft
133
Table 1 (continued)
Fiore/Birch/Haid
Patient Age/sex Presentation Diagnosis Treatment Follow-up Outcome
7 34/M Neck pain, Irreducible type III odontoid Bilateral C1, C3 LMS, C2 pars 9 months Osseous fusion,
neurologically fracture screws, plate construct; F/E films stable,
intact interspinous cable/autograft neurologically
intact
8 72/F Hand weakness, Congenitally narrow canal at C1 laminectomy; bilateral C1, 3 months F/E films stable,
myelopathy C1, cervicomedullary C3 LMS, plate construct; C2 clinically stable
compression, C1-C2 sublaminar cables; lateral
subluxation, bilateral medial autograft arthrodesis
VAs
dramroo
9 69/F Neck pain, Leiomyosarcoma metastasis to Tumor resection; bilateral C1, 16 months F/E films stable,
neurologically R C2 pars, medial L VA C3, C4 LMS, L C2 pars screw, neurologically
intact plate construct; lateral autograft intact, died from
arthrodesis diffuse
metastases
10 74/M R hemiparesis, Metastatic colon cancer, Occipitocervical fusion to C5 9 days Died from
neck pain pathologic C2 burst fracture, with polyaxial screw/rod pneumonia
spinal cord compression system, bilateral C1 LMS
LMS Lateral mass screw, TAS transarticular screw, F/E flexion/extension, VA vertebral artery.
134
Fig. 6. Patient 1. Preoperative lateral cervical radiograph demonstrating anteriorly dis
placed type II odontoid fracture.
Illustrative Case: Patient 1

A neurologically intact 16-year-old girl was referred for management of a
type II odontoid fracture. She had sustained the fracture 2 months earlier in a
motor vehicle accident and had been treated with external immobilization by
the referring physician. She was referred when follow-up radiographs showed a
1-cm subluxation at C1-C2 (fig. 6). After 3 days of halo traction, follow-up
radiographs demonstrated minimal reduction of the subluxation, and she was
taken to the OR for internal fixation. Intraoperative attempts to reduce the
fracture under fluoroscopy were unsuccessful. Internal fixation was achieved
using a polyaxial screw-rod system with C1 lateral mass screws and C2 pars
interarticularis screws, supplemented with interspinous iliac crest autograft and
sublaminar cable. The patient was mobilized postoperatively in a hard cervical
collar. Postoperative radiographs revealed solid fixation at C1-C2 (fig. 7).
The patient is neurologically intact at 6 months follow-up.
Discussion
We have described a technique to achieve solid fixation of the C1 lateral

mass that can be utilized in a variety of instrumentation constructs for varying

dramroo
Fig. 7. Patient 1. Postoperative AP and lateral radiographs demonstrating C1-C2
construct with polyaxial screw-rod system and interspinous cable and autograft.
indications. In our experience, the most common indication for the use of C1 lat
eral mass screws is atlantoaxial instability. Although a variety of techniques exist
to treat atlantoaxial instability, certain anatomic factors may preclude their appli
cation in specific situations. C1-C2 interspinous fusion techniques using either
sublaminar cables or interlaminar clamps in combination with iliac crest auto
graft require the presence of intact posterior elements. These techniques cannot
be applied when the C1 arch or C2 laminae have been disrupted by trauma, neo
plasm, or other pathologic processes, or when resection of these elements is nec
essary to achieve neural decompression. C1-C2 transarticular screw fixation is
likewise precluded by a variety of factors. In up to 20% of patients, a medially
located or high-riding vertebral artery will preclude safe passage of C1-C2
transarticular screws unilaterally. In 3% of patients, vertebral artery anatomy will
preclude passage of screws bilaterally [9, 10]. Irreducible C1-C2 subluxation will
likewise preclude optimal placement of C1-C2 transarticular screws. In this case,
a screw trajectory traversing the articular surfaces of C1 and C2 cannot be
achieved. Severe cervicothoracic kyphosis may preclude C1-C2 transarticular
screw placement by obstructing the trajectory of the instruments used to insert
the screws. Destruction or erosion of the osseous substrate for screw fixation by
trauma, neoplasm, or other pathologic processes will likewise preclude transar
ticular screw placement. In these situations occipitocervical fusion may be con
sidered as an alternative means to treat atlantoaxial instability. Occipitocervical
fusion may be avoided by using C1 lateral mass screws to achieve atlantocervi
cal fixation. By avoiding occipitocervical fixation, patients avoid the risk of
intracranial bleeding which may occur with placement of occipital hardware
[11]. Additionally, range of motion at the atlantooccipital joint is maintained,
dramroo
reducing morbidity from craniocervical malalignment that may occur following
occipitocervical fusion. Clinical studies have also suggested that avoidance of
occipitocervical fusion may decrease the incidence of delayed subaxial subluxa
tion [12, 13]. In this series we achieved atlantocervical fixation in 7 patients who
demonstrated various anatomic characteristics that precluded traditional methods
of atlantoaxial fixation. These patients were all mobilized in hard cervical collars,
avoiding postoperative halo vest immobilization.
It is likely that C1 lateral mass screws will also prove to be an extremely
useful technique for occipitocervical fixation. C1 lateral mass screws provide
additional fixation points for occipitocervical constructs, possibly increasing
resistance to construct failure. This additional construct integrity is achieved
without fusing additional cervical levels, thus preserving cervical motion
segments. In the one patient in this series who underwent occipitocervical
fixation, we were able to achieve solid C1 fixation and integration into the
occipitocervical construct using C1 lateral mass screws. As the use of polyax
ial screw-rod systems for occipitocervical fixation becomes more widespread,
we anticipate that C1 lateral mass screws will be used more frequently, since the
contourable rods used in these systems will allow C1 screws to be easily
incorporated into occipitocervical constructs.
In 6 of the 9 patients in this series who underwent atlantocervical fixation,
constructs using C1 lateral mass screws were supplemented with a posterior
C1-C2 fusion using sublaminar cable and interspinous autograft. In recent case
series, Dickman et al. [4] reported an 86% fusion rate after interspinous fusion
with cables and autograft, while Farey et al. [5] reported a 58% fusion rate. In two
smaller series, fusion rates of 100% were achieved using the Brooks method of
interspinous fusion, but all patients were immobilized in a halo vest for 3 months
[14, 15]. It is not clear from this series whether C1 lateral mass screw constructs
will increase fusion rates when applied in addition to interspinous fusion tech
niques, but it seems likely that the additional rigidity conferred by these constructs
should result in improved outcomes. The ability of unilateral C1 lateral mass
screw constructs to increase fusion rates when applied together with contralateral
C1-C2 transarticular screws is also unclear. Song et al. [16] reported a 95% fusion
rate after unilateral transarticular screw placement combined with posterior inter
spinous fusion in a group of patients with high-riding vertebral arteries. In the
present series, 1 patient had a unilateral C1 lateral mass screw construct in com
bination with a contralateral transarticular screw and interspinous fusion. Again,
it seems likely that the supplemental fixation provided by the C1 lateral mass
screw construct will increase rigidity and result in higher fusion rates.
It is also unclear whether C1 lateral mass screw constructs can be used as
a stand-alone method for achieving atlantocervical fusion in the absence of
interspinous fusion or unilateral transarticular screw fixation. Harms and

dramroo
Melcher [2] suggest that temporary C1-C2 constructs using C1 lateral mass
screws may be used in selected cases, including rotatory subluxation and young
patients with displaced odontoid fractures, allowing preservation of rotation at
C1-C2 after instrumentation removal. In addition, they state that C1-C2 fixa
tion with C1 lateral mass screws eliminates the morbidity associated with
passage of C1 sublaminar cables. In this series we treated 3 patients with dis
placed odontoid fractures, adding interspinous cable and autograft to increase
construct rigidity and provide additional substrate for bony fusion. We feel that
odontoid fractures with irreducible subluxation are best treated with C1-C2
interspinous arthrodesis in addition to instrumentation, to provide optimal rates
of long-term fixation. In our experience, passage of C1 sublaminar cables can
be performed with minimal morbidity when neural compression is not present.
In this series 3 patients were treated with stand-alone constructs because
absence of the C1 or C2 laminae precluded interspinous arthrodesis. Two
patients had stable radiographs at follow-up, and rod-screw separation occurred
in 1 patient with a unilateral construct, although there was no overt radiographic
instability. When stand-alone constructs are used, it is important to achieve
lateral arthrodesis by decorticating the lateral masses and C1-C2 joint space,
with placement of cancellous autograft laterally. In the future, the decision to
employ a C1 lateral mass screw construct without interspinous fusion or con
tralateral transarticular screw fixation should be considered on an individual
basis in the context of the pathologic process causing instability, bone quality,
and other comorbidities influencing bone fusion, as well as the potential mor
bidity of the alternative treatment, occipitocervical fusion. Larger studies with
long-term follow-up will be necessary to determine the safety and efficacy of
C1 lateral mass screw constructs.
To date, only one study has examined the biomechanical characteristics
of atlantoaxial constructs using C1 lateral mass screws. Lynch et al. [17]
evaluated an atlantoaxial construct with C1 lateral mass screws and C2
pedicle screws with and without supplemental interspinous cable and graft,
and compared this construct with atlantoaxial transarticular screws. The C1
lateral mass screw construct was most resistant to lateral bending and axial
rotation, and less resistant to flexion and extension. Adding a posterior cable
and graft reduced motion slightly. Compared to transarticular screws, the C1
lateral mass screw construct was slightly less rigid, allowing an average of
0.6 more motion. This study suggests that C1 lateral mass screw constructs
are a reasonable alternative to transarticular screws for achieving atlantoaxial
stabilization.
Harms and Melcher [2] used a specially modified screw at C1 with an
unthreaded proximal shaft, in order to reduce the risk of greater occipital nerve
irritation as well as screw breakage. In our series we used standard screws with
dramroo
threads along the entire shaft. We did not observe any cases of occipital
neuralgia or screw breakage. These results indicate that standard screws may be
used at C1. We believe the risk of greater occipital nerve irritation is small as
long as there is adequate space caudal to the C1 screw for passage of the nerve.
In 1 patient not included in this series, placement of C1 lateral mass screws was
planned. However, intraoperatively the C2 nerve roots were found to be much
larger than usual, occupying the entire space between the inferior aspect of the
C1 dorsal arch and the superior aspect of the C2 pars. In this case, we elected
not to place C1 screws, since the risk of C2 irritation was unacceptably high.
In this case the alternative would be to place C1 screws directly into the dorsal
aspect of the C1 arch, rather than into the inferior surface of the arch. This
would place the screw shafts away from the C2 nerve roots. However, this can
only be considered when the rostrocaudal dimension of the C1 arch is large
enough to accommodate a screw. In addition, the vertebral artery in the C1
groove must be retracted rostrally during drilling, increasing the risk of embolic
complications or direct injury to the vertebral artery by the drill, retractor, or
other instruments.
The risk of vertebral artery injury must always be assessed when place
ment of lateral mass screws or transarticular screws is planned. In this small
case series, there were no vertebral artery injuries. To minimize this risk,
preoperative assessment of the path of the vertebral artery using CT scanning
is mandatory prior to placement of C1 lateral mass screws. Magnetic resonance
angiography or catheter angiography may be performed to provide additional
information concerning the path and patency of the vertebral arteries, although
in our experience we have not found this to be necessary. The surgeon must note
that the trajectory of the C1 lateral mass screw is very different from that of
lateral mass screws placed in the subaxial cervical spine. Particularly important
is that the C1 screw is placed with a slight medial angulation to avoid the
vertebral artery laterally and the spinal canal medially. We consider the use of
intraoperative fluoroscopy or CT-based image guidance mandatory to safely
place C1 lateral mass screws. Fluoroscopy allows safe placement of bicortical
screws under direct visualization, while CT-based image guidance provides
additional three-dimensional information about the vertebral artery and spinal
canal. Virtual fluoroscopy may also prove to be a useful adjunct to screw
placement.
Conclusion
The placement of C1 lateral mass screws provides a useful alternative

method to achieve atlantocervical fixation when anatomic factors preclude the

dramroo
placement of atlantoaxial transarticular screws. This method achieves immediate
rigid stabilization of the atlantoaxial joint and obviates the need for halo vest
immobilization. This technique may be used in certain cases as an alternative to
occipitocervical fusion, and may also be used to increase construct stability
when occipitocervical fixation is employed. Evaluation of the course of the
vertebral artery with preoperative CT scanning and use of intraoperative fluo
roscopy or image guidance are mandatory when using this technique.
Placement of C1 lateral mass screws is a technically demanding procedure that
may result in grave complications from vertebral artery injury if improperly
performed. We thus advocate that this procedure only be performed by surgeons
who are highly experienced in the treatment of atlantoaxial instability, and who
have an intimate understanding of the anatomy of the region. The uninitiated
surgeon can minimize the possibility of complications during C1 lateral mass
screw placement by first performing this procedure in a cadaveric setting.
Further biomechanical analysis of this technique should be performed to quan
tify the strength of constructs employing C1 lateral mass screws as compared
with other fixation methods. Further clinical studies should be performed to
determine the safety and efficacy of this technique.
References
1 Fiore AJ, Haid RW, Rodts GE, Subach BR, Mummaneni PV, Riedel CJ, Birch BD: Atlantal lateral
mass screws for posterior spinal reconstruction: Technical note and case series. Neurosurg Focus
2002:12/1:article 5.
2 Harms J, Melcher RP: Posterior C1-C2 fusion with polyaxial screw and rod fixation. Spine
2001;26:24672471.
3 Stokes JK, Villavicencio AT, Liu PC, Bray RS, Johnson JP: Posterior atlantoaxial stabilization:
New alternative to C12 transarticular screws. Neurosurg Focus 2002;12/1:article 6.
4 Dickman CA, Sonntag VK, Papadopoulos SM, Hadley MN: The interspinous method of posterior
atlantoaxial arthrodesis. J Neurosurg 1991;74:190198.
5 Farey ID, Nadkarni S, Smith N: Modified Gallie technique versus transarticular screw fixation in
C1-C2 fusion. Clin Orthop 1999;35:126135.
6 Dickman CA, Sonntag VK: Posterior C1-C2 transarticular screw fixation for atlantoaxial
arthrodesis. Neurosurgery 1998;43:275281.
7 Haid RW, Subach BR, McLaughlin MR, Rodts GE, Wahlig JB: C1-C2 transarticular screw
fixation for atlantoaxial instability: A 6-year experience. Neurosurgery 2001;49:6570.
8 Moskovich R, Crockard HA: Atlantoaxial arthrodesis using interlaminar clamps. Spine
1992;17:261267.
9 Madawi AA, Casey AT, Solanki GA, Tuite G, Veres R, Crockard HA: Radiological and anatomi
cal evaluation of the atlantoaxial transarticular screw fixation technique. J Neurosurg
1997;86:961968.
10 Paramore CG, Dickman CA, et al: The anatomic suitability of the C1-2 complex for transarticular
screw fixation. J Neurosurg 1996;85:221224.
11 Vale FL, Oliver M, Cahill DW: Rigid occipitocervical fusion. J Neurosurg 1997;91(suppl 2):144150.
12 Clark CR, Goetz DD, Menezes AH: Arthrodesis of the cervical spine in rheumatoid arthritis.
dramroo
13 Kraus DR, Peppelman WC, Agarwal AK, DeLeeuw HW, Donaldson WF: Incidence of subaxial
subluxation in patients with generalized rheumatoid arthritis who had previous occipital cervical
fusions. Spine 1991;16(suppl):486489.
14 Boden SD, Dodge LD, Bohlman HH, Rechtine GR: Rheumatoid arthritis of the cervical spine.
A long-term analysis with predictors of paralysis and recovery. J Bone Joint Surg Am
1993;75:12821297.
15 McCarron RF, Robertson WW: Brooks fusion for atlantoaxial instability in rheumatoid arthritis.
South Med J 1988;81:474476.
16 Song GS, Theodore N, et al: Unilateral posterior atlantoaxial transarticular screw fixation.
J Neurosurg 1997;87:851855.
17 Lynch JJ, Crawford NR, Chamberlain RH, Bartolomei JC, Sonntag VK: Biomechanics of lateral
mass/pedicle screw fixation at C1-2. Proceedings of the 2002 Annual Meeting of the American
Association of Neurological Surgeons, Chicago, 2002.

Department of Neurosurgery, Emory University
1365B Clifton Road, NE, Ste. 6400, Atlanta, GA 30322 (USA)
Tel. 1 404 778 5770, Fax 1 404 778 4472, E-mail [email protected]

dramroo
Cervical Laminoplasty
Adrian T.H. Caseya,b, H. Alan Crockarda
a
Victor Horsley Department of Neurosurgery, National Hospital for Neurology
and Neurosurgery, London, and
b
Spinal Surgery Unit, The Royal National Orthopaedic Hospital, Stanmore, UK
Cervical laminoplasty is very much a Japanese invention. Its popularity in

Japan arises from the formidable challenges of anterior decompression for ossi
fication of the posterior longitudinal ligament [15]. These anterior multilevel
surgeries would be frequently complicated by dural tears as the dura is usually
intimately associated with the ossified ligament [1, 69]. There was also a sig
nificant risk of instrumentation or graft failure [2]. It has been reported that the
rate of these complications including cerebrospinal fluid leakage and dislodg
ment or pseudarthrosis of the strut grafted bone was 24% and the rate of the
salvage operation required was 12.5% [10].
In the past, laminectomy has been the most common method to achieve pos
terior decompression of the cervical spine in these patients. However, the proce
dure has been complicated by postoperative instability resulting in deformity,
particularly kyphosis [1115], which may exacerbate neurological symptoms.
Kyphosis and instability may leave the spine more vulnerable to cervical
spine trauma, especially flexion injuries. In addition, postlaminectomy mem
branes have been implicated in arachnoiditis and restenosis after simple laminec
tomy [16, 17]. To avoid the disadvantages of laminectomy, several authors have
described the technique of cervical laminoplasty whereby decompression is
achieved without removal of the posterior spinal elements, maintaining the bio
mechanical integrity of the cervical spine and the spinal cord-protective fea
tures of the posterior elements [3, 1825]. This is a more physiological solution.
Expansive open-door laminoplasty was first described by Hirabayashi et al.
[26] as a development of the air drill laminectomy technique of Kirita [27] and
has since been modified by Hirabayashi et al. and many others [3, 1825, 28].
The authors of this chapter present a modification of the technique using
titanium miniplates to stabilize the posterior elements in the open position [22].
dramroo
Table 1. Indications for expansive cervical laminoplasty
OPLL over multiple levels (with maintained cervical lordosis)
Congenital canal stenosis (with maintained cervical lordosis)
Multilevel cervical spondylosis (with maintained cervical lordosis)
Posterior compression from ligamentous hypertrophy (with maintained cervical lordosis)
As part of a staged anterior and posterior canal expanding procedure
OPLL Ossification of the posterior longitudinal ligament.
Table 2. Contraindications for expansive cervical laminoplasty
Significant anterior compression

Established absolute kyphosis
Isolated radiculopathy
Loss of anterior column support resulting from tumour, trauma or infection
This method has the advantage of technical simplicity and allows for postoper
ative magnetic resonance imaging because the use of stainless steel implants is
avoided. It has been used in the authors departments in more than 300 cases
now, with good short- and long-term results. The technique is no longer limited
to ossification of the posterior longitudinal ligament, which is rare in
Caucasians. Its most frequent use is now for multilevel cervical spondylotic
myelopathy associated with varying degrees of constitutional canal stenosis.
The indications and contraindications are listed in tables 1 and 2.
Kyphosis is a contraindication to laminectomy [29] and probably lamino
plasty [30, 31].
There are several technical variations. The two main differences are
whether the opening is in the midline (French door) or to one side (open door).
Then there are differences in how to keep the door open. Initially the techniques
were quite cumbersome and involved suturing or wiring the bony posterior
elements to muscle. Inevitably they did not always keep the door as open as it
may have been originally at the time of surgery. The Queen Square technique
uses titanium miniplates with no bone graft to keep the door open. In our expe
rience this has been a simple and reliable technique. Alternative techniques
have used ceramic spacers or bone graft to keep the doors open.
Three types of Z plasty may be performed: on the lamina using the method
of Hattori as reported by Oyama et al. [32], between each lamina using the
reciprocal method of Tomimura and Morizono [33] and between two segmen
tal laminae using the Chiba University technique [34]. The variety of reported
techniques are illustrated in figure 1.
Laminoplasty 143
dramroo
Itoh 1984 Koyama 1985 Lin 1986
Hirabayashi 1987 N. Nakano 1988
Hattori 1989
Tomita 1988 Shikata 1989
Yoshida 1992 K. Nakano 1992
Fig. 1. Different techniques reported.
Casey/Crockard 144
dramroo
Tomita et al. have [3] championed the use of their thread wire saw to
achieve the opening. This is similar to a very fine Gigli saw. It was developed
for en bloc vertebrectomy. This is a clever technical innovation, however it does
require sublaminar passage of the wire. This is very much finer than sublami
nar wires and indeed cables which have been implicated in neurological deteri
oration. The majority of the other techniques require a high-speed drill.
Queen Square Technique [22]
Surgical technique (C3C6): The patient is placed in a prone position with

the head slightly flexed secured by Mayfield pins. The body is supported by a
Montreal mattress. Somatosensory evoked potentials are recorded. A standard
posterior midline approach allows exposure of the cervical laminas from the
caudal edge of C2 to the cranial edge of C7 and laterally to the lateral aspect of
the facet joints. Care is taken to preserve the facet capsules and soft tissue
attachments to the lateral masses. The spinous processes and interspinous
ligaments are preserved.
Using a 3-mm spherical cutting burr, two channels are drilled bilaterally.
These are placed at the junction of the lamina and the medial aspect of the lateral
masses (fig. 2a). Care is taken to avoid damage to the facet joints. Using this
technique, the canal is entered laterally, where spinal cord compression is less
severe. Inadvertent plunging penetration of the canal is prevented by the shelf
created in the medial aspect of the lateral masses. The depth of the channels is
increased until the ventral cortex of the lamina is identified. The laminoplasty is
opened on the side with clinical evidence of unilateral radiculopathy or with asym
metric canal or foraminal stenosis identified on the preoperative scans. Specific
radiculopathies can also be addressed with foraminal decompressions. If there is
no unilateral radiculopathy or asymmetric stenosis, either side may be opened.
On the side to be opened, the channel is completed through the ventral cortex
of the lamina with a diamond burr or 1-mm Kerrison up-cutting rongeur.
Alternatively, the craniotome attachment for the Midas Rex (Midas Rex
Pneumatic Tools, Fort Worth, Tex., USA) may be used to transect the lamina on
the opening side. This is in fact our most usual technique. There have been no
dural tears or neurological deteriorations in our series, using the craniotome
attachment to open the door. On the hinge side, the ventral cortex of the lamina
is thinned until the posterior elements can be rotated dorsally to effect decom
pression of the spinal cord (fig. 2b). This is done en bloc because the interspinous
ligament and ligamentum flavum are intact from the occiput through C7.
Hirabayashi et al. [26] suggest a limited osteotomy of the spinous process
of C7 (or most caudal vertebra) to allow a sufficient opening of the door.
Laminoplasty 145
dramroo
a b c
Fig. 2. Operative sequence showing preparation of the bilateral channels for the hinge
side (right) and opening side (left) of the laminoplasty (a), mobilization and rotation of the
posterior elements (b), and stabilization of the open-door laminoplasty with titanium
miniplates (c) [modified from 22].
Release of ligamentum flavum, dural adhesions, and bridging vessels is per

formed on the opening side as required to elevate the posterior elements.
At C2C3 and C6C7, partial resection of the intraspinous ligaments on the
opening side and partial subperiosteal release of the intraspinous ligaments of
C2 and C7 on the hinge side may be necessary to allow adequate mobilization
and rotation of the posterior elements, achieving full decompression (fig. 2b).
Vigorous epidural bleeding is occasionally encountered on release of the
stenotic canal. This can be controlled with bipolar coagulation and haemostatic
material. After adequate decompression, the dura is typically pulsatile, indicat
ing satisfactory decompression. When the open door has been adequately mobi
lized, it is stabilized with titanium miniplates (Lehbinger, Freiburg, Germany).
Hirabayashi et al. [35] unflex the neck prior to fixing their laminoplasty.
This they claim helps maintain and secure cervical lordosis. We most often use
a straight five-hole plate bent into sigmoid shape/open Z configuration (fig. 2c)
to allow one hole for fixation into the lamina and 2 screws into the corre
sponding lateral mass. Longer or shorter plates can be fashioned as needed to
achieve adequate opening. The plates are held in position with 2 or 3 screws
59 mm in length. An attempt is made to angle the screws away from the facet
joints. We typically use one plate per level, if possible, to distribute the forces
over multiple fixation points (fig. 3). No formal attempt is made to graft bone
to the operative site. However, on the hinge side, where the adjacent lamina
and lateral masses are still connected, the dorsal cortex of these structures are
Casey/Crockard 146
dramroo
Fig. 3. Schematic diagram showing the orientation of the titanium miniplates in a
C3C6 laminoplasty and adjacent anatomic structures.
brought into contact as the hinge side trough is closed during elevation and
rotation of the posterior elements into the open position. This is analogous to a
green stick fracture. The morbidity associated with harvesting bone graft is
avoided. A soft collar is used to facilitate patient comfort while mobilizing
during the first few days after surgery. No other external orthosis is used after
surgery.
Discussion
Anterior and posterior decompression are established techniques in the man

agement of multilevel cervical canal stenosis resulting in myeloradiculopathy.
Anterior cervical discectomy and fusion immobilize segments of the cervical spine
and result in high mechanical demands on the adjacent intervertebral segments.
Over the patients lifetime, abnormal mechanical stresses can produce significant
radiographic and clinical evidence of deterioration. Hilibrand et al. [36] in a con
secutive series of 374 patients reported that symptomatic adjacent-segment disease
occurred at a relatively constant incidence of 2.9%/year (range 0.04.8%/year)
during the 10 years after the operation. Survivorship analysis predicted that 25.6%
of the patients (95% confidence interval, 2032%) who had an anterior cervical
arthrodesis would have new disease at an adjacent level within 10 years after the
operation. However, contrary to their hypothesis, they found that the risk of new
disease at an adjacent level was significantly lower following a multilevel
arthrodesis than it was following a single-level arthrodesis (p 0.001).
Laminoplasty 147
dramroo
Nonetheless there is an intellectual reluctance to perform a multilevel
trench procedure with fusion in a young patient. The incidence of complications
for anterior multilevel corpectomy has been reported to be high even in expert
hands. Saunders et al. [37] did a retrospective analysis of 31 cases of cervical
spondylotic myelopathy treated by four-level subaxial cervical corpectomy.
Three patients died within 3 weeks of surgery (9.7%). Delayed radiculopathy
occurred in 4 patients after surgery, 3 had acute graft complications, and 1 had
pseudomeningocoele, resulting in a morbidity rate of 25.8%. There were no
cases of infection or increasing myelopathy. In another series by these authors,
on 40 cases of cervical corpectomy, they reported a perioperative complication
rate of 47.5%, with a 7.5% incidence of persistent sequelae [3739]. Fessler
et al. [40] have reported their extensive experience of cervical corpectomy in a
retrospective series of 93 cases over 10 years with a lower complication rate.
In another more recent series Edwards et al. [41] have attempted to compare the
neurological outcome and complications of cervical corpectomy and cervical
laminoplasty. Medical records of all patients treated for multilevel cervical
myelopathy with either multilevel corpectomy or laminoplasty between 1994
and 1999 at the Emory Spine Center were reviewed. From a pool of 38 patients
meeting stringent inclusion and exclusion criteria, 13 patients who underwent
multilevel corpectomy were blindly matched with 13 patients who underwent
laminoplasty based on known prognostic criteria. Improvement in function
averaged 1.6 Nurick grades after laminoplasty and 0.9 grades after multilevel
corpectomy (p 0.05). Subjective improvements in strength, dexterity, sensa
tion, pain, and gait were similar for the two operations. The prevalence of axial
discomfort at the latest follow-up was the same for each cohort, but the anal
gesic requirements tended to be greater for patients who underwent multilevel
corpectomy. This is the opposite finding to the larger study of Hosono et al. [42]
on pain. Sagittal motion from C2 to C7 decreased by 57% after multilevel cor
pectomy and by 38% after laminoplasty. One complication (C6C7 herniated
nucleus pulposus requiring anterior discectomy with fusion) occurred in the
laminoplasty group [41]. Multilevel corpectomy complications included the
progression of myelopathy, non-union, persistent dysphagia, persistent dyspho
nia, and subjacent motion segment ankylosis.
An alternative posterior decompressive procedure to laminoplasty is cervi
cal laminectomy with posterior fixation/fusion. Today this would be performed
using the lateral mass screw/rod system. This will also prevent postoperative
kyphus. It will more reliably immobilize the spine. Adams and Logue [43, 44]
have shown that one of the reasons for delayed deterioration following cervical
laminectomy is hypermobility. Screw/rod fixation is therefore an attractive
option. However the extra rigidity is probably not required as Herkowitz
[45, 46] concluded from his biomechanical study that stability of the cervical
Casey/Crockard 148
dramroo
spine after laminoplasty was not significantly different from that of the intact
control. Laminoplasty affords some rigidity, but still preserves motion [47]. The
long-term effects of laminoplasty on cervical movement and alignment were
investigated by radiography and CT scans in a study of 56 patients with multi-
segmental myelopathy who had undergone a C3 to C7 open-door laminoplasty.
Follow-up averaged 5.8 years. Satisfactory neurological improvement occurred
in 73%. Cervical flexion decreased by 35% and extension by 57%; the decrease
of both movements was statistically significant. Decreased vertebral slip as well
as slightly reduced lordosis was seen after operation.
Cervical laminectomy and laminoplasty have been compared in a retrospec
tive study [48]. This is another study from Emory, with similar design principles
to the corpectomy/laminoplasty study described above (an independent matched
cohort analysis), involving 13 patients in each arm. Both objective improvement in
patient function (Nurick score) and the number of patients reporting subjective
improvement in strength, dexterity, sensation, pain, and gait tended to be greater in
the laminoplasty cohort [48]. Whereas no complications occurred in the lamino
plasty cohort, there were 14 complications in 9 patients that underwent laminec
tomy with fusion. Complications included progression of myelopathy, non-union,
instrumentation failure, development of a significant kyphotic alignment, persis
tent bone graft harvest site pain, subjacent degeneration requiring reoperation, and
deep infection. The marked difference in complications and functional improve
ment between these matched cohorts suggests that laminoplasty may be preferable
to laminectomy with fusion, as a posterior procedure for multilevel cervical
myelopathy [48]. Similar findings were found in a less well-controlled study by
Herkowitz [49]. In this retrospective study he compared the results and complica
tions of 45 patients with at least a 2-year follow-up, who had undergone anterior
fusion, cervical laminectomy, or cervical laminoplasty for the surgical manage
ment of multiple level cervical radiculopathy due to cervical spondylosis. Eighteen
patients (58 levels) underwent anterior fusion, 12 patients (38 levels) had a cervi
cal laminectomy, and 15 patients (57 levels) underwent a cervical laminoplasty. In
another study by Baisden et al. [50], radiographic and biomechanical results in the
goat model revealed that laminoplasty was superior to laminectomy in maintain
ing cervical alignment and preventing postoperative spinal deformities. Quite how
this relates to humans is debatable.
Complications
There are two notable complications with cervical laminoplasty radicu

lopathy and pain. These have been the subject of several studies [51]. C5,C6
radicular pain and/or paresis are the most frequent complications that occur in
Laminoplasty 149
dramroo
approximately 510% of the patients in the series of Hirabayashi et al. [35]
of 350 patients although most complications resolve spontaneously within 2
years. Of 365 patients who had undergone laminoplasty, 20 patients (5.5%)
developed postoperative radiculopathy. Using data from postoperative com
puted tomography scans and other sources, these patients were compared with
211 patients with no radiculopathy, who had undergone laminoplasty during the
same period, to identify risk factors related to patient characteristics and surgi
cal techniques. Of various risk factors studied, the narrowest level of the spinal
canal, preoperative symptomatic severity, flatness of the spinal cord assessed by
computed tomography myelopathy at C4C5, cervical curvature, anterior pro
trusion of the superior articular process as assessed by computed tomography
scan, laterality of the osteophytes, and ossification of the posterior longitudinal
ligament could not significantly discriminate between patients with and with
out postoperative radiculopathy. The angle of lamina as measured by using
computed tomography scans obtained after expansion in the patients with
radiculopathy was greater on both the opened and hinged sides and was signif
icantly greater than the angle in patients without radiculopathy (p 0.05). The
incidence of radiculopathy on both the opened and hinged sides was signifi
cantly higher in patients in whom the bony gutter had been cut on the lateral
side of the medial aspect of the zygapophyseal joint. An alternative theory is
that too radical a decompression allows slumping backwards of the spinal cord,
putting traction on the anatomically vulnerable C5 nerve root. This is the theory
espoused for C5 radiculopathy following the trench corpectomy procedure.
Here it now recommended limiting the decompression to 15 mm from right to
left. Friction heat generated by drilling on the open and hinge sides, traumatic
use of Kerrison rongeurs, and a drop of the hinge into the canal are also by
some considered as causes of such injury. However, these types of trauma dur
ing the operation are likely to damage the posterior root rather than the anterior
root; therefore, the sensory disturbance is expected to be stronger. Nevertheless,
in most cases, sensory disturbance at C5 or C6 was absent. This lends credence
to the tethering action on the anterior root [35].
Pain
There is an increasing recognition that this is quite a painful procedure in
the short term, mainly with pain in the trapezius region. In the long run a fair
number of patients experience neck pain. This has been studied by Hosono
et al. [42]. Ninety-eight patients had surgery for their disability secondary to
cervical spondylotic myelopathy. Of those patients, 72 had laminoplasty and 26
had anterior interbody fusion. The presence or absence of axial symptoms was
investigated before and after surgery. The duration, severity, and laterality of
symptoms were also recorded. The prevalence of postoperative axial symptoms
Casey/Crockard 150
dramroo
was significantly higher after laminoplasty than after anterior fusion (60 vs.
19%; p 0.05). In 18 patients (25%) from the laminoplasty group, the chief
complaints after surgery were related to axial symptoms for more than 3 months,
whereas in the anterior fusion group, no patient reported having such severe
pain after surgery. In this group shoulder pain developed exclusively on the
hinged side.
In conclusion, open door expansive laminoplasty is a versatile, easy and
effective method for achieving multilevel decompression of the cervical spine
affected by ossification of the posterior longitudinal ligament or cervical
spondylosis [51].
References
1 Banerji D, Acharya R, Behari S, Chhabra DK, Jain VK: Corpectomy for multi-level cervical
spondylosis and ossification of the posterior longitudinal ligament. Neurosurg Rev 1997;20/1:
2531.
2 Epstein NE: Reoperation rates for acute graft extrusion and pseudarthrosis after one-level anterior
corpectomy and fusion with and without plate instrumentation: Etiology and corrective manage
ment. Surg Neurol 2001;56/2:7381.
3 Tomita K, Kawahara N, Toribatake Y, Heller JG: Expansive midline T-saw laminoplasty (modified
spinous process-splitting) for the management of cervical myelopathy. Spine 1998;23/1:3237.
4 Nakano N, Nakano T: Clinical results following enlargement of the cervical spinal canal by means
of laminoplasty. Nippon Seikeigeka Gakkai Zasshi 1988;62:11391147.
5 Morimoto T, Uranishi R, Nakase H, Kawaguchi S, Hoshido T, Sakaki T: Extensive cervical
laminoplasty for patients with long segment OPLL in the cervical spine: An alternative to the ante
rior approach. J Clin Neurosci 2000;7/3:217222.
6 Kojima T, Waga S, Kubo Y, Kanamaru K, Shimosaka S, Shimizu T: Anterior cervical vertebrec
tomy and interbody fusion for multi-level spondylosis and ossification of the posterior longitudi
nal ligament. Neurosurgery 1989;24:864872.
7 Onari K, Akiyama N, Kondo S, Toguchi A, Mihara H, Tsuchiya T: Long-term follow-up results of
anterior interbody fusion applied for cervical myelopathy due to ossification of the posterior
longitudinal ligament. Spine 2001;26:488493.
8 Sridhar K, Ramamurthi R, Vasudevan MC, Ramamurthi B: Surgery for ossification of the poste
rior longitudinal ligament of the cervical spine. Neurol India 2001;49/2:116123.
9 Epstein NE: Identification of ossification of the posterior longitudinal ligament extending through
the dura on preoperative computed tomographic examinations of the cervical spine. Spine 2001;
26/2:182186.
10 Shinomiya K, Okamoto A, Kamikozuru M, Furuya K, Yamaura I: An analysis of failures in pri
mary cervical anterior spinal cord decompression and fusion. J Spinal Disord 1993;6/4:277288.
11 Callahan RA, Johnson RM, Margolis RN, Keggi KJ, Albright JA, Southwick WO: Cervical facet
fusion for control of instability following laminectomy. J Bone Joint Surg Am 1977;59:9911002.
12 Fager CA: Laminectomy and kyphotic deformity. J Neurosurg 2001;95(suppl 1):157158.
13 Herman JM, Sonntag VK: Cervical corpectomy and plate fixation for postlaminectomy kyphosis.
J Neurosurg 1994;80:963970.
14 Zdeblick TA, Bohlman HH: Cervical kyphosis and myelopathy. Treatment by anterior corpectomy
and strut-grafting. J Bone Joint Surg Am 1989;71/2:170182.
15 Yonenobu K, Hosono N, Iwasaki M, Asano M, Ono K: Neurologic complications of surgery for
cervical compression myelopathy. Spine 1991;16:12771282.
16 Skowronski J, Bielecki M: The results of laminectomy and laminoplasty in cervical myeloradicu
lopathy. Rocz Akad Med Bialymst 1992;37:7173.
Laminoplasty 151
dramroo
17 Yonenobu K, Wada E, Suzuki S, Kanazawa A: The dorsal approach in degeneratively changed cer
vical spine. Orthopde 1996;25:533541.
18 Kawai S, Sunago K, Doi K, Saika M, Taguchi T: Cervical laminoplasty (Hattoris method).
Procedure and follow-up results. Spine 1988;13:12451250.
19 Lee TT, Manzano GR, Green BA: Modified open-door cervical expansive laminoplasty for spondy
lotic myelopathy: Operative technique, outcome, and predictors for gait improvement. J Neurosurg
1997;86/1:6468.
20 Ishibashi K: Expansive laminoplasty by sagittal splitting of the spinous process for cervical
myelopathy: Correlation of clinical results with morphological changes in the cervical spine.
Kurume Med J 2000;47/2:135145.
21 Morimoto T, Yamada T, Okumura Y, Kakizaki T, Kawaguchi S, Hiramatsu K, et al: Expanding
laminoplasty for cervical myelopathy-spinous process roofing technique. Acta Neurochir 1996;
138:720725.
22 OBrien MF, Peterson D, Casey AT, Crockard HA: A novel technique for laminoplasty augmenta
tion of spinal canal area using titanium miniplate stabilization. A computerized morphometric
analysis. Spine 1996;21:474484.
23 Gillett GR, Erasmus AM, Lind CR: CG-clip expansive open-door laminoplasty: A technical note.
Br J Neurosurg 1999;13:405408.
24 Hirabayashi K, Bohlman HH: Multilevel cervical spondylosis. Laminoplasty versus anterior
decompression. Spine 1995;20:17321734.
25 Wang JM, Roh KJ, Kim DJ, Kim DW: A new method of stabilising the elevated laminae in open-
door laminoplasty using an anchor system. J Bone Joint Surg Br 1998;80:10051008.
26 Hirabayashi K: Expansive open-door laminoplasty for cervical spondylotic myelopathy. Jpn J
Surg 1978;32:11591163.
27 Kirita Y: Posterior decompression for cervical spondylosis and ossification of the posterior longi
tudinal ligament. Shujutsu 1976;30:287302.
28 Nakano K, Harata S, Suetsuna F, Araki T, Itoh J: Spinous process-splitting laminoplasty using
hydroxyapatite spinous process spacer. Spine 1992;17(suppl 3):S41S43.
29 Batzdorf U, Batzdorff A: Analysis of cervical spine curvature in patients with cervical spondylo
sis. Neurosurgery 1988;22:827836.
30 Matsunaga S, Sakou T, Nakanisi K: Analysis of the cervical spine alignment following lamino
plasty and laminectomy. Spinal Cord 1999;37/1:2024.
31 Fujimura Y, Nishi Y: Atrophy of the nuchal muscle and change in cervical curvature after expan
sive open-door laminoplasty. Arch Orthop Trauma Surg 1996;115/34:203205.
32 Oyama M, Hattori S, Noriwaki N: A new method of cervical laminectomy. Chubu Nippon
Seikeigeka Gakkai Zasshi 1984;16:792794.
33 Tomimura KS, Morizono Y: Osteoplastic enlargement of the cervical canal for cervical spondylo
sis. Rinsho Seikeigeka 1984;19:491497.
34 Watanabe TI, Goto S: Expansive laminoplasty (Chiba modification) for cervical myelopathy.
Shujutsu 1987;41:519527.
35 Hirabayashi K, Toyama Y, Chiba K: Expansive laminoplasty for myelopathy in ossification of the
longitudinal ligament. Clin Orthop 1999;359:3548.
36 Hilibrand AS, Carlson GD, Palumbo MA, Jones PK, Bohlman HH: Radiculopathy and myelopa
thy at segments adjacent to the site of a previous anterior cervical arthrodesis. J Bone Joint Surg
Am 1999;81:519528.
37 Saunders RL, Pikus HJ, Ball P: Four-level cervical corpectomy. Spine 1998;23:24552461.
38 Saunders RL, Bernini PM, Shirreffs TG Jr, Reeves AG: Central corpectomy for cervical spondy
lotic myelopathy: A consecutive series with long-term follow-up evaluation. J Neurosurg 1991;
74/2:163170.
39 Saunders RL: Anterior reconstructive procedures in cervical spondylotic myelopathy. Clin
Neurosurg 1991;37:682721.
40 Fessler RG, Steck JC, Giovanini MA: Anterior cervical corpectomy for cervical spondylotic
myelopathy. Neurosurgery 1998;43/2:257265.
41 Edwards CC 2nd, Heller JG, Murakami H: Corpectomy versus laminoplasty for multilevel cervi
cal myelopathy: An independent matched cohort analysis. Spine 2002;27/11:11681175.
Casey/Crockard 152
dramroo
42 Hosono N, Yonenobu K, Ono K: Neck and shoulder pain after laminoplasty. A noticeable compli
cation. Spine 1996;21:19691973.
43 Adams CB, Logue V: Studies in cervical spondylotic myelopathy. II. The movement and contour
of the spine in relation to the neural complications of cervical spondylosis. Brain 1971;94:
568586.
44 Adams CB, Logue V: Studies in cervical spondylotic myelopathy. III. Some functional effects of
operations for cervical spondylotic myelopathy. Brain 1971;94:587594.
45 Herkowitz HN: The surgical management of cervical spondylotic radiculopathy and myelopathy.
Clin Orthop 1989/239:94108.
46 Herkowitz HN: Cervical laminoplasty: Its role in the treatment of cervical radiculopathy. J Spinal
Disord 1988;1/3:179188.
47 Baba H, Maezawa Y, Furusawa N, Imura S, Tomita K: Flexibility and alignment of the cervical
spine after laminoplasty for spondylotic myelopathy. A radiographic study. Int Orthop 1995;19/2:
116121.
48 Heller JG, Edwards CC 2nd, Murakami H, Rodts GE: Laminoplasty versus laminectomy and
fusion for multilevel cervical myelopathy: An independent matched cohort analysis. Spine 2001;
26:13301336.
49 Herkowitz HN: A comparison of anterior cervical fusion, cervical laminectomy, and cervical
laminoplasty for the surgical management of multiple level spondylotic radiculopathy. Spine
1988;13:774780.
50 Baisden J, Voo LM, Cusick JF, Pintar FA, Yoganandan N: Evaluation of cervical laminectomy and
laminoplasty. A longitudinal study in the goat model. Spine 1999;24:12831289.
51 Uematsu Y, Tokuhashi Y, Matsuzaki H: Radiculopathy after laminoplasty of the cervical spine.
Spine 1998;23:20572062.
Professor Alan Crockard, DSc FRCS, FRCS (Ed)
Department of Surgical Neurology
National Hospital for Neurology and Neurosurgery
Box 3, Queen Square, London WC1N 3BG (UK)
Tel. 44 20 7829 8714, Fax 44 20 7813 1138
E-Mail [email protected], [email protected]
Laminoplasty 153
dramroo
Cervical Pedicle Screws: Advances
in Spinal Stabilization
Jay Y. Chuna, Praveen V. Mummanenib, Barry D. Birchc,

Rick C. Sassod
a
Department of Neurosurgery, University of California, San Francisco, Calif.,
b
Department of Neurosurgery, Emory University, Atlanta, Ga.,
c
Department of Neurosurgery, The Mayo Clinic, Scottsdale, Ariz., and
d
Indiana Spine Group, Indianapolis, Ind., USA
Since Hadra [8] first used silver wires to internally fixate the cervical
spine in 1891, cervical spinal fixation has undergone significant transforma
tion. Posterior instrumentation systems have evolved from wires to facet
screws, lateral mass plates, and ultimately to cervical pedicle screws. These
advances in cervical stabilization techniques have been accompanied by inno
vative spine image guidance systems to assist with appropriate placement.
Transpedicular fixation of the cervical spine poses a particular challenge
to surgeons due to the close proximity of the cervical pedicle to the spinal cord,
nerve roots and the vertebral arteries. Nevertheless, possible biomechanical
advantages afforded by cervical pedicle screws over lateral mass screws and
posterior cervical wires are the primary factor in the continuing study and
advancement in pedicle screw application for cervical spine fixation.
Anatomy
Due to the small margin of error allowed for ideal cervical pedicle screw
placement, even small miscalculations can result in a vascular or neural injury.
Therefore, intimate knowledge of the cervical pedicles is important.
Pedicle Measurement Studies

It should be noted that some surgeons feel that C2 does not have a true pedi
cle but rather has a large elongated pars; we disagree with this. Panjabi et al. [17]
dramroo
Table 1. Pedicle dimensions
C2 C3 C4 C5 C6 C7 T1
Panjabi et al. [17], 1991

Width, mm 7.7 5.8 5.7 6.1 6.3 6.6
Height, mm 9.4 7.6 7.4 6.7 7.1 7.5
Cross-sectional area, mm2 32.3 24.2 24.7 23.8 24.5 30.4
Rampersaud et al. [24], 2001
Width, mm 7.4 4.9 4.8 5.1 5.4 6.5 8.0
Height, mm 8.0 6.2 6.8 6.5 6.4 7.1 9.4
in 1991 were among the first to understand the three-dimensional anatomy of the
cervical pedicle and to propose that the cervical pedicle could tolerate pedicle
screw placement. In their cadaveric study of 12 cervical spines, Panjabi et al.
systematically demonstrated that both the width and height of the cervical pedicle
was the greatest at C2 (C2 width and height were on the average 7.7 and 9.4 mm,
respectively) (table 1). In addition, they demonstrated that the cross-sectional area
of the C2 pedicle was the greatest of all the cervical pedicles. From C3C7, the
pedicle angles to the transverse plane ranged from an average of 9.2 below to
13.4 above the transverse plane. In the sagittal plane, there was a decrement of
the pedicle angle from an average of 41.6 at C3 to 33.1 at C7.
Other cadaveric studies have also demonstrated the feasibility of placing
screws within the cervical pedicle. An et al. [6] in 1991 utilized a cadaveric
study to investigate the pedicle anatomy from C7T2. In terms of C7 pedicle
anatomy, they demonstrated that the medial angulation averaged 34 at C7,
while the mediolateral and superoinferior outer pedicle diameters were on aver
age 6.9 and 7.5 mm, respectively. The pedicle distances (from the entry point
to the posterior vertebral body line) measured 9.1 mm on average. An et al. rec
ommended that for pedicle screw placement, the entry point should be 1 mm
inferior to the midpoint of the facet with a 2530 medial angle.
Shin et al. [19] further defined the cervical pedicle anatomy by addressing
cross-sectional variability of the cervical pedicles. They demonstrated that the
medial pedicle walls are consistently thicker than the lateral pedicle walls and
that there was a substantial variability in the composition and shape of the cer
vical pedicle cross section.
Both Xu et al. [23] and Ugur et al. [22] expanded on prior anatomical stud
ies by evaluating the relationship between the cervical pedicles and the adjacent
neural structures. These studies showed that there was no gap between the pedi
cle and the superior portion of the nerve root and between the pedicle and the
thecal sac from C3 to C7. The average distances between the pedicle and the
Cervical Pedicle Screws 155

dramroo
Fig. 1. Entry point of C2 pedicle screw (B) is 2 mm superior and 12 mm lateral to
the C2 pars screw entry point. The C2 pedicle screw is angled more medially (C). A C1
lateral mass screw entry point; D vertebral artery; E C2 nerve root exit.
inferior nerve root margins ranged from 1.4 to 1.6 mm. Consequently, both
studies concluded that the risk of neurological injuries may be higher in screw
penetration of the medial or superior cortex of the pedicle rather than in screw
penetration of the inferior cortex of the pedicle (fig. 1).
Technique Studies
Ludwig et al. [15] compared the accuracy of three different techniques for
placing pedicle screws in cadaveric specimens. With screws placed based on
morphometric data alone, 12.5% of the screws were placed entirely within
the pedicle; 21.9% had a noncritical breach, and 65% had a critical breach
(critical: encroachment of vertebral artery, nerve root or spinal cord by the
screw, noncritical: violation of the pedicle cortex without injury to surrounding
vital structure). In the second technique (laminoforaminotomy), 45% of the
screws were within the pedicle; 15.4% had a noncritical breach, and 39.6% had
a critical breach. In the third technique (computer-assisted surgical guidance
system), 76% of the screws were placed entirely within the pedicle; 13.4% had
a noncritical breach, and 10.6% had a critical breach.
Biomechanical Studies
Once the feasibility of screw placement within the cervical pedicle was
demonstrated, biomechanical studies were required to justify the use of cervical
pedicle screw fixation in light of the technically challenging nature of pedicle
screw placement. Kotani et al. [13] in 1994 compared the biomechanical stability
Chun/Mummaneni/Birch/Sasso 156
dramroo
of seven different cervical fixation methods, including transpedicular screw
fixation. They demonstrated that the three-column fixation of the cervical spine
using cervical pedicle screws offered increased stability over other posterior
cervical fixation systems. Even when both the anterior and middle columns were
compromised, the stability provided by cervical pedicle fixation was similar to
combined anterior plate and posterior triple wiring in one-level fixation.
Additionally, Jones et al. [10] in 1997 demonstrated in cadaveric speci
mens that cervical pedicle screws had a significantly higher pullout strength
than lateral mass screws. The load failure mean for cervical pedicle screws was
677 N in contrast to 355 N for lateral mass screws.
Kowalski et al. [14] in 2000 further evaluated the pullout strengths of
pedicle screws. They compared the standard method of pedicle screw place
ment (decortication of the lateral mass and passage of a hand drill prior to tapping)
to the Abumi insertion method (decortication of the entire lateral mass, which
provides a direct view of the pedicle introitus). There was no significant differ
ence in the mean pullout resistance between the Abumi (696 N) and standard
(636.5 N) insertion techniques (p 0.41).
These studies demonstrated that cervical pedicle screw constructs are
biomechanically stronger than lateral mass screw or wire fixation systems.
Clinical Studies
In 1989 Roy-Camille [25] described the technique and the indication for
the placement of a transpars screw at C2 for Hangmans fractures. For C2, he
recommended drilling approximately 15 in the medial direction and 35 in the
superior direction.
Abumi et al. [1] in 1994 was the first to report placement of pedicle screws
in the subaxial spine. Thirteen patients with fractures/dislocation of the middle
and lower cervical spine underwent transpedicular screw fixation. The angle of
the cervical pedicle screws of Abumi et al. ranged from 25 to 45 medial to the
midline in the transverse plane. All patients had solid fusion without loss of cor
rection at an average of 22 months follow-up. Despite three cortical breaches
of the 52 screws that were placed, no neurological or vascular complications
were observed. This study demonstrated that safe and successful placement of
cervical pedicle screws was possible.
Abumi and Kaneda [2] further utilized pedicle screw fixation for nontrau
matic lesions of the cervical spine. They analyzed the clinical results in 45
patients and demonstrated that the solid fusion was obtained in all patients
except 8 patients who did not receive bone graft. There was one case of
transient radiculopathy.

dramroo
Abumi et al. [3] in 2000 followed this study with another large retro
spective study analyzing the complications associated with pedicle screw
fixation of the cervical spine. Seven hundred twelve screws were inserted
into the cervical pedicles, and the locations of 669 screws were radiologi
cally evaluated in 180 consecutive patients. Forty-five screws (6.7%) were
found to penetrate the pedicle, and 2 of the 45 screws caused a postoperative
radiculopathy. Abumi et al. concluded that the incidence of clinically
significant complications caused by cervical pedicle screw insertion was
extremely low.
Since then other studies have reported successful placement of pedicle
screws in the cervical spine. Albert et al. [5] demonstrated successful use of
C7 pedicle screws in 21 patients. Pedicle screws were placed after direct
palpation of the pedicle with a right angle nerve hook after laminoforamino
tomy at C7. There were no neurological complications related to pedicle
screw placement and no failures of fixation or complications at 1-year
follow-up.
More recently, Harms and Melcher [9] in 2001, Mummaneni et al. [16] in
2002, and Fiore et al. [7] in 2002 demonstrated a novel technique of atlantoax
ial stabilization using lateral mass fixation at C1 and C2 pars screw fixation
with minipolyaxial screws and rods. No neural or vascular damage related to
this technique was observed in these studies. The early clinical and radiologic
follow-up data indicated solid fusion in all patients.
Surgical Technique
For posterior cervical pedicle screw (and lateral mass screw) fixation, we
prefer to use a polyaxial screw-rod system (VERTEX, Medtronic Sofamor
Danek, Memphis, Tenn., USA). This system is more versatile than standard lat
eral mass plating systems and allows for more varied screw entry points and
screw angles because the screw placement is not dependent on the predeter
mined plate entry holes.
C2 Pars Screw Placement

We recommend a screw entry point 3 mm superior and 3 mm lateral to the
C2/3 facet joint. We drill approximately 15 in the medial direction and 35 in
the superior direction with direct visualization and palpation of the medial and
superior aspect of the C2 pars with a Penfield 4 to decrease the chance of cortex
violation (fig. 24). We use a handheld drill to create the screw pathway. We
then tap and place a polyaxial VERTEX screw.
dramroo
2 3
Fig. 2. A C2 pars screw (A) has a higher risk of vertebral artery injury than a C2 pedi
cle screw (B) because the vertebral artery runs occasionally through the inferior portion of
the pars of C2. SAP Superior articulating process; IAP inferior articulating process.
Fig. 3. The trajectory of a C2 pedicle screw (A) and the trajectory of a C2 pars screw
(B) are shown.
Fig. 4. Lateral cervical x-ray shows

C1 lateral mass screws and C2 pars screws.
C2 Pedicle Screw Placement

The entry point for C2 pedicle screws is 12 mm superior and 12 mm
more lateral than that of the C2 pars screw. We expose and palpate the medial
portion of the C2 pars to guide our medial trajectory, which is approximately 25

dramroo
Fig. 5. The difference in the angulation
of pedicle screws versus lateral mass screws
in the subaxial spine is shown.
(this is more medial angulation than the pars screw as the entry point is more
lateral). We also angle approximately 25 in the superior direction. We use
fluoroscopy or image guidance to help with screw trajectory.
Since the vertebral artery occasionally runs within the inferior pars of C2,
the entry point of the C2 pedicle screw is safer than the entry point of the C2
pars screw because the C2 pedicle screw entry point is more superior than the
entry point of the C2 pars screw.
Violation of the medial pedicle wall is unlikely with a C2 pedicle screw
because the bone here is cortical and quite strong. We prefer to use the tap from
the VERTEX set to enter and create a screw pathway because the tap is less
likely to create a cortical wall breach than is the drill.
C3C6 Pedicle Screw Placement

The entry point of C3C6 pedicle screws is slightly lateral to the center of
the facet and close to the posterior margin of the superior articular surface. After
decorticating the lateral mass, the pedicle can then be probed to validate screw
trajectory. The tap from the VERTEX (Medtronic Sofamor Danek) instrumenta
tion set is particularly good for this maneuver because it is delicate with fine
cutting edges and has a tendency to be sucked down the pedicle (fig. 58).
dramroo
Fig. 6. Lateral cervical spine x-ray of subaxial pedicle screws.
Fig. 7. Anterior-posterior cervical spine x-ray of subaxial pedicle screws.
Fig. 8. Axial CT of cervical pedicle screws.
Laminotomies are not routinely performed to identify the medial aspect of

the C3C6 pedicle unless prior facetectomy or laminotomy has been performed
for decompressive purposes. Image guidance with Stealth is helpful for appro
priate screw placement. Based on measurements from preoperative CT images,

dramroo
the intended angle of the pedicle screw (usually 3040 medial to the midline
in the transverse plane) can be determined, and this is confirmed intraopera
tively with image guidance.
C7 and T1 Pedicle Screw Placement

For C7 and T1 pedicle screw placement, we prefer to create lamino
foraminotomies at C6C7 and at C7T1, respectively. This window allows for
direct palpation of the medial, superior, and inferior walls of the pedicle with a
right angle nerve hook (fig. 8).
We use a 2-mm burr to establish an entry point that is based on the direct
palpation of the medial wall of the pedicle. We then use the tap from the
VERTEX instrumentation system to cut and tap the pedicle. The VERTEX tap
has a tendency to be sucked down the pedicle with minimal downward force.
If the pedicle is sclerotic, then a drill with an automatic stop at 18 mm can
be used instead. It is especially important to sound the pedicle through the
laminoforaminotomy when using the drill as perforations are more likely with
this instrument than with the VERTEX tap.
If a cortical perforation is made at C7 or T1, the safest place to perforate
the pedicle is directly lateral to it since the vertebral artery is not present at C7
or T1. The important structures are the thecal sac medially and the nerve roots
superior and inferior to the pedicle. The lateral cortex of the pedicle is a rela
tively safe zone. In fact, at T1, a lateral cortical penetration may be performed
and a longer screw used to fixate the end of the screw into the costotransverse
junction, which can increase the pullout strength of the screw [7].
Image Guidance Technology
Safe placement of cervical pedicle screws requires knowledge of the

three-dimensional anatomical structure of the pedicles and entry points, and
the pedicles relationship to neural and vascular structures. Given the wide vari
ability of cervical pedicle dimensions and questionable reliability of topo
graphical surface anatomy, greater reliance on visualization/palpation of pedicle
and/or image guidance is preferred.
Ludwig et al. [15] in 2000 investigated comparative accuracy of three
different techniques of pedicle screw placement in cadaveric specimens. They
compared screw placement using morphometric data versus laminoforamino
tomy versus image guidance. They showed that the computer-assisted, image-
guided placement had the lowest error rate.
It is our preference to use image guidance for pedicle screw placement
from C3C6. For C2, C7, and T1 it has been our experience that anatomic
dramroo
landmarks and medial pedicle wall palpation with a Penfield 4 are adequate for
accurate screw placement.
Conclusion
The purpose of posterior cervical fixation is to provide adequate fixation

to resist deforming forces until solid bony fusion occurs. To this end, lateral
mass plating has been widely utilized. The procedure has been shown to be rel
atively easy and complications are minimal. Unfortunately, there are circum
stances, such as when lateral masses are eroded or not available, that require an
alternative fixation technique. Under these circumstances, cervical pedicle
screws may be a viable alternative for fixation.
The biomechanical advantages of cervical pedicle screws have been
demonstrated. Nevertheless, current data seems to indicate that unless one is
intimately familiar with cervical pedicle anatomy and is well versed in cervi
cal pedicle screw placement, this fixation option should be performed
sparingly.
Use of image guidance for accurate cervical pedicle screw placement seems
promising, and we have successfully utilized the Stealth system for this purpose.
References
1 Abumi K, Itoh H, Taneichi H, Kaneda K: Transpedicular screw fixation for traumatic lesions of the
middle and lower cervical spine: Description of the techniques and preliminary report. J Spinal Disord
1994;7:1928.
2 Abumi K, Kaneda K: Pedicle screw fixation for nontraumatic lesions of the cervical spine. Spine
1997;22:18531863.
3 Abumi K, Kaneda K, Shono Y, Fujiya M: One-stage posterior decompression and reconstruction
of the cervical spine by using pedicle screw fixation systems. J Neurosurg 1999;90:1926.
4 Abumi K, Shono Y, Ito M, Taneichi H, Kotani Y, Kaneda K: Complications of pedicle screw
fixation in reconstructive surgery of the cervical spine. Spine 2000;25:962969.

5 Albert TJ, Klein GR, Joffe D, Vaccaro AR: Use of cervicothoracic junction pedicle screws for
reconstruction of complex cervical spine pathology. Spine 1998;23:15961599.
6 An HS, Gordin R, Renner K: Anatomic considerations for plate-screw fixation of the cervical
spine. Spine 1991;16:S548S551.
7 Fiore A, Haid R, Rodts G, Subach B, Mumanneni P, Reidel C, Birch B: Atlantal lateral mass
screws for posterior spinal reconstruction. Neurosurg Focus 2002;12/1:16.
8 Hadra BE: The classic: Wiring of the vertebrae as a means of immobilization in fracture and Potts
disease. Berthold E. Hadra. Med Times and Register, Vol22, May 23, 1891. Clin Orthop 1975;
112:48.
9 Harms J, Melcher RP: Posterior C1C2 fusion with polyaxial screw and rod fixation. Spine
2001;26:24672471.
10 Jones EL, Heller JG, Silcox DH, Hutton WC: Cervical pedicle screws versus lateral mass screws.
Anatomic feasibility and biomechanical comparison. Spine 1997;22:977982.

dramroo
11 Karaikovic EE, Daubs MD, Madsen RW, Gaines RW Jr: Morphologic characteristics of human
cervical pedicles. Spine 1997;22:493500.
12 Karaikovic EE, Yingsakmongkol W, Gaines RW Jr: Accuracy of cervical pedicle screw placement
using the funnel technique. Spine 2001;26:24562462.
13 Kotani Y, Cunningham BW, Abumi K, McAfee PC: Biomechanical analysis of cervical stabiliza
tion systems. An assessment of transpedicular screw fixation in the cervical spine. Spine 1994;19:
25292539.
14 Kowalski JM, Ludwig SC, Hutton WC, Heller JG: Cervical spine pedicle screws: A biomechani
cal comparison of two insertion techniques. Spine 2000;25:28652867.
15 Ludwig SC, Kowalski JM, Edwards CC 2nd, Heller JG: Cervical pedicle screws: Comparative
accuracy of two insertion techniques. Spine 2000;25:26752681.
16 Mummaneni P, Haid R, Traynelis V, Sasso R, Subach B, Fiore A, Rodts G: Posterior cervical
fixation using a new polyaxial screw and rod system: Technique and surgical results. Neurosurg
Focus 2002;12:15.
17 Panjabi MM, Duranceau J, Goel V, Oxland T, Takata K: Cervical human vertebrae. Quantitative
three-dimensional anatomy of the middle and lower regions. Spine 1991;16:861869.
18 Roy-Camille R, Saillant G, Berteaux D, Bisserie M: Severe strains of the cervical spine operated
on by a posterior approach. Rev Chir Orthop Reparatrice Appar Mot 1978;64:677684.
19 Shin EK, Panjabi MM, Chen NC, Wang JL: The anatomic variability of human cervical pedicles:
Considerations for transpedicular screw fixation in the middle and lower cervical spine. Eur Spine J
2000;9:6166.
20 Smith MD, Anderson P, Grady MS: Occipitocervical arthrodesis using contoured plate fixation.
An early report on a versatile fixation technique. Spine 1993;18:19841990.
21 Stanescu S, Ebraheim NA, Yeasting R, Bailey AS, Jackson WT: Morphometric evaluation of the
cervico-thoracic junction. Practical considerations for posterior fixation of the spine. Spine
1994;19:20822088.
22 Ugur HC, Attar A, Uz A, Tekdemir I, Egemen N, Caglar S, Genc Y: Surgical anatomic evaluation
of the cervical pedicle and adjacent neural structures. Neurosurgery 2000;47:11621168.
23 Xu R, Kang A, Ebraheim NA, Yeasting RA: Anatomic relation between the cervical pedicle and
the adjacent neural structures. Spine 1999;24:451454.
25 Roy-Camille R, Salient G, Mazel C: Internal fixation of the unstable cervical spine by posterior
osteosynthesis with plates and screws; in The Cervical Spine Research Society (ed): The Cervical
Spine, ed. 2, Philadelphia, Lippincott, 1989, pp 390493.
Jay Y. Chun, MD, PhD
UCSF, Department of Neurosurgery
505 Parnassus Ave., M-779, Box 0112
San Francisco, CA 94143 (USA)
Tel. 1 415 753 1772, E-Mail [email protected]
dramroo
Cervical Lateral Mass Advances

Praveen V. Mummaneni a, Vincent C. Traynelisb, Rick C. Sassoc
a
b
Division of Neurosurgery, University of Iowa, Iowa City, Iowa, and
c
Department of Orthopaedic Surgery, Indiana University School of Medicine,
Indianapolis Neurosurgical Group, Indianapolis, Ind., USA
Posterior cervical fixation utilizing lateral mass plates has been shown to
be a safe and efficacious method to achieve cervical fusion [13]. Lateral mass
plating is biomechanically superior to laminar wiring or clamping in limiting
cervical motion [47]. In addition, unlike posterior laminar wiring or clamp
ing, lateral mass plating does not require the presence of the posterior
elements.
However, lateral mass plates have numerous drawbacks. They are difficult
to contour, and the screw positions are dictated by the fixed plate entry holes.
In addition, the screw trajectories are divergent from the plate entry holes, and
the connection of the screw to the plate is not rigid. There is no space to pack
autograft bone under the screw-plate connection. Screws placed medially or lat
erally cannot be captured by the plate. Successive screws cannot be compressed
or distracted because of the fixed plate hole distances. Moreover, if the plate
needs to be revised, the screws must be removed. Finally, most of the systems
currently available do not easily allow for extension of fusion up to the occiput
or down to the thoracic spine [8].
The ideal posterior cervical instrumentation system will address these
shortfalls from the lateral mass plate systems. Currently, there are three com
mercially available systems that address the problems of lateral mass plates.
These systems do allow for initial screw placement with subsequent rod con
touring. The three systems are Starlock/Cervifix (Synthes USA, Paoli, Pa.,
USA), SUMMIT (DePuy Acromed, Raynham, Mass., USA), and VERTEX
(Medtronic Sofamor Danek, Memphis, Tenn., USA).
dramroo
Instrumentation
The three systems that allow for initial screw placement with subsequent
rod attachment differ from each other in several ways.
Starlock/Cervifix Systems (Synthes)

Unlike the VERTEX and SUMMIT systems, the Starlock/Cervifix sys
tems do not have polyaxial screw heads. These systems require the use of an
intervening closed-loop eyebolt to connect the screw to the rod and, as a result,
the connection of the rod to the screw is not a locked, rigid attachment.
Starlock and Cervifix are two instrumentation sets from Synthes that have
interchangeable hardware. Cervifix is more restrictive than Starlock because it
requires the surgeon to place the lateral mass screws through closed loop eye
bolts, which are already mounted on a rod. The closed-loop eyebolts have the
capability of sliding up and down the rod, thus enabling the screws to be placed
variably in the sagittal plane [9]. Starlock is an improved version of this system
that requires attaching closed-loop eye bolts to standard lateral mass screws and
then subsequently threading the rod through the closed-loop eyebolts. However,
the threading of a contoured rod through closed-loop eyebolts can be tedious
and sometimes impossible [9].
Both the Starlock and Cervifix systems have features that distinguish them
from standard lateral mass plate systems. They both have occipital plate exten
sions to accommodate occipitocervical fusions and laminar hooks for use when
lateral mass screw attachment is not possible. It is possible to accommodate
thoracic pedicle screws into the systems with appropriate rod contouring and
attachment of the closed-loop eyebolts to upper thoracic pedicle screws.
Starlock/Cervifix are also compatible with the thoracic hook and rod system
marketed by Synthes, and can be attached to a separate thoracic instrumenta
tion system via an adaptive connector.
VERTEX System (Medtronic Sofamor Danek)

The VERTEX system allows for initial polyaxial screw placement with
subsequent multiplanar rod contouring and attachment with or without offset
connectors. The capability of direct connection of the polyaxial screw to the rod
allows for a locked, rigid attachment of the rod to the screw. This is biome
chanically a sounder construct than the closed-loop eyebolt connection in the
Starlock/Cervifix system.
The novel VERTEX polyaxial screw heads are especially useful for facili
tating rod attachments in patients with severely abnormal cervical curvatures.
The polyaxial screw heads lock to the rod via top loading locking cap screws
Mummaneni/Traynelis/Sasso 166
dramroo
that fit inside the screw head. The VERTEX systems polyaxial screw heads are
very low profile as a result of this attachment scheme [10].
The VERTEX system is easily adaptable for occipital and thoracic exten
sions. It consists of 6- to 10-mm titanium occipital screws (which are not
polyaxial) and 14- to 18-mm titanium polyaxial cervicothoracic screws. In
addition, noncannulated 4050 mm transarticular titanium polyaxial screws are
available for C1/C2 fusion. The systems titanium rods are malleable in three
dimensions, and there is an available rod with an occipital plate on one end to
allow for occipitocervical fusions.
With the VERTEX system, occipitocervical fusion can be performed, but
the nonpolyaxial occipital screws must be placed through the apertures in the
occipital plates. The polyaxial cervicothoracic screws, however, are placed
independently of the rod system. These polyaxial 14- to 18-mm screws are ideal
for placement in the C1 lateral mass, C2 pars, C37 lateral masses, and upper
thoracic pedicles [10, 11]. The contoured rods are then linked either directly to
the polyaxial screw heads with a locking cap screw or are linked via an offset
connector.
SUMMIT System (DePuy Acromed)

The SUMMIT system shares many of the features of the VERTEX system.
SUMMIT polyaxial screw heads are placed independently of the rod and are
then connected directly to the rod through a rigid locked connection. Recently,
an adaptive occipital extension has been made available, and an extension into
the thoracic spine is also possible.
However, the SUMMIT polyaxial screw heads are higher profile than are
the VERTEX screws, and the locking cap fits around the polyaxial screw head
instead of within it, thus further increasing the profile of the system.
Consequently, thin patients who have the SUMMIT system implanted in the
posterior cervical spine may experience discomfort from the higher profile
instrumentation. In addition, in our experience, the higher profile of the polyax
ial screw heads of the SUMMIT system makes it more difficult to fit the con
toured rod into the screw heads in severely spondylotic patients whose lateral
masses are very close together as a result of their hyperlordosis.
Surgical Technique for Placement of Posterior Cervical Instrumentation

Exposure
A standard midline posterior cervical exposure is performed to reveal the
lateral aspects of the cervical facets. The exposure is extended for one to two
levels below the inferior end of the planned arthrodesis to allow for optimal
screw placement. In patients with marked degenerative changes, the osteo
phytes on the dorsal facets are removed to provide better visualization, to help
Cervical Lateral Mass Advances 167

dramroo
define the anatomy of the facets, and to provide a suitable surface to allow the
polyaxial screw heads to rotate. However, care should be taken to preserve,
where possible, the posterior cortex of the articular mass in order to provide for
better screw purchase.
In cases where posterior cervical decompression is necessary, we drill and
tap pilot holes for the cervical lateral mass screws prior to performing the full
laminectomies in order to preserve the normal anatomic landmarks for the
screw trajectories. In addition, the lamina serve to protect the neural elements
during screw hole preparation.
For occipitocervical fusions, we expose the suboccipital area up to the
inion.
For cervicothoracic fusions, we expose the thoracic transverse processes.
At C7 and T1, when canal decompression is not necessary, at least minimal
laminoforaminotomies are performed to expose the medial walls of the C7 and
T1 pedicles to help guide C7 and T1 pedicle screw placement.
Screw Placement (VERTEX System)

We prefer to use the VERTEX system for posterior occipitocervicothoracic
arthrodesis. This system has the advantage of having low profile polyaxial
screw heads that can attach either directly or via an offset connector to the rod.
After the exposure is completed, we turn our attention to cervical polyax
ial screw placement. Initially, we perforate the posterior cortices of the lateral
masses with a high-speed drill. Our screw trajectories for C37 are based on the
guidelines set by Haid, Papadopoulos, and Sonntag (unpubl. data, presented at
the American Association of Neurological Surgeons Annual Meeting, 1991)
and further elucidated by McCafferty et al. [12]. Entry points are 1 mm medial
to the center of the lateral mass and trajectories are 2030 cephalad and 2030
lateral (fig. 1, 2). We normalize the entry point and screw trajectory at each
lateral mass to allow for changes in the orientation of the lateral masses sec
ondary to accentuated cervical lordosis or kyphosis and to allow for each
patients unique pathoanatomy.
Some surgeons elect to place screws with the assistance of fluoroscopy or
image guidance. However, we do not routinely use fluoroscopy or image guid
ance (except for screw placement into C1 and C2). Consequently, attention to
the patients unique cervical anatomy is of paramount importance for us.
For screw placement into the lateral mass of C1, we utilize the technique
described by Harms and Melcher [13] and refined by Fiore et al. [11]. The
screw entry point is at the junction of the posterior arch of C1 and the center of
the posterior, inferior C1 lateral mass. The C2 nerve root is gently retracted
inferiorly with a Penfield 4 to expose the screw entry point. The screw trajec
tory is parallel to the plane of the C1 lamina and is aimed straight anterior from
dramroo
a 2025 010 b
Fig. 1. Illustration of the lateral mass screw trajectory in the axial plane. a Appropriate
angle. b The screw trajectory puts the nerve root at risk.
Fig. 2. Lateral mass screw trajectory

in the sagittal plane.
the entry point. Screw depth is guided by lateral fluoroscopy and can be esti
mated by preoperative Stealth CT planning.
For screw placement into the C2 pars, we pay close attention to the preop
erative CT to assess the course of the vertebral artery. In addition, we palpate
the medial pars with a blunt probe to help guide our screw trajectory. We utilize
a screw entry point 3 mm superior and lateral (3 mm up and out) to the medial
aspect of the C2/3 facet joint. The screw trajectory is 1015 medial and 35
cephalad. We typically use 4-mm width and 16-mm length screws for the C2
pars [10, 11].
For C1/2 transarticular screw placement, the entry point and trajectory are
the same as for C2 pars screws; the screw length, however, is longer (up to 50 mm

dramroo
in some cases) [14, 15]. We prefer to use lateral fluoroscopy for guidance when
placing C1/2 transarticular screws.
Screw placement at C7 is dependent on the bony anatomy. We scrutinize the
preoperative CT scan to assess if the patients C7 lateral mass has a typical cer
vical anatomy or has a transitional thoracic anatomy with a well-formed pedicle.
When the C7 anatomy is transitional, we prefer to place a C7 pedicle screw.
For pedicle screw placement at C7 or in the upper thoracic spine, we
expose and palpate the medial walls of the pedicles and utilize an entry point
1 mm below the center of the facet joint. Our screw trajectory is typically
2530 medial while maintaining a perpendicular angle in the sagittal plane [10].
We have found it useful to use the drill to create a small pilot hole for entry
of the tap into the pedicle. We then use the 4.0-mm tap from the VERTEX set
to create the screw path and simultaneously tap the pedicle. This instrument is
delicate and particularly useful to find an appropriate trajectory into the C7 and
thoracic pedicles. It has a tendency to be sucked down the pedicle with gentle
manipulation in experienced hands.
In the thoracic spine, pedicle screws also can be placed laterally into the
costotransverse joint to achieve greater cortical purchase [16].
After polyaxial screw placement, the appropriate posterior decompressions
are performed based on the patients symptoms. In addition, the facet joints to
be fused are stripped of cartilage and decorticated with a high-speed drill and
packed with autograft.
The final step is rod contouring and attachment. A rod template is used to
estimate the rod length and rod contour required. A small endotracheal tube
stylet is particularly good for use as a rod template. The titanium rods are then
measured, cut, contoured, and directly attached to the polyaxial screw heads
with locking cap screws. In cases where the patients pathoanatomy requires
significantly different lateral or medial screw positions at successive levels, we
utilize small offset connectors in order to facilitate rod attachment.
When occipitocervical fusions are planned, we utilize a specialized rod
with an occipital plate on the cephalad end. The rod and occipital plate are
contoured based on the rod trial. We position the occipital plate near the midline
occipital keel to provide for the most bony purchase for the occipital screws.
The occipital screws are not polyaxial and must be placed through the apertures
in the occipital plates. Ideally six occipital screws (three on each side) should be
placed near the midline of the suboccipital bone. This has been shown to be the
most biomechanically favorable position for suboccipital fixation [17] (fig. 3).
After the instrumentation construct is placed, but before final tightening of
the construct, we compress, distract, or laterally rotate each successive segment
as needed to restore normal cervical lordosis and sagittal balance. Autograft is
then packed over the tops of the fusion sites.
dramroo
Fig. 3. Optimal screw placement
positions for occipital fixation.
Discussion and Conclusion
Lateral mass plating has been shown to be an effective method of achiev

ing posterior cervical arthrodesis and stabilization [3]. Reported complication
rates are low. Injuries to the cervical spinal cord or vertebral artery have not
been reported in recent large published series on lateral mass plating [3, 5,
1820]. The reported rate of radiculopathy from malpositioned screws has
ranged from 0 to 6% of patients [3, 5, 1820].
However, lateral mass plates are of limited use when fusing from the
occiput to the thoracic spine in patients with abnormal cervical anatomy
because of their lack of malleability and their predetermined screw hole trajec
tories. The Starlock/Cervifix systems overcome some of these problems but are
suboptimal because they require threading of a contoured rod through closed
loop eyebolts, which is not a rigid connection and which can be tedious and
sometimes impossible [9].
The ideal posterior cervical instrumentation system will allow for initial
screw placement with subsequent multiplanar rod contouring and attachment
via a rigid connection with a locking cap screw. In addition, offset connectors
are needed to allow for screw capture at any distance laterally away from the
rod (as is the case with C7 and T1 pedicle screws (fig. 4). Finally, the system
should be low profile so that the polyaxial screw heads will not be crowded in
cases with cervical hyperlordosis.
The new VERTEX and SUMMIT posterior polyaxial cervical screw and
rod systems satisfy many of the criteria for the ideal posterior cervical instru
mentation system. They have the versatility to accommodate occipitocervical
fusions with C1 lateral mass screws and C2 pars screws. In addition, they allow

dramroo
Fig. 4. Different screw trajectory and
entry point needed for C7 and T1 pedicle
screws as compared to cervical lateral mass
screws.
Fig. 5. Photograph of the VERTEX system. Note the polyaxial screw heads, the offset
connector, and the laminar hook.
for lateral mass fixation from C37 as well as pedicle fixation and laminar
hook placement in the lower cervical and upper thoracic spine without the
limitations inherent in placing screws through holes in lateral mass plates
(fig. 58). Finally, the rods are easily contoured and attached directly and
rigidly to the polyaxial screw heads.
dramroo
Fig. 6. Photograph of occipitocervicothoracic fixation in a saw bone model utilizing
the VERTEX system.
7 8
Fig. 7. Preoperative x-ray in a severely kyphotic and myelopathic patient.

Fig. 8. Postoperative x-ray in the same patient following posterior cervical decom
pression and occipitocervicothoracic fusion with the VERTEX system.
The VERTEX system has advantages over the SUMMIT system. VERTEX
is lower profile than the SUMMIT system and more easily accommodates com
plex cervical curvatures and spondylosis as a result. In addition, VERTEX has
top-loading offset connectors to easily accommodate cervicothoracic pedicle
screws.
Further studies will be required to establish long-term results and fusion
rates with the cervical screw-rod systems.

dramroo
Acknowledgments
We are grateful to Drew Imhulse for assistance with the radiographic images in the figures.
Disclosure Statement
The following authors are consultants for Medtronic Sofamor Danek: Vincent C.
Traynelis, MD and Rick C. Sasso, MD.
References
1 Anderson PA, Henley MB, Grady MS, Montesano PX, Winn HR: Posterior cervical arthrodesis
with AO reconstruction plates and bone graft. Spine 1991;16(suppl 3):S72S79.
2 Fehlings MG, Cooper PR, Errico TJ: Posterior plates in the management of cervical instability:
Long-term results in 44 patients. J Neurosurg 1994;81:341349.
3 Graham AW, Swank ML, Kinard RE, Lowery GL, Dials BE: Posterior cervical arthrodesis and
stabilization with a lateral mass plate: Clinical and computed tomographic evaluation of lateral
mass screw placement and associated complications. Spine 1996;21:323329.
4 Gill K, Paschal S, Corin J, Ashman R, Bucholz RW: Posterior plating of the cervical spine: A bio
mechanical comparison of different posterior fusion techniques. Spine 1988;13:813816.
5 Heller JG, Silcox DH 3rd, Sutterlin CE 3rd: Complications of posterior cervical plating. Spine
1995;20:24422448.
6 Sutterlin CE, McAfee PC, Warden KE, Rey RM, Farey ID: A biomechanical evaluation of cervi
cal spinal stabilization methods in a bovine model: Static and cyclical loading. Spine 1988;13:
795802.
7 Ulrich C, Woersdoerfer O, Kalff R, Claes L, Wilke HJ: Biomechanics of fixation systems to the
cervical spine. Spine 1991;16(suppl 3):S4S9.
8 Smith MD, Anderson P, Grady MS: Occipitocervical arthrodesis using contoured plate fixation:
An early report on a versatile fixation technique. Spine 1993;18:19841990.
9 Horgan MA, Kellogg JX, Chesnut RM: Posterior cervical arthrodesis and stabilization: An early
report using a novel lateral mass screw and rod technique. Neurosurgery 1999;44:12671272.
10 Mummaneni PV, Haid RW, Traynelis VC, Sasso RC, Subach BR, Fiore AJ, Rodts GE: Posterior
cervical fixation using a new polyaxial screw and rod system: Technique and surgical results.
Neurosurg Focus 2002;12/1:15.
11 Fiore AJ, Haid RW, Rodts GE, Subach BR, Mummaneni PV, Riedel CJ, Birch BD: C1 lateral mass
screws for posterior spinal reconstruction: Technical note and case series. Neurosurg Focus
2002;12/1:16.
12 McCafferty RR, Haid RW, Martin G, Rodts GE: Techniques for lateral mass plating of the post
erior cervical spine in patients with cervical spinal stenosis. Tech Neurosurg 1999;5/2:101105.
13 Harms J, Melcher RP: Posterior C1C2 fusion with polyaxial screw and rod fixation. Spine
2001;26:24672471.
14 Haid RW Jr: C1C2 transarticular screw fixation: Technical aspects. Neurosurgery 2001;49:
7174.
15 Haid RW Jr, Subach BR, McLaughlin MR, Rodts GE Jr, Wahlig JB Jr: C1C2 transarticular screw
fixation for atlantoaxial instability: A 6-year experience. Neurosurgery 2001;49:6568.
17 Oda I, Abumi K, Sell LC, Haggerty CJ, Cunningham BW, McAfee PC: Biomechanical evaluation
of five different occipito-atlanto-axial fixation techniques. Spine 1999;24:23772382.
dramroo
18 Schultz KD Jr, McLaughlin MR, Haid RW Jr, Comey CH, Rodts GE Jr, Alexander J: Single-stage
anterior-posterior decompression and stabilization for complex cervical spine disorders. J Neurosurg
2000;93(suppl 2):214221.
19 Swank MA, Sutterlin CE 3rd, Bossons CR, Dials BE: Rigid internal fixation with lateral mass
plates in multilevel anterior and posterior reconstruction of the cervical spine. Spine 1997;22:
274282.
20 Wellman BJ, Follett KA, Traynelis VC: Complications of posterior articular mass plate fixation of
the subaxial cervical spine in 43 consecutive patients. Spine 1998;23:193200.
Praveen V. Mummaneni, MD
Clinical Instructor/Complex Spine Fellow
Department of Neurosurgery, The Emory Clinic
1365 Clifton Road, Building B, Suite 6400, Atlanta, GA 30322 (USA)

dramroo
Interbody Carbon Fiber

Rudolf Bertagnoli
Klinikum St. Elisabeth, Spinal Department, Straubing, Germany
The fusion of spinal segments is one of the major goals in surgical treat
ment of degenerative disc disease (DDD). To obtain the best biomechanical
support and fusion rates, interbody fusion is the preferred method. The main
advantages of implants made of carbon fiber reinforced plastic (CFRP) are the
radiolucency and the fact that there is no distortion on CT and MRI. The
surgeon can chronologically follow the biological reaction in the fusion section
and can clearly detect bony fusion. Regions which are hidden using metal
implants could now be analyzed on standard radiographs. In implants made of
CFRP matrix materials such as thermoset epoxy resin systems (EPN/DDS) or
thermoplastic systems PEAK (PEKEKK, PEEK, ULTRAPEK) are
used. The first ones have been clinically used since 1988 following detailed
in vitro and in vivo tests starting 1975 according to ISO 10993-1 [1]. In the
spine these implant materials have been used since 1993 [2].
These CFRP implants are manufactured using a special fiber winding process
of carbon fiber roving which is impregnated with resin and laid on a rotating rod.
Due to that and to the design-adapted machining processes the properties of the
material could be chosen for each kind of implant design. Hardening, tempering,
and machining lead to the final implant geometry and must be controlled.
Interbody fusion with cage support is an ideal situation for the application of
resorbable biomaterials [3, 4]. To monitor the production of the new bone masses
and to determine the degree of bone fusion, cages made from carbon composite
materials are superior to metallic cages due to their radiolucent characteristics.
In the literature some mechanical and in vitro tests performed on CFRP
implants could be found: Brantigan et al. [5], Ciapetta et al. [6] (vertebra
replacement), Jost et al. [7], Kandziora et al. [8] (80 cervical spines of sheep
with different cage designs), Shono et al. [9] (18 calf spines in compression and
rotation), and Steinhauser et al. [10] (static and dynamic compression and shear
dramroo
Fiber Matrix
Fig. 1. Cross section and SEM (artificial fracture) of CFRP implants: carbon fiber
with a diameter of 5 m surrounded by matrix material.
testing) all showed very good results. The first animal tests of CFRP implants
as interbody lumbar fusion devices were done by Brantigan et al. [5]; they
reported 100% fusion after 6, 12 and 24 months in 27 Spanish goats. A sum
mary of clinical experience with CFRP cages is shown in table 1. In all cases
high fusion rates were detected with no device-related complications. One case
report showed infection and a broken CFRP cage (see table 1) [17].
Material and Methods
Mechanical Tests
To test the axial compression and shear behavior the spinal implants were loaded
between two parallel stainless steel plates according to ASTM F2077-00. An axial force rate
of 500 N/min was used for quasistatic testing. The loading was stopped when a permanent
failure of the specimen occurred or a displacement of 3.0 mm was reached.
For dynamic testing the specimens were tested under cyclic fatigue using a sinusoidal
loading waveform at a constant frequency of 5 Hz (lumbar) and 12 Hz (cervical), with an
R ratio (Pmin/Pmax) of 0.1. The maximum load cycle was 5,000,000 cycles and a displace
ment limit error of 3 mm was established.
To determine the pull-out strength, the implants were placed in between two blocks
made of Rohacell RC 300WL with mechanical properties comparable to cancellous bone. An
axial preload of 100 N was applied. The instrument intended for clinical use was attached
Interbody Carbon Fiber 177

dramroo
Table 1. Summary of the clinical experience with CFRP cages
Location Cases Follow-up Results Ref. No.
Lumbar (PLIF) 71 28 months 90% fusion rate, 11

66% overall satisfaction rate
Lumbar (PLIF) 11 648 months 90% fusion rate after 12
48 months
Lumbar (PLIF) 221 24 months 98.9% fusion rate, 13
86% clinical success
Lumbar (PLIF) 51 12 months 86% fusion rate 14
(89% of levels)
Lumbar (L4L5) 15
Bilateral 83 1224 months 97% fusion rate
Unilateral 80 100% fusion rate
Lumbar (ALIF and PLIF), 70 312 months 98% fusion rate 2
cervical
Cervical 19 1221 months 100% fusion rate 16
Lumbar (PLIF) 1 4 week to 25 months Case report of infection 17
and broken CFRP cage
PLIF Posterior lumbar interbody fusion; ALIF anterior lumbar interbody fusion.
according to the procedure defined by the manufacturer. A constant deflection rate of

5 mm/min was used. All tests were performed dry at room temperature.
To determine the ultimate torque load, the implants were placed in between two blocks
made of Rohacell RC 300WL with mechanical properties comparable to cancellous bone.
Prior to testing, an axial preload of 1,000 N was applied to enable settlement of the contact
area. The upper as well as the lower block were fixed versus the torsion measurement sys
tem. A constant axial preload of 100 N was applied during the test. Torsion was applied at a
constant rate of about 100/min.
All static tests were performed dry at room temperature. The dynamic tests were per
formed under Ringer solution at 37C temperature and all specimens were preconditioned.
Clinical Experience
The fusion devices to be assessed at the postoperative evaluation were the CORNER
STONE-SR C (lordotically shaped anterior cervical interbody fusion device, Medtronic
Sofamor Danek) and UNION or UNION-L (lordotically shaped anterior or lateral lumbar
interbody fusion device, Medtronic Sofamor Danek) spinal fusion cages.
The implant system CORNERSTONE-SR C consists of a basic body (12 13 mm),
various heights of 48 mm (1016 mm) and two lines of fins (see fig. 2). The anterior
retroperitoneal implant system UNION or UNION-L consists of a basic body (24 26 or
26 31 mm, respectively) with two chambers, various heights of 1016 mm and three lines
of fins (see fig. 2). For both types of implants x-ray contrasts are layers of BaSO4 embedded
into the CFRP material. Prior insertion of the implant the fins must be prepared in the bone
with a special instrument.
Bertagnoli 178
dramroo
a b
Fig. 2. CFRP implants: cervical interbody fusion cage (CORNERSTONE-SR C, a)

and anterior lumbar interbody fusion cage (UNION, b).
Table 2. Primary diagnoses of all studies
Diagnosis Retrospective study Prospective study
cervical lumbar cervical lumbar
Failed back disc surgery syndrome 2 40 18

Segmental instability 55 19 6 8
Disc herniation 77 2 28 10
Spondylolisthesis (grade 13) 25
Spinal stenosis 2 3
Osteochondrosis 2
Fracture 1
In a retrospective study preoperative, peroperative, and postoperative data were col

lected and entered in case report forms and a comparison was made between preoperative
and postoperative clinical status (diagnoses see table 2).
In the cervical group 206 cages were used in 139 patients. There were 69 men and
70 women (36.2% smokers), with a mean age of 46.5 years (from 17 to 68), a mean weight
of 75.4 kg (from 48 to 114 kg) and a mean height of 170.2 cm (from 151 to 196 cm). 25.2%
of the patients had 1 previous spinal surgery and 4.2% had 2 previous spinal surgeries.
In 91.4% of cases tibia was used as an autologous graft.
In the lumbar group 146 cages were used in 89 patients. There were 54 men and
35 women (36.8% smokers), with a mean age of 44.2 years (from 21 to 65), a mean weight
of 76.3 kg (from 51 to 114 kg) and a mean height of 172.6 cm (from 155 to 197 cm). 64.0%
of the patients had 1 previous spinal surgery and 6.7% had 2 previous spinal surgeries.
In 94.4% of cases the pelvic bone was used as an autologous graft.

dramroo
45
CORNERSTONE-SR C
40
UNION
35
30
25
20
15
10
5
0
12 months 24 months 36 months 48 months
Fig. 3. Minimum follow-up of the CFRP cervical and lumbar fusion implants:
CORNERSTONE-SR C (n 126) and UNION (n 89).
In 2% of the cervical cases there was a supplemental anterior fixation; in 97.8% an

anterior-lateral approach was used. C3C4 was operated on in 1.5%, C4C5 in 11.7%,
C5C6 in 52.4%, C6C7 in 34.0%, and C7T1 in 0.5% of the cases. In 91% of the lumbar
cases there was a supplemental posterior fixation; in 94.4% an anterior approach was used.
L5S1 was operated on in 50.7%, L4L5 in 39.0%, L3L4 in 8.9% and L2L3 in 1.4% of
the cases.
In addition a prospective study was performed on 34 cervical and 36 lumbar patients
(diagnoses see table 2). The group of 34 cervical patients consisted of 19 males and
15 females with an average age of 47.1 years (range 2677), an average weight of 76.5 kg
and an average height of 169.6 cm. The group of lumbar patients consisted of 17 male and
19 female patients with an average age of 48 years (range 1673), an average weight of
74.3 kg and an average height of 170.6 cm. In the cervical group 27 patients had no and
7 patients had 1 or 2 previous spinal operations in the affected areas; in the lumbar group
18 patients had no and 18 patients more than 1 operation.
In 7 cervical patients a supplemental screw and plate fixation was performed. The
patients were treated with fusion surgery using the CFRP CORNERSTONE-SR C cages
filled with the bone regeneration substance Colloss, an osteoinductive natural protein com
plex. Monosegmental surgery was performed; in 10 patients a multisegmental procedure was
carried out. There were 28 patients for the 3-month follow-up, 17 patients for the 6-month
follow-up and 8 patients for the 12-month follow-up.
In the lumbar group, 34 patients were treated with the CFRP UNION cages and poste
rior pedicle fixation devices were used; the other 2 patients underwent stand-alone proce
dures. In 29 patients, one camber of the cage was filled with autologous cancellous bone and
the other camber with Colloss. In 7 patients both cambers were filled exclusively with
Colloss. In 26 of the patients, monosegmental surgery was performed; in 10 patients a mul
tisegmental procedure was carried out.
The following radiographical criteria were used to detect fusion: bone in the fusion area
is more dense and more mature than originally achieved in surgery, no interface between
donor bone and vertebral bone; sclerotic line between graft and vertebral bone, mature bone
trabeculae bridging the fusion area, resorption of anterior vertebral traction spurs, fusion of
facet joints and ring phenomenon on CT.
Bertagnoli 180
dramroo
The neurological status was assessed postoperatively using a comprehensive neurolog
ical status scale ranging from 0 (worst) to 100 (best). Pre- and postoperatively each patient
rated the frequency of their back/leg pain on a scale of 110. Patient satisfaction questions
and success are defined as either a completely recovered, much improvement, or slightly
improved response.
For the prospective study the arm/neck score was used to evaluate and compare the pain
intensity and frequency in the cervical groups (with 1 equal to no pain or never and
10 equal to extremely painful or always). The results of the lumbar group were evaluated
and compared using the Oswestry score. In this scale patients can receive a maximum value
of 5 points and a minimum value of 0 points on each of the 10 criteria being used.
Clinical examination and comparative measurements were performed on AP and lateral
radiographs as well as on flexion/extension films in order to estimate the position of the
cages, the degree of restoration of the disc height of the intervertebral discs and the density
of the fusion area. MRIs were taken at defined times to verify possible edema that had been
reported when using graft material. CTs and regular tomograms were also used to evaluate
new bone formation inside the cages.
Results and Discussion
The standard mechanical tests showed the superior strength of the CFRP
fusion devices. The cervical fusion device (CORNERSTONE-SR C) has a high
compression and shear load (see fig. 7) which is more than 20 times above the
normal load in the cervical spine [18]. Pull-out and torsion tests show significant
differences between designs with and without the fins (see fig. 8) but even for the
nonfin version the values reached are above the daily loads [18]. Figure 9 shows
the compression and shear behavior of the CFRP lumbar fusion device UNION.
No significant differences could be detected between different sizes (small
and medium) and different heights (10 up to 20 mm) of the implants. The com
pression load is more than 4 times higher than the load during daily activities
[18]. The fatigue behavior due to compression and shear load is shown in
figure 10. Run-out load is much higher than could be detected during daily
activities. In the clinical studies no intraoperative or device-related complications
were observed in either group. In all cervical and lumbar cases of the retro
spective study, 100% fusion was determined.
In the cervical group the mean preoperative back/leg pain frequency score
was 7.3 and the mean postoperative back/leg pain frequency score was 5.2
(28.8% improvement) (one clinical case, see fig. 4af). In the lumbar group the
mean preoperative back/leg pain frequency score was 7.9 and the mean post
operative back/leg pain frequency score was 6.2 (21.5% improvement) (two
clinical cases, see fig. 5ac, 6ad).
In the cervical group patients were satisfied with the results of surgery in
81.6%, surgery helped as much as patient thought it would in 79.0%, and

dramroo
a b c d e f
Fig. 4. af Radiography showing the CFRP fusion implant CORNERSTONE-SR C

(visible due to a BaSO4 layer in the implants): 40-year-old female, vertical segmental insta
bility C5C6, radicular pain with dysesthesia C5, severe neck pain, unsuccessful conservative
treatment. Good realignment of cervical lordosis and secure fusion at the 6-month checkup.
a b c
Fig. 5. ac Conventional tomogram and CT scan show excellent visualization also

inside the UNION cage. Secure determination of fusion showing the CFRP fusion implant
UNION (visible due to a BaSO4 layer in the implants); here at 6 months checkup is possible.
patients would have the same surgery again in 80.7% of the cases. In the lum
bar group patients were satisfied with results of surgery in 71.4%, surgery
helped as much as patient thought it would in 57.2%, and patient would have
the same surgery again in 67.5% of the cases. The independent physicians
assessment showed in 95.3% of the cervical and 97.7% of the lumbar cases
excellent or good results (see table 3). 6% of the cervical and 13% of the lum
bar patients in the retrospective study showed general and local complications.
Bertagnoli 182
dramroo
a b c d
Fig. 6. ad Radiography showing the CFRP fusion implant UNION-L (L4L5, visible
due to a BaSO4 layer in the implants): 42-year-old female, failed back disc surgery in 1999,
vertical segmental instability L4L5, severe low back pain, minor radicular pain, unsuccess
ful conservative treatment. Posterior neurolysis L5 and stabilization L4L5, ALPA approach.
Good realignment of lumbar lordosis and secure fusion at the 12-month checkup.
4
Load (kN)
0
CORNERSTONE-SR C 5mm CORNERSTONE-SR C 4mm
Fig. 7. Compression and shear strength of the CFRP cervical interbody fusion cage
CORNERSTONE-SR C tested according to ASTM F2077-00.
In the cervical group of the prospective study 97% fusion could be detected
radiographically; in 1 patient (3%) due to only one early check-up at 6 weeks
after surgery fusion could not be evaluated. The mean arm/neck pain score
improved significantly from 7.6 preoperatively to 4.2 after 12 months (45%
improvement), the arm/neck pain frequency score from 8.8 to 5.4 after 12 months
(39% improvement). 87.5% of the patients felt there was an improvement

dramroo
100 N 100 N CSR-C Serrated cage 4mm
100 CSR-C 7mm 5
p 0.05
80 4
p0.05
Torque (Nm)
Load (N)
60 3
40 2
20 1
0 0
Pull-out test Torsion
Fig. 8. Pull-out force and torsion behavior of CFRP cervical interbody fusion devices:
comparison of implants with rails (CORNERSTONE-SR C) and without rails (CORNER
STONE-SR C serrated).
60
40
Load (kN)
20
0
UNION small 14mm UNION medium 20mm UNION large 15mm
Fig. 9. Compression and shear strength of the CFRP anterior lumbar interbody fusion
cage UNION tested according to ASTM F2077-00.
and were satisfied after 12 months, the results of surgery met the expectations of
75%, and 87.5% would have the same surgery again.
In the lumbar group 100% fusion could be detected radiographically; no
pseudarthrosis was determined. The mean Oswestry score was 52.3 before
surgery and improved after 6 months to 16.2 (69% improvement). 90.0% of the
Bertagnoli 184
dramroo
30
UNION medium 11mm
UNION large 15mm
25
Maximum load (kN)
20
15
10
Run-out
0
10,000 100,000 1,000,000 10,000,000
Cycles
Fig. 10. S/N curve of compression and shear strength of the CFRP anterior lumbar
interbody fusion cage UNION tested according to ASTM F2077-00.
Table 3. Independent physicians

assessment of the retrospective study Cervical, % Lumbar, %
Excellent 73.4 86.5

Good 21.9 11.2
Fair 4.7 2.3
Poor 0.0 0.0
patients felt there was an improvement and were satisfied after 6 months, the
results of surgery met the expectations of 80%, and 60.0% would have the same
surgery again. Only 1 patient was not satisfied with the result.
Conclusion
Spine stabilization is an important area of orthopedic surgery, especially

due to an increasing number of patients with chronic low back pain and DDD.
For interbody fusion, besides autologous or allograft bone grafts, fusion cages
composed of metal or polymer as a space holder are increasingly used.
Especially CFRP are used in the medical field when radiolucency, high

dramroo
mechanical strength and innovative design are required. During the manufac
turing process of the material CFRP carbon fibers are embedded into a ther
moset (e.g. epoxy resin EPN/DDS) or thermoplastic (e.g. PEAK) resin matrix.
CFRP is biocompatible, radiolucent and has a higher mechanical potential
compared to other implant materials and has clinically been used since 1993 for
fusion implants of the cervical and lumbar spine. The clinical experience of
cervical and lumbar fusion systems showed high fusion rates and no carbon
fiber-related complications. In addition prospective studies have shown that the
use of osteoinductive bone graft material in conjunction with CFRP cages is a
safe and effective method with high fusion rates in the treatment of DDD.
References
1 Boenisch B, Ascherl R, Liebendrfer A, Geissdoerfer K, Schmeller ML, Kerschbaumer S, Scherer

MA, Erhardt W, Scheer W, Bluemel G: Five-year functionality study of carbon fibre reinforced
acetabular implants in dogs; in Heimke G, Soltsz U, Lee AJC (eds): Clinical Implant Materials.
Advances in Biomaterials. Amsterdam, Elsevier, 1990, vol 9, pp 391396.
2 Bertagnoli R: Continuum of care in degenerative disc disease. Backup 2001;1:48.
3 Sandhu HS: Anterior lumbar interbody fusion with osteoinductive growth factors. Clin Orthop
2000;371:5660.
4 Summers BN, Eisenstein SM: Donor site pain from the ilium. A complication of lumbar spine
fusion. J Bone Joint Surg Br 1989;71:677680.
5 Brantigan JW, McAfee PC, Cunningham BW, Wang H, Orbegoso CM: Interbody lumbar fusion
using a carbon fiber cage implant versus allograft bone. An investigational study in the Spanish
goat. Spine 1994;19:14361444.
6 Ciapetta P, Boriani S, Fava GP: A carbon fiber reinforced polymer cage for vertebral body replace
ment: Technical note. Neurosurgery 1997;41:12031206.
7 Jost B, Cripton PA, Lund T, Oxland TR, Lippuner K, Jaeger P, Nolte LP: Compressive strength of
interbody cages in the lumbar spine: The effect of cage shape, posterior instrumentation and bone
density. Eur Spine J 1998;7:132141.
8 Kandziora F, Pflugmacher R, Schafer J, Born C, Duda G, Haas NP, Mittlmeier T: Biomechanical
comparison of cervical spine interbody fusion cages. Spine 2001;26:18501857.
9 Shono Y, McAfee PC, Cunningham BW, Brantigan JW: A biomechanical analysis of decompres
sion and reconstruction methods in the cervical spine. Emphasis on a carbon-fiber-composite
cage. J Bone Joint Surg Am 1993;75:16741684.
10 Steinhauser E, Bader R, Rechl H, Bertagnoli R, Mittelmeier W, Gradinger R: Mechanical testing
of spinal interbody implants Characteristics and limitations of a standardized testing procedure.
Biomed Tech 2001;46:325332.
11 Agazzi S, Reverdin A, May D: Posterior lumbar interbody fusion with cages: An independent
review of 71 cases. J Neurosurg 1999;91:186192.
12 Diedrich O, Perlick L, Schmitt O, Kraft CN: Radiographic characteristics on conventional radio
graphs after posterior lumbar interbody fusion: Comparative study between radiotranslucent and
radiopaque cages. J Spinal Disord 2001;14:522532.
13 Brantigan JW, Steffee AD, Lewis ML, Quinn LM, Persenaire JM: Lumbar interbody fusion using
the Brantigan I/F cage for posterior lumbar interbody fusion and the variable pedicle screw place
ment system: Two-year results from a Food and Drug Administration investigational device
exemption clinical trial. Spine 2000;25:14371446.
14 Tullberg T, Brandt B, Rydberg J, Fritzell P: Fusion rate after posterior lumbar interbody fusion
with carbon fiber implant: 1-year follow-up of 51 patients. Eur Spine J 1996;5:178182.
Bertagnoli 186
dramroo
15 Commarmond J: One-segment interbody lumbar arthrodesis using impacted cages: Posterior
unilateral approach versus posterior bilateral approach. Rev Chir Orthop Reparatrice Appar Mot
2001;87:129134.
16 Brooke NSR, Rorke AW, King AT, Gullan RW: Preliminary experience of carbon fibre cage
prostheses for treatment of cervical spine disorders. Br J Neurosurg 1997;11:221227.
17 Tullberg T: Failure of a carbon fiber implant. A case report. Spine 1998;23:18041806.
18 White AA, Panjabi MM: Clinical Biomechanics of the Spine, ed 2. Philadelphia, Lippincott, 1990.
Dr. med. Rudolf Bertagnoli
Kay 2a, D94315 Straubing (Germany)
Tel. 49 171 4325300, Fax 49 9421 530475, E-Mail [email protected]

dramroo
Thoracic Pedicle Screw Placement

Lawrence G. Lenke
Orthopaedic Surgery, Pediatric and Adult Spinal, Scoliosis and Reconstructive
Surgery, Washington University Medical Center, Shriners Hospital in St. Louis,
St. Louis, Mo., USA
The use of thoracic pedicle screws in the treatment of various spinal

disorders has shown a dramatic increase in the last decade [14]. The bio
mechanical stability afforded by transpedicular fixation, the ability to correct
various sagittal and coronal plane spinal malalignments and deformities, and the
available room for bone grafting around the implants provide an optimal envi
ronment for spinal fusion success [510]. This chapter will discuss the tech
nique of free-hand thoracic pedicle screw placement using anatomic landmarks,
and a special thoracic pedicle probe combined with appropriate surgical
3-dimensional orientation and feel to access the vertebral body via the pedicle
channel [11]. This free-hand technique can be adopted by any surgeon comfort
able placing lumbar pedicle screws and experienced with thoracic spinal
anatomy and surgical techniques, and does not require the use of continuous or
intermittent intraoperative fluoroscopy, radiography or navigational systems.
Preoperative Assessment
One of the first assessments that needs to be made by the surgeon contem
plating the use of thoracic pedicle screws for the treatment of various surgical
spinal disorders is whether the thoracic pedicle is large enough to probe down
with a pedicle probe and subsequently accept a screw of appropriate and avail
able diameter and length. There is a wide spectrum of individual pedicle dimen
sions, with the limiting dimension always being the medial to lateral diameter at
the isthmus in millimeters [1215]. This can range from only 23 mm in the mid
thoracic region up to 10 mm at the lower thoracic levels. Interestingly, the
smallest pedicle dimensions are usually found at the T5T7 levels. Pedicle
dramroo
isthmus dimensions increase both above and below these mid thoracic levels,
being consistently larger in the lower thoracic region (T10T12), and somewhat
smaller in the proximal thoracic region (T1T3). Because of the varied anatomy
and pedicle sizes, it is extremely important to have a variety of choices for ped
icle screw diameters available ranging from a 4.0-mm all the way through a
7.5-mm diameter in 0.5-mm increments. The goal should be to maximize fit into
the pedicle cortex with the largest screw that will safely fit. The lengths of the
screws will vary but range from 4050 mm in length in the lower thoracic
regions to 2535 mm in length in the proximal thoracic levels. The lengths will
depend on patient size, and orientation of the screw, and will be shorter if the
trajectory is somewhat lateral. It is also helpful to have both fixed head and
multiaxial screws available. I tend to use fixed head screws in the thoracic spine,
for they are lower profile and allow more room for bone grafting. However,
multiaxial screws can easily be used as well.
In most instances, standard AP and lateral radiographs will adequately
demonstrate the pedicle dimensions required to confirm the size of the thoracic
pedicles and the ability of the surgeon to navigate down them into the body.
One must remember that in a true frontal plane radiograph of the thoracic spine,
the pedicles that are perfectly perpendicular to the x-ray beam will have the
most accurate representation of their overall dimensions. With increasing
thoracic kyphosis, the ability to adequately assess the pedicles above and below
the apex will be diminished and occasionally additional views may be helpful
in assessing pedicle dimensions in these specific cases. As an alternative, a
preoperative CT scan can be obtained with a single slice through each mid
pedicle level to check on overall isthmus dimensions. In my experience, these
CT dimensions somewhat underestimate the diameter of the screw that can be
placed since there does appear that cortical expansion occurs with aggressive
tapping and final screw placement. This will be discussed further in the tech
niques section of this report. I would caution against using a preoperative MRI
to assess pedicle dimensions as it appears to underestimate the true cortical size
due to artifacts that can occur during the acquisition process.
It is much more difficult to assess rotated pedicles, such as those located at
the concave apex of scoliosis deformities. For scoliosis cases, we rely on the
assessment of pedicle dimensions noted at proximal and distal levels away from
the apex where the vertebrae become more neutral in orientation. Routinely, if
the pedicle dimensions are adequate at these more neutral levels, the pedicles will
be accessible at the concave apical levels. Also, analyzing the convex-sided ped
icle dimensions can provide information on the corresponding concave pedicles.
The major difference is that the concave pedicles will tend to be more cortical
than the more cancellous centered convex pedicles, although they appear to be
relatively the same size at least in adolescent idiopathic scoliosis patients [16].
Thoracic Pedicle Screw Placement 189

dramroo
Free-Hand Pedicle Screw Technique
The technique of free-hand thoracic pedicle screw placement is identical to

that utilized in the lumbar spine and includes eight specific steps: (1) exposure,
(2) locating the appropriate starting point with burring of the dorsal cortex,
(3) use of the thoracic pedicle probe to navigate down the pedicle into the verte
bral body, (4) intraosseous palpation, (5) undertapping by 0.5 mm of the intended
screw diameter, (6) repalpating to confirm intraosseous borders, (7) screw place
ment, and (8) confirmation of intraosseous screw placement (table 1, fig. 18).
It is extremely important to have a well-exposed posterior spine with
minimal oozing of blood into the field. The free-hand technique requires the
identification of specific posterior element landmarks to help confirm the
appropriate starting point. In addition, the vertebral body can often bleed a fair
amount through the pedicle and thus blood loss can be an issue if multilevel
screws are placed throughout the thoracic spine. So the initial spinal dissection
must be meticulous and the pedicle tract should be liberally plugged with bone
wax to minimize oozing from the body through the pedicle. It is also recom
mended to perform a partial inferior facetectomy which serves several purposes
including to remove the cartilage off the facet joint, to improve the fusion rate,
and also to gain access to the base of the facet joint which will be utilized as
the starting point for the mid thoracic pedicle levels (T7, T8, T9).
Next a specific starting point on the posterior elements must be located for
the creation of a cortical defect which should be at the superficial base of the
pedicle. We have discovered that there are specific starting points for each
thoracic pedicle between T1 and T12 that are fairly consistent for patients with
or without a coronal and/or sagittal spinal deformity present [11]. Beginning
distal at T12, the starting point is the down slope of the medial portion of the
bisected transverse process where it meets the lamina which should be just at or
lateral to the lateral portion of the pars. And as one proceeds to the T11 and T10
spinal levels, the starting point moves slightly more proximal and more medial
such that by T10 the starting point is the proximal edge of the transverse process
where it meets the lamina and facet. The starting points for T7, T8 and T9 are
fairly consistent being the proximal portion of the transverse process where it
meets the facet joint, just lateral to the mid portion of the superior facet. Then as
one proceeds proximal from T7, the starting point becomes more lateral and dis
tal, based off the anatomic landmarks of the thoracic transverse process-lamina
junction. Thus at T6, similar to T10, the starting point is the proximal edge of the
transverse process base at the junction of the lamina. Then by T3, the starting
point is the bisected transverse process where the transverse process meets the
lamina. This also holds true for T1 and T2, which are slightly larger pedicles, but
have a greater medial inclination in the transverse plane [13].
Lenke 190
dramroo
Table 1. 8 steps of the free-hand technique of thoracic pedicle screw placement
Step Description
Exposure (see fig. 1) Thorough and meticulous exposure to the tips of all transverse processes to
be included in the instrumentation/fusion.
Starting point and cortical Visualize the starting point based upon as much anatomical information as
burr (see fig. 2) possible. For noncontiguous levels, this may be limited to a review of the
pre- and intraoperative radiographs and orientation of the posterior elements.
However, with successive levels, much information is provided by making
fine adjustments to the trajectory of the previous levels screw or contralateral
screw. Use a 5.0-mm acorn-tipped burr (or smaller burr in smaller patients) to
thin the posterior cortex and enter the pedicle. The pedicle blush should be
visualized suggesting entrance into the cancellous bone of the pedicle. In
cases with smaller pedicles, especially in the apical concavity of scoliosis
patients, there will be very limited intrapedicular cancellous bone and
therefore a pedicle blush may not be observed.
Pedicle gearshift-lateral Use a 2-mm blunt-tipped, slightly curved pedicle finder (thoracic gearshift)
(see fig. 3) facing laterally to enter the pedicle. Often the endosteal diameter of the
pedicle is quite small, so allowing the finder to fall into the pedicle. The
thoracic gearshift should be perpendicular to the plane of the superior facet
and/or lamina. Aiming slightly laterally initially will help to avoid medial
wall perforations. Rotate the pedicle finder to a medially faced orientation
after a depth of approximately 1520 mm is attained (the average length of a
typical pedicle). Carefully place the tip to the base of the prior hole before
advancing the pedicle finder. Advance the finder to approximately the length
of the desired screw and then rotate the finder 180 to make room for the
screw. Make sure you feel bone the entire length of the pedicle. Any sudden
advancement of the gearshift suggests penetration into soft tissue and thus a
pedicle wall violation or vertebral body violation. These should be
investigated immediately in order to possibly salvage the pedicle and avoid
complications.
Pedicle palpation Use a flexible ball-tipped sounding device to palpate all 5 walls of the pedicle
(see fig. 4) (cephalad, caudad, medial, lateral, and floor). Pay special attention to the
junction of the middle and upper portions of the tract as this is the region of
the pedicle where the spinal canal is located. If any wall besides the medial
wall has been breached, the pedicle may be salvageable. In this circumstance,
place bone wax in the pedicle hole to limit bleeding and reangle the pedicle
finder with a more appropriate trajectory.
Pedicle tapping Tap the pedicle with a tap that is 0.5 mm smaller than the proposed screw.
(see fig. 5) If there is difficulty passing the tap, use the next smaller tap and then retap
the pedicle. If there is no anterior wall violation, one can use a K wire to
assist with guiding the path of the cannulated taps. If there is any question of
whether the anterior wall is intact, do not use a K wire as cardiac tamponade
due to K-wire-induced trauma to a coronary artery has been reported.

dramroo
Table 1 (continued)
Step Description
Repeat pedicle palpation Use the palpating device a second time to assess the bony
(see fig. 6) pedicle walls and remeasure the tract length with a hemostat. Compare this
measurement directly adjacent to the screw to be placed to ensure appropriate
screw length.
Screw placement Place the screw slowly to confirm it is threaded properly and
(see fig. 7) allow for viscoelastic expansion of the pedicle. Make sure the angle of screw
insertion matches the tract previously palpated and tapped.
Confirmation of After all the screws have been placed, intraoperative
intraosseous screw radiographs are repeated prior to placing the rods in order to
placement (see fig. 8) confirm accurate placement of the screws. Some surgeons prefer to use
fluoroscopy in order to obtain a true AP of each level. The coronal plane
should show harmonious screw positions from proximal to distal. The sagittal
plane should demonstrate parallel orientation to the superior end-plate so that
no screw tip is anterior to the anterior vertebral body line. Perform EMG
assessment of the screws based upon rectus abdominis data. Frequently we use
active EMGs via the same approach through the pedicle finder or cannulated
tap to assess real-time monitoring of the thoracic nerve root. The screws with
the lowest impedance are removed and rechecked with a ball-tipped sounder
prior to replacement of the screw if intraosseous borders are confirmed.
Fig. 1. Complete exposure of the bony anatomy.
Lenke 192
dramroo
T1 Thoracic Pedicle Screw Starting Points
T2 Cephalad-Caudad Medial-Lateral
Level
Starting Point Starting Point
T3
T1 Midpoint TP Junction: TP-Lamina
T4
T5
T6 Junction: Proximal
T4 Junction: TP-Lamina
third-Midpoint TP
T5 Proximal third TP Junction: TP-Lamina

T7
Junction: Proximal
Junction: TP-Lamina-
T6 edge-Proximal third
Facet
TP
T8
T7 Proximal TP Midpoint Facet

T9

T10
Junction: Proximal
Junction: TP-Lamina-
T11 T10 edge-Proximal third
Facet
TP
Just medial to lateral
T11 Proximal third TP
pars
T12
At the level of lateral

T12 Midpoint TP
pars
Fig. 2. Thoracic pedicle screw starting points.
A specially designed thoracic pedicle probe (Lenke Probe, Medtronic

Sofamor-Danek Memphis, Tenn., USA) helps probe down the small-
sized pedicle channel into the vertebral body. The pedicle probe has a 2-mm
tip, which will be required for navigating smaller pedicles versus a conven
tional 4- to 5-mm tip used as a lumbar pedicle probe. The probe tip becomes
thicker as one proceeds away from the tip of the probe, which will make the
probe somewhat snugger as one proceeds down into the body past the pedicle.
It is also calibrated in 5-mm increments with a rectangular tip going to a
round shaft at 35 mm in length. The probe is also shaped with a slight

dramroo
3
Fig. 3. Use a 2-mm blunt-tipped, slightly curved pedicle finder (thoracic gearshift)
facing laterally to enter the pedicle.
Fig. 4. Sounding device to palpate all 5 walls of the pedicle (cephalad, caudad, medial,
lateral, and floor).
Fig. 5. Tap the pedicle with a tap that is 0.5 mm smaller than the proposed screw.
Lenke 194
dramroo
6
Fig. 6. Use the palpating device a second time to assess the bony pedicle walls.
Fig. 7. Place the screw slowly.
Fig. 8. Intraoperative x-rays.

dramroo
curvature, which is utilized to avoid penetration of the spinal canal at all times
[17]. We almost always start our probe with the curve facing lateral for this
purpose.
After a cortical window is drilled with a 4.0-mm burr, the gearshift is taken
and placed perpendicular to the plane of the lamina at the level being instru
mented. This perpendicularity is an extremely important feature for obtaining
both correct sagittal plane and transverse plane angulation to allow placement
of the screw parallel to the superior end-plate of the instrumented level.
The pedicle probe tip starts out with the curve facing lateral in the pedicle tract.
A soft spot of the cancellous pedicle is identified with an initial light push of
the probe tip into the burred surface. Once this soft spot is acknowledged by
feel, then the probe is advanced deeper into the pedicle past the spinal canal.
Once the pedicle probe has been engaged down to 1520 mm (i.e. past the
spinal canal and into the posterior portion of the body), the probe is removed,
and turned in a medially curved direction and the probe is then advanced
medially into the vertebral body. If the probe was left continually in the lateral
curved position, it would deflect out of the vertebra laterally, which is less than
desirable. It is important as one is advancing the probe to rotate it slightly so as
to allow the tip to fall into the soft portion of the pedicle and then into the body
quite readily. If resistance is being met, then the probe tip is usually positioned
either too medial or too lateral. One should make multiple small permutations
of probe tip positioning and orientation to seek out the best fit of the pedicle
probe down the shaft of the pedicle without breaking out of the pedicle.
Certainly, this feel is an extremely important component to the free-hand
screw placement technique. Thus, one should ideally have mastered this in the
lumbar spine before proceeding up to the thoracic spine where medial penetra
tion may render the thoracic spinal cord at risk during probe placement and/or
screw placement.
After the probe has been advanced to a depth between 30 and 35 mm
(proximal thoracic region) or up to 4045 mm (lower thoracic region), the probe
is removed and the pedicle tract palpated with a fine ball-tipped palpation
device. It is absolutely essential to feel the floor and four bony walls (medial,
lateral, superior, and inferior) confirming intraosseous placement. One must
carefully rule out a ledge that will be indicative of a medial wall breech that will
be present in the proximal to mid third of the pedicle tract where the spinal
canal is located. A tip-off of a medial breach is fairly aggressive bony oozing
of venous blood from the pedicle tract out of proportion to other levels.
Although this bleeding can be indicative of intraosseous bleeding, it may also
signal epidural bleeding secondary to a medial pedicle wall defect. Liberal use
of bone wax will plug the hole between the various maneuvers before the screw
is placed, if intraosseous borders are confirmed.
Lenke 196
dramroo
Next, the pedicle tract is tapped with a tap undersized from the final
pedicle screw by 0.5 mm. I use a variety of color-coded cannulated taps that
range in diameter from 4.0 up to 6.5 mm with the threaded portion 35 mm in
length (Medtronic Sofamor-Danek). If necessary, a short K wire can be placed
into the pedicle tract as long as there is a floor for the K wire not to advance
through while tapping is undertaken. The K wire can be especially helpful if
more than one attempt is made to probe down the pedicle and a false passage has
been created. A K wire leading down the correct pathway can guide the tap into
the appropriate position. We have also found clinically, and confirmed biome
chanically, that undertapping by 0.5 mm will improve the screw ultimate pullout
strength. It is not necessary to tap the entire length of the pedicle screw tract,
only the portion deep into the isthmus of the pedicle usually meaning approxi
mately 2025 mm.
Following tapping, the pedicle is palpated again with a fine ball-tipped
sounding probe. Once again the floor and four walls (medial, lateral, superior
and inferior) should be confirmed intact to be completely intraosseous. There
is often a very nice cortical bony ridge that is palpated following the tapping
that absolutely confirms an intraosseous screw tract. It is important at this time
to confirm the appropriate length of the screw by measuring the length of the
sounding device while it is held deep against the floor of the vertebral body. We
will place a screw of a slightly smaller length so as not to penetrate anterior or
lateral to the vertebral body with an inadvertently long screw. The screw is then
placed in the same orientation as the pedicle tract was probed and tapped.
If the initial pedicle probe tract is found to be either medial or lateral, it
often can be redirected with the probe down the appropriate pedicle shaft. Even
with a medial wall defect, a new medial wall can be created with a pedicle
probe placed correctly down the pathway of the pedicle. However, this is not
always the case and in very tiny pedicles, often only one pass will be allowed
to successfully navigate the pedicle. Almost always if the initial pedicle probe
is lateral, the true pedicle pathway can be salvaged with a more medially
directed probe. Similar to the lumbar spine, if one is having difficulty finding
the start of the pedicle, it is best to err slightly more lateral. If one does exit
lateral initially, virtually no structures are at risk from minor lateral wall
penetrations, and the correct pedicle tract can almost always be salvaged in
these circumstances.
Confirmation of Intraosseous Screw Placement
Following screw placement, it is imperative that the surgeon confirm accu

rate screw placement. This is done intraoperatively by several methods including

dramroo
previously mentioned tract palpation, visual orientation and line-up of the screws,
intraoperative AP and lateral radiographs and/or fluoroscopy, and pedicle screw
stimulation with EMG recordings [18]. The use of CT scans postoperatively to
confirm appropriate screw position is strongly recommended to those beginning
their experience [19].
Probably the more important intraoperative confirmation of an intraosseous
screw is the palpation method. This is utilized twice: initially after the probe is
in place and then following tapping of the pedicle. Once the screw is placed,
intraoperative coronal and sagittal plane radiographs or fluoroscopy should be
used to check all screws. In the frontal plane, the screws should be fairly har
monious with a lack of crossing of the screw tips in the midline. Although this
does not absolutely predict medially directed screws, it certainly heightens ones
suspicion. Also, lateral placement needs to be checked because this is fairly
common in the thoracic spine especially since the vertebral bodies thin out lat
erally in mid and proximal thoracic vertebral levels. In the lateral view, I prefer
to have the screw parallel to the superior end-plate. And also, the screw tips
should not extend past the anterior border of the vertebral body so as to make
sure the screws are not excessively long. It is important to confirm that the top
and bottom screws are not penetrating the superior end-plate extending into the
adjacent disc space. This is especially true for the most proximal level thoracic
screw, as it is often difficult to angle ones pedicle probe, tap and screw in the
appropriate kyphotic angulation of the proximal thoracic region, with resultant
cephalad positioning.
We perform thoracic pedicle screw stimulation with EMGs recorded
from the rectus abdominis muscles to confirm intraosseous screw placement
[20]. An identical technique is used for screws placed between L1 and S1
with recording in the various lower extremity myotomes. Screws are stimu
lated with an ascending method of simulation to obtain a compound muscle
action potential from the rectus abdominis muscles bilaterally. A recent study
of 677 screws placed between T6 and T12 (innervation of the rectus) pre
dicted a medially placed screw by a combination of threshold response less
than 6 mA in absolute value and also a threshold intensity averaging 65%
below the mean of all other thoracic screw responses in the same patient.
Although this is not absolutely indicative for a medially misplaced screw, it is
another method utilized to heighten ones suspicion of a pedicle wall defect
intraoperatively.
It also is advisable for surgeons beginning with the technique of thoracic
pedicle screw placement to obtain postoperative CT scans of their patients [19].
It can be quite enlightening to see exactly where the screws are placed, where
the tips of the screws are located, and it can only help the surgeon improve his
or her technique over time.
Lenke 198
dramroo
Results and Correlations of Thoracic Pedicle Screws
Placed by the Free-Hand Technique
All thoracic pedicle screws placed by the free-hand placement technique

have been reviewed by an independent spine surgeon performing a 2-year clin
ical research fellowship (Y.J. Kim, MD, Seoul, Korea) [21]. 2,199 thoracic
pedicle screws have been placed by this technique in the last 9 years, the vast
majority of which over the last 4 years. The majority of screws have been
placed in the treatment of pediatric and adult spinal deformity, but also include
use in nondeformity conditions such as spinal trauma, tumor and infection.
Our results in spinal deformity treatment include increasing correction rates
with a decreased use of preliminary anterior release and fusions in some cir
cumstances, and the rare use of any postoperative external immobilization
(fig. 9).
There were no screws removed for any type of neurologic, vascular or
visceral complications [11, 21]. Specifically, there have been no screw-related
neurologic deficits, unexplained thoracic chest wall pain or radicular lower
extremity pain, or unexplained neurologic symptoms. There have been no cases
of revision surgery needed up to this stage in the cases with thoracic pedicle
9a 9b
(For legend see p. 201.)

dramroo
9c 9d
9e 9f
Lenke 200
dramroo
9g 9h
Fig. 9. a ER is a 14 11-year-old female who presented with a severe 113 right

thoracic idiopathic scoliosis. Proximal thoracic curve measured 46 and lumbar compen
satory curve 52. b Her sagittal plane showed 43 of thoracic kyphosis between T5 and T12
and 80 of lumbar lordosis between T12 and the sacrum. c Her left side bender showed a
structural proximal thoracic curve bending out to only 35, with a flexible lumbar curve
bending out to 17. d Her right side bender showed a stiff main thoracic region as expected
with correction to only 84. e She underwent an open anterior release and fusion from T5 to
T12 and a posterior segmental screw instrumentation and fusion from T3 to L3 and in a
staged fashion. Her upright postoperative coronal radiograph demonstrates excellent correc
tion of her main thoracic curve to 36, with well balanced curves both above and below. Note
the marked apical translation occurring with use of multisegmental screw instrumentation.
f Her upright lateral x-ray shows normalized thoracic kyphosis of 33 and overall good sagit
tal balance. g Her preoperative upright photograph demonstrates her right thoracic trunk shift
and rib prominence on the right side. h Her postoperative clinical photo demonstrates excel
lent balance of her trunk on her pelvis and level shoulders with a diminished rib prominence.
screw constructs, except for 1 patient 2 years postoperatively from an adoles

cent idiopathic scoliosis surgery with a chronic deep wound infection treated by
implant removal. Because of these results, we have continued to utilize pedicle
screws as our main anchor attachment to the thoracic and lumbar spine in all
forms of pediatric and adult deformity and nondeformity conditions requiring
spinal instrumentation and fusion.

dramroo
Conclusions
The free-hand placement of thoracic pedicle screws in both pediatric and

adult patients with or without spinal deformity can be performed in a safe and
reliable manner. This technique should be mastered in the lumbar spine and tho
racolumbar junction prior to extending it more proximal in the mid and proxi
mal thoracic spine. Strict and meticulous attention to detail is required along
with a precise surgical technique to safely place thoracic pedicle screws for a
variety of spinal conditions.
References
1 Gaines RW: The use of pedicle-screw internal fixation for the operative treatment of spinal disor
ders. J Bone Joint Surg Am 2000;82:14581476.
2 Liljenqvist UR, Halm HF, Link TM: Pedicle screw instrumentation of the thoracic spine in idio
pathic scoliosis. Spine 1997;22:22392245.
3 Suk SI, Kim WJ, Lee SM, Kim JH, Chung ER: Thoracic pedicle screw fixation in spinal deform
ities: Are they really safe? Spine 2001;26:20492057.
4 Suk SI, Lee CK, Kim WJ, Chung YJ, Park YB: Segmental pedicle screw fixation in the treatment
of thoracic idiopathic scoliosis. Spine 1995;20:13991405.
5 Hamill CL, Lenke LG, Bridwell KH, Chapman MP: The use of pedicle screw fixation to improve
correction in the lumbar spine of patients with idiopathic scoliosis. Is it warranted? Spine 1996;
21:12411249.
6 Hirano T, Hasegawa K, Takahashi HE, Uchiyama S, Hara T, Washio T, Surgiura T, Yokaichiya M,
Ikeda M: Structural characteristics of the pedicle and its role in screw stability. Spine 1997;22:
25042510.
7 Krag MH, Weaver DL, Beynnon BD, Haugh LD: Morphometry of the thoracic and lumbar spine
related to transpedicular screw placement for surgical spinal fixation. Spine 1998;13/1:2732.
8 Liljenqvist UR, Link TM, Halm HF: Morphometric analysis of thoracic and lumbar vertebrae in
9 Lonstein JE, Denis F, Perra JH, Pinto MR, Smith MD, Winter RB: Complications associated with
pedicle screws. J Bone Joint Surg Am 1999;81:15191528.
10 Vaccaro AR, Rizzolo SJ, Balderston RA, Allardyce TJ, Garfin SR, Dolinskas C, An HS:
Placement of pedicle screws in the thoracic spine. II. An anatomical and radiographic assessment.
11 Kim Y, Lenke LG, Bridwell KH, Riew KD, Rhee JM, Hanson DS: Free hand pedicle screw place
ment in the thoracic spine (poster). Scoliosis Research Society 36th Annual Meeting, Cleveland,
2001.
12 Cinotti G, Gumina S, Ripani M, Postacchini F: Pedicle instrumentation in the thoracic spine.
A morphometric and cadaveric study for placement of screws. Spine 1999;24/2:114119.
13 Ebraheim NA, Xu R, Ahmad M, Yeasting RA: Projection of the thoracic pedicle and its morpho
metric analysis. Spine 1997;22/3:233238.
14 Kothe R, OHolleran JD, Liu W, Panjabi MM: Internal architecture of the thoracic pedicle.
An anatomic study. Spine 1996;21/3:264270.
15 Vaccaro AR, Rizzolo SJ, Allardyce TJ, Ramsey M, Salvo J, Balderston RA, Cotler JM: Placement
of pedicle screws in the thoracic spine. I. Morphometric analysis of the thoracic vertebrae. J Bone
Joint Surg Am 1995;77:11931199.
16 OBrien MF, Lenke LG, Mardjetko S, Lowe TG, Kong Y, Eck K, Smith D: Pedicle morphology in
thoracic adolescent idiopathic scoliosis: Is pedicle fixation an anatomically viable technique?
Spine 2000;25:22852293.
Lenke 202
dramroo
17 Ebraheim NA, Jabaly G, Xu R, Yeasting RA: Anatomic relations of the thoracic pedicle to the
adjacent neural structures. Spine 1997;22:15531557.
18 Brown CA, Lenke LG, Bridwell KH, Geideman WM, Hasan SA, Blanke K: Complications of
pediatric thoracolumbar and lumbar pedicle screws. Spine 1998;23:15661571.
19 Kim YJ, Lenke LG, Bridwell KH, Riew KD, OBrien M, Rhee JM, Hanson DS: CT scan accuracy
of free hand thoracic pedicle screw placement in pediatric spinal deformity (poster). Scoliosis
Research Society Annual Meeting, Cleveland, 2001.
20 Raynor BL, Lenke LG, Kim Y, Bridwell KH, Hanson DS, Padberg AM: Can triggered EMG
thresholds accurately predict thoracic pedicle screw placement? Spine, in press.
21 Kim Y, Lenke LG, Bridwell KH, Riew KD, Rhee JM, Hanson DS: Thoracic pedicle screw place
ment in deformity: Is it safe? (poster). Scoliosis Research Society 36th Annual Meeting,
Cleveland, 2001.
Lawrence G. Lenke, MD
One Barnes-Jewish Hospital Plaza, Suite 11300, St. Louis, MO 63110 (USA)

dramroo
Percutaneous Lumbar Pedicle Screws:

Indications, Technique, Results
Gerald E. Rodts, Jr.
Department of Neurosurgery, Emory University, Atlanta, Ga., USA
Lumbar pedicle fixation has proven to be a safe, effective means of lumbar

stabilization. Prospective, randomized trials and community clinical experience
using lumbar instrumentation have demonstrated an increased fusion rate.
Lumbar instrumentation has also improved the ability to mobilize patients after
surgery and perhaps brace them for shorter periods of time or not at all. To date,
techniques of lumbar pedicle screw placement have involved a single midline or
bilateral paramedian incision. Some degree of muscular tendon detachment from
the bone and muscle retraction has been necessary. Without question, this is a
source of postoperative pain and subsequent muscle atrophy and scar formation.
Until recently, techniques of percutaneous lumbar fixation have involved
external instrumentation or epifascial techniques [13]. Described herein is a
new technique to allow for direct subfascial lumbar pedicle fixation, rod inser
tion, distraction/compression, and final tightening, all with percutaneous entry
points. Compared to open techniques, the presumed benefit of the percutaneous
technique is avoidance of muscular tendon detachment, elimination of large-
scale muscle retraction, diminished postoperative pain, less blood loss, and
improved cosmetic result. In the future, improved minimally invasive fusion
techniques and the use of bone morphogenetic protein may increase the appli
cability of percutaneous techniques.
Indications
The indications for percutaneous lumbar pedicle fixation initially are

limited to a few specific clinical conditions. Currently, percutaneous lumbar
pedicle fixation may be most useful when a surgeon wishes to create a posterior
dramroo
tension band. As posterior endoscopic or other minimally invasive techniques
progress, percutaneous fixation may be combined with simultaneous laminec
tomy, discectomy, interbody or posterolateral fusion.
One indication for percutaneous posterior fixation is following a laparo
scopic or open (retroperitoneal or transperitoneal) anterior lumbar interbody
fusion (ALIF). Many surgeons do not rely on a stand-alone ALIF for the treat
ment of a mobile (unstable) lumbar spondylolisthesis. Whether threaded tita
nium cylinders, vertical carbon or titanium or ceramic cages, allograft femoral
ring wedges or threaded allograft bone dowels are used for the ALIF, the place
ment of posterior instrumentation in the setting of a mobile spondylolisthesis
provides greater biomechanical stability. Furthermore, studies have shown a
higher rate of pseudoarthrosis following stand-alone ALIF using allograft bone.
With posterior fixation, the rate of fusion following ALIF with bone-only is
improved.
Another consideration for the use of posterior percutaneous fixation is in
the case of a pseudoarthrosis following previous stand-alone ALIF. If a
nonunion is present yet the surgeon is satisfied with the structural integrity of
the previously implanted bone graft, posterior instrumentation ultimately can
help to achieve a successful arthrodesis. An example would be a case of previ
ous femoral ring ALIF with an intact graft, absence of subsidence or vertebral
body lysis, but no evidence of bone union on plain radiographs or CT with
reconstructions.
Percutaneous pedicle fixation is also a reasonable alternative to open
placement of instrumentation in the setting of a pseudoarthrosis following pre
vious posterolateral fusion. In this setting, decortication of the previous fusion
mass and placement of bone graft can be done through the same portals used
for percutaneous screw placement.
More recently, advances in less invasive techniques using tubular retractors
and blunt dissection through muscle have allowed for lumbar bone decompres
sion (laminectomy, laminotomy, medical facetectomy, discectomy), posterolat
eral fusion, and interbody fusion. Percutaneous pedicle screw fixation using the
same or different stab incisions may be a beneficial adjunct to these newer
approaches to decompression and arthrodesis.
Materials and Techniques
For percutaneous pedicle screw fixation, radiographic image guidance is essential.

Several options may be considered. Placing the instrumentation using plain radiographs is not
recommended. A single fluoroscopic C-arm can be alternated between anterior-posterior (AP),
oblique or owls eye, and straight lateral views. The owls eye view allows a coaxial view
down the barrel of the pedicle. Performing percutaneous screws with a single fluoroscopic
Percutaneous Lumbar Pedicle Screws 205

dramroo
C-arm is possible; however, it is tedious and significant time must be devoted to changing
sterile drapes and moving the C-arm into different positions.
Simultaneous biplanar fluoroscopy offers immediate feedback in two planes.
Ergonomically, it is challenging to operate inside of and around two C-arms placed for lateral
and AP or lateral and oblique views. Simultaneous dual views allow for instant feedback
when an instrument trajectory is altered manually by the surgeon in one plane. One limita
tion of biplanar fluoroscopy in percutaneous procedures, however, is the difficulty in seeing
the tip of an instrument at the level of the skin surface. This is particularly true in patients
with a large body habitus where there is a great distance between skin surface and the spine.
Computer-assisted, virtual fluoroscopic systems offer a tremendous advantage. We
have had extensive experience using the FluoroNav virtual fluoroscopic system
(Medtronic Sofamor Danek, Memphis, Tenn., USA). With this technology, a stereotactic ref
erence arc with light-emitting diodes is attached rigidly to the patients spine through a small
stab incision. The arc can also be attached to the ileum using a screw. Routine lateral, oblique
or owls eye, and AP images are obtained. The images are obtained with an overlying cali
bration array of fiducials attached to the C-arm. Images are then transferred automatically
from the C-arm monitor to the Stealth or ION (intraoperative navigation) computer
monitor. The images are automatically registered and ready for navigation. Left and right
oblique images are recommended for each level to be instrumented. Up to four images can
be displayed and monitored simultaneously on the same screen. Digitized instruments with
light-emitting diodes are then recognized by the camera and their location in the room is
thereby known by the computer. In short, a surgeon can navigate on two-dimensional images
as if the C-arm were operating live continuously. No further irradiation is necessary. Virtual
cartoons or representations of the surgical instruments move real time on the monitor as the
surgeon moves them in surgery.
Perhaps the most powerful component of the virtual fluoroscopic system is the ability
to virtually extend the tips of the various instruments from the skin surface down to the
spine. Thus, when an instrument is held on the skin surface, the tip is extended down to
recognizable radiographic landmarks that represent the entry point for pedicle fixation. This
is not possible with conventional fluoroscopy or radiography unless a wire is actually passed
through the skin and soft tissues to simulate a trajectory. The lateral view identifies the center
of the pedicle and appropriate trajectory in the sagittal plane. The owls eye or oblique view
gives the surgeon feedback in the axial plane. Straight AP images represent the coronal
plane. One helpful technique is to extend the tip of the virtual instrument on the computer
screen to the point where the pedicle enters the posterior aspect of the vertebral body
(fig. 1, 2). In the sagittal view, therefore, the instrument would have virtually probed the
entire length of the pedicle in the AP direction. When one then looks simultaneously at the
oblique view, the very tip of the extended instrument should be within the cortical confines
(margins) of the pedicle. If the tip is medial, medial wall perforation may be possible. In
short, the surgeons brain is able to imagine three-dimensional trajectories based on multiple
two-dimensional fluoroscopic images. Once the desired skin entry point is identified, a small
stab incision is made (approximately 1012 mm) (fig. 3).
In open placement of lumbar pedicle screws, slight variations of the same technique are
used to identify the entry point for pedicle fixation. It is recommended that the proximal or
medial-most aspect of the transverse process be considered for the entry point. This is as
opposed to performing a partial inferior-medial facetectomy for a more medial entry point.
For obvious reasons, one would favor a technique for percutaneous placement that does not
Rodts 206
dramroo
Fig. 1. Virtual representation of probe extending through skin down to the level of
vertebral body (AP and lateral views) (photo courtesy of Kevin Foley, MD).
Fig. 2. Schematic drawing of extend

ing virtual tip of probe beyond length of
pedicle to help stay within confines of cortex.
Fig. 3. Reference arc and digitized

probe just beneath skin surface.
require bone removal. Another advantage of using the medial transverse process is that the
surgeon can palpate with a wire or other image-guided probe the superior and inferior edges
of the transverse process at the same time that visual feedback is given from the AP virtual
fluoroscopic view. One can then feel the slight groove where the transverse process meets the
superior facet process.

dramroo
4 5
Fig. 4. Percutaneous placement of cannulated pedicle screw with screw extender (over
K wire).
Fig. 5. Pedicle screw extenders.
Using an image-guided drill guide, a K wire is drilled into the pedicle approximately
20 mm, allowing for entry into the posterior vertebral body. A cannulated pedicle probe can
then be passed or a cannulated drill bit can be used to drill the pedicle. A cannulated tap is
passed (usually) to the depth of the posterior vertebral body (fig. 4). Patients with sclerotic
or very firm bone may require tapping the full length of the intended screw. The pedicle is
now ready for screw placement.
An important part of the percutaneous instrumentation is the screw extender. A multi
axial screw (M8, Medtronic Sofamor Danek) is attached to a long arm that will cross the dis
tance between the posterior pedicle and the skin edge (fig. 5). A locking screw is preloaded
into the chamber of the screw extender. The outer end of the screw extender will attach to the
Sextant device.
For L5S1 fixation, a single longer stab incision can be made to accommodate place
ment of both the L5 and S1 screws and screw extenders. For L4L5 and higher levels, sepa
rate stab incisions will be necessary. Thus, once one screw has been placed, the sequence of
steps is repeated for placement of the second screw. When both screws have been placed, the
ends of the screw extenders are snapped together. The Sextant device with detachable trocar
is then attached to the end of the screw extender complex (fig. 6). The arm of the Sextant is
rotated down to the skin surface. This site is identified and a small stab incision is made. The
trocar is passed through the adipose layer, fascia, and bluntly through the muscle to the heads
of the multiaxial screws (fig. 7, 8). The arm is rotated back out. This step helps prepare for
Rodts 208
dramroo
6 7
Fig. 6. Sextant device with rod attached and screw extenders prior to percutaneous
delivery.
Fig. 7. Fluoroscopic view of rod inserted with tip of Sextant device visible inferiorly.
Fig. 8. Sextant device at moment of

final rod engagement.
passage of the rod. The trocar is removed and a lordotic rod is attached to the arm of the
Sextant. The rod is delivered through the same fixed arc into the side openings of the multi
axial screws. One additional real-time fluoroscopic image is recommended to ensure proper
placement of the rod, though the fixed geometry of the screw extender Sextant system pro
vides for consistent accurate placement. The rod is detached from the delivery arm, and the
arm removed. The locking screws previously loaded into the screw extender chambers are
tightened to the appropriate torque tension with a screwdriver. Then, the complete apparatus

dramroo
9 10
Fig. 9. Postoperative scars from unilateral percutaneous pedicle screw-rod fixation.

Fig. 10. L5S1 ALIF with allograft femur wedge and unilateral percutaneous pedicle
fixation, 14 months postoperatively.
is pulled out of the three (or if L5S1, two) stab incisions. Subcuticular closure is then
possible (fig. 9). The entire process can be repeated on the opposite side, though some
surgeons may perform unilateral fixation when the biomechanical goal is to provide a
posterior tension band following ALIF (fig. 10).
Results
The preliminary results of this technique confirm that it is a safe, effective

way to place lumbar pedicle screws and rods. Lefkowitz and Foley [4] pre
sented their results for 10 patients with spondylolisthesis. Patients underwent
ALIF followed by posterior percutaneous pedicle screw/rod fixation. There
were no complications related to the percutaneous technique. The average total
operating time for both the ALIF and the percutaneous screw placement in this
initial series was 7.3 h. Blood loss was markedly decreased compared to open
techniques. The average hospital stay was 2.9 days.
Nockels et al. [5] presented their experience in 15 patients with a wide range
of pathology. A total of 64 screws with rods were placed. Two cases of a two-level
instrumentation were done. Patients were evaluated postoperatively with plain
radiography and CT scan. Most cases were performed in combination with an
ALIF. The average time for the percutaneous placement of lumbar screws with
Rodts 210
dramroo
rods was 65 min. No complications were reported due to the percutaneous
technique. There were no cases of implant failure during the initial follow-up.
Lefkowitz et al. [6] also reported on their experience using percutaneous
pedicle screw fixation following minimally invasive posterior lumbar interbody
fusion (PLIF). This latter technique was performed through endoscopic tubular
retractors but with the use of the operating microscope or surgical loupes. Six
patients underwent this technique. The indications were spondylolisthesis or
degenerative disc disease. One dural tear was reported but was not related to the
screw technique (it occurred during the PLIF technique). The average operating
time for the entire procedure was 6 h 45 min. The average blood loss was
183 ml. The average duration of hospitalization was 2.3 days.
Rodts [7] presented preliminary data of 5 patients who underwent unilat
eral posterior percutaneous pedicle screw fixation following ALIF. The indica
tion for surgery was spondylolisthesis in 4 patients and recurrent disc
herniation in a fifth. All patients received unilateral instrumentation alone.
Follow-up ranged from 10 to 22 months. Solid arthrodesis has been achieved in
all patients. Postoperative CT scans were performed showing satisfactory screw
placement in all screws except one. One S1 screw had 3 mm of lateral cortical
perforation of no clinical significance.
Conclusion
Lumbar instrumentation has proven to be a useful component of current

lumbar fusion surgery. The advantages include higher rates of arthrodesis, pro
vision of immediate spinal stability, and the ability to more effectively reduce
or correct deformity [811]. For patients who require extensive posterior
decompression with fusion, open placement of pedicle instrumentation remains
a useful technique.
In some patients, however, the goals of arthrodesis may be achieved with an
ALIF, a minimally invasive PLIF, or a minimally invasive posterolateral fusion
technique. In cases of mobile spondylolisthesis, previous surgery (and subopti
mal fusion bed, vascular supply), metabolic bone disease, or smoking, a surgeon
may wish to supplement the ALIF or PLIF arthrodesis with instrumentation. In
this setting, the percutaneous technique may offer an advantage over open tech
niques. Several very small stab incisions may be cosmetically superior to one
longer midline or paramedian incision. The paraspinal musculature is spared the
ill-effects of midline tendinous release and wide retraction causing ischemia and
subsequent atrophy and scar formation. Blood loss is markedly less compared to
open techniques. In their initial experience, surgeons using the percutaneous
pedicle screw technique have found that their patients have substantially less
back pain in the postoperative period.

dramroo
The key component to safe, effective passage of percutaneous instrumenta
tion is image guidance. Whether a surgeon chooses to use a single or biplane live
fluoroscopy or computer-assisted virtual fluoroscopy, the proper radiographic
identification of important landmarks is critical. Virtual fluoroscopy offers sev
eral advantages, including the ability to view four images at once (e.g. right and
left oblique, AP, lateral), and the ability to virtually extend various instruments
from the surface of the skin down to the spine on the radiographic images.
Further experience with percutaneous and minimally invasive techniques
will not only lead to less postoperative pain, smaller incisions and soft tissue
trauma, shorter operating time, and less blood loss, but it will also enable the
surgeon to achieve the same desired result as in open procedures: successful
arthrodesis. With the advent of new biological products such as bone morpho
genetic protein, and with the refinement of interbody fusion implants, the role
of minimally invasive surgery may expand.
References
1 Wiesner L, Kothe R, Ruther W: Anatomic evaluation of two different techniques for the percuta
neous insertion of pedicle screws in the lumbar spine. Spine 1999;24:15991603.
2 Lowery GL, Kulkarni SS: Posterior percutaneous spine instrumentation. Eur Spine J 2000;9:
S126S130.
3 Magerl F: External skeletal fixation of the lower thoracic and the lumbar spine; in Uhtoff HK,
Stahl E (eds): Current Concepts of External Fixation of Fractures. Berlin, Springer, 1982,
pp 353366.
4 Lefkowitz M, Foley KT: Percutaneous pedicle fixation for spondylolisthesis. Annual Meeting of
the AANS/CNS Joint Section on Disorders of the Spine and Peripheral Nerves, Orlando, 2002.
5 Nockels R, et al: Percutaneous pedicle screw fixation. Annual Meeting of the AANS/CNS Joint
Section on Disorders of the Spine and Peripheral Nerves, Orlando, 2002.
6 Lefkowitz M, Palmer S, Foley KT: Percutaneous pedicle fixation following minimally-invasive
PLIF. 2002 Annual Meeting of the AANS/CNS Joint Section on Disorders of the Spine and
Peripheral Nerves, Orlando, 2002.
7 Rodts GE: Unilateral percutaneous lumbar pedicle screw/rod fixation following ALIF. Global
Spine Meeting, Mexico, 2002.
8 Tribus CB, Belanger TA, Zdeblick TA: The effect of operative position and short-segment fusion
on maintenance of sagittal alignment of the lumbar spine. Spine 1999;24/1:5861.
9 Zuckerman JF, Zdeblick TA, Bailey SA, et al: Instrumented laparoscopic spinal fusion:
Preliminary results. Spine 1995;20:20292034.
10 Yuan HA, Garfin SG, Dickman CA, Mardjetko SM: A historical cohort study of pedicle screw
fixation in thoracic, lumbar and sacral spine fusions. Spine 1999;19(suppl 20):2279S2296S.
11 Foley KT, Gupta SK: Percutaneous pedicle screw fixation of the lumbar spine: Preliminary
results. J Neurosurg 2002;97:712.
Gerald E. Rodts, Jr., MD

Department of Neurosurgery, Emory University
550 Peachtree St., NE, Suite 806, Atlanta, GA 30308 (USA)
Rodts 212
dramroo
Thoracolumbar Deformity Advances

1. Nonoperative Treatment of Thoracolumbar Deformity
Praveen V. Mummanenia, Stephen L. Ondrab, Rick C. Sassoc

a
b
Department of Neurosurgery, Northwestern University, Chicago, Ill., and
c
The terms scoliosis, kyphosis, and lordosis were first coined by the Greek
physician Galen in the second century AD [1]. Since that time, significant
advances have occurred in the classification and management of patients with
spinal deformities.
Early physicians attempted to correct thoracolumbar deformities with non-
operative treatments. Hippocrates, and later Galen, unsuccessfully used longitu
dinal traction to try to pull the deformed spine back into alignment [2]. Ambrose
Pare was the first physician to use an orthosis to brace a scoliotic patient
(approximately 1,500 AD). Pare soon realized that bracing was not useful once
a patient had reached skeletal maturity [1].
During the past several centuries, more sophisticated and effective non-
operative treatment modalities for spinal deformity have been developed.
Evaluation of Deformity
There are multiple causes of thoracolumbar deformity; a full discussion

of all of these entities is beyond the scope of this chapter. We will limit our dis
cussion to the most frequently encountered entities: adolescent idiopathic scoli
osis, adult scoliosis, and thoracic hyperkyphosis. When the surgeon encounters
any of these entities, the first step in the treatment paradigm is to evaluate the
patient fully.
dramroo
Anterior superior iliac spine
Medial malleolus
Umbilicus
Fig. 1. Measurement of leg length from the umbilicus or the anterior superior iliac
spine to the medial malleolus.
History
The patients medical and family history should be documented.
Documentation of the age of onset of the deformity and the rate of progression
of the deformity are important factors that may influence treatment. Since the
rate of curve progression often accelerates after puberty, the age of onset of
menarche should also be noted. In addition, since spinal deformities can be
inherited, it is important to identify all relatives with deformity and to attempt
to classify the type of deformity and the rate of progression of the deformity in
family members.
Physical Examination
Initially, a global survey is performed to evaluate overall spinal balance.
Obvious curvatures are noted. Pelvic obliquities and abnormal skin folds are
often seen with moderate to severe scoliotic curvatures. In the pediatric popu
lation, secondary sex characteristics are also noted. Additionally, in scoliotic
patients, leg lengths are measured and discrepancies are noted. Leg lengths can
be measured either from the umbilicus or the anterior superior iliac spine to the
medial malleolus (fig. 1). In addition, overall coronal balance can be quantified
in the upright position by dropping a plumb line from the C7 spinous process
(fig. 2). Normally the line should extend down the intergluteal crease.
Patients with scoliotic curvatures may be in overall coronal balance if their
secondary curves are of sufficient magnitude to recenter C7 over the inter-
gluteal crease in the coronal plane. Patients are asked to perform side bends to
assess if their curves are flexible or rigid.
Mummaneni/Ondra/Sasso 214
dramroo
Fig. 2. C7 plumb line. Normally line
should fall into intergluteal crease indicating
coronal plane balance. The plumb line in this
illustration falls to the right of the intergluteal
crease indicating coronal plane imbalance.
In order to further evaluate scoliosis, the patient is asked to stand upright

with their palms opposed in the praying position and with their arms extended
perpendicular to their torso. The patient is asked to bend forward at the waist
until the shoulders and hips are in the same axial plane. The physician then
views the patient from both the front and the back noting any spinal curvature
or asymmetric rib/flank elevation. For lumbar curves, more forward flexion is
required to bring out the maximal asymmetry; for thoracic curves, less forward
flexion is required to bring out the maximal asymmetry. Some physicians elect
to use an inclinometer to quantify their patients scoliotic curvature. The patient
is maneuvered to a flexed position where the asymmetry is most striking before
the inclinometer is used.
Patients with kyphotic deformities are observed in the neutral position to
see if they are able to look straight ahead when facing forward, and to assess for
compensatory extension at the hips and flexion at the knees. In addition, these
patients are asked to perform the forward bending test, which makes thora
columbar apex kyphosis obvious. Finally, these patients are asked to extend and
flex their spines at the apex of the curvature to check for curve flexibility.
Nonoperative Treatment of Thoracolumbar Deformity 215

dramroo
Fig. 3. Postural and structural thoracic kyphosis. The illustration on the left shows a
patient with a postural kyphosis. Note the rounded, smooth curve. The illustration of the patient
on the right has a structural kyphosis with a pronounced gibbus in the midthoracic spine.
Kyphotic curves are classified as either postural or structural. Kyphotic

curves that correct in extension are classified as postural. Postural kyphotic
curves are smooth curves and are usually due to poor posture and not to a
pathological process. Structural kyphotic curves, on the other hand, do not
correct in extension, often have a sharp, angular gibbus at the apex of the curve
on forward flexion, and are due to a pathological process (fig. 3).
Radiographic Evaluation
Standard radiographic evaluation for deformity begins with standing full
length (36 inch) posterior-anterior (P/A) and lateral spine x-rays. In patients
who cannot stand, recumbent x-rays are acceptable. The flexibility of the
patients scoliotic curve is evaluated with recumbent side-bending x-rays. Side-
bending x-rays are taken in the recumbent position in order to avoid locking the
facet joints during the bending motion. In patients who do not have adequate
muscle mass to perform side bends, a manual push may be administered during
the x-ray.
Scoliosis x-rays differ from nonscoliosis x-rays in several ways. First of all,
they are shot in the P/A direction in order to minimize the radiation dose to the
dramroo
5
2 3
4
1
Fig. 4. The Risser classification (grades 05) is used to quantify skeletal maturity.
Risser grade 1 patients have ossified only 25% of their iliac apophysis. Risser grade
4 patients have ossified 100% of their iliac apophysis, but have not fused their iliac wing to
their ileum. Risser grade 5 patients are skeletally mature and have ossified their entire iliac
apophysis and fused their iliac wing to the ileum.
breast tissue in adolescent girls. Secondly, scoliosis x-rays are placed on the
x-ray viewer in the opposite orientation from other x-rays. The reader looks at
scoliosis x-rays as if he is looking at the patients back, i.e. the right side of the
x-ray is on the right side of the viewer.
In younger patients, special attention is paid to the pelvis on the P/A x-ray
to assess skeletal maturity. In 1958, Risser [3] noted that the growth of the ver
tebral body endplate parallels the ossification of the iliac apophysis. The Risser
classification (grades 05) is used to quantify skeletal maturity. Risser grade
1 patients have ossified only 25% of their iliac apophysis. Risser grade 4 patients
have ossified 100% of their iliac apophysis but have not fused their iliac wing
to their ileum. Risser grade 5 patients are skeletally mature and have ossified
their entire iliac apophysis and fused their iliac wing to the ileum (fig. 4).
Furthermore, the pelvic x-ray is examined to see if the triradiate acetabular
cartilage is still open (indicating skeletal immaturity).
The lateral x-rays are inspected to assess overall sagittal balance.
Normally, a plumb line dropped from the posterior vertebral body of C7 should
pass through a point within 2.5 cm of the posterosuperior corner of S1. If the
plumb line passes more than 2.5 cm anterior to this point, then the patient is
noted to have a positive sagittal imbalance. Likewise, if the plumb line passes
more than 2.5 cm posterior to this point, then the patient is noted have a nega
tive sagittal imbalance.
The P/A x-rays are evaluated to assess coronal balance and rotatory defor
mity. Overall coronal balance is assessed by dropping a plumb line from the C7
spinous process. Normally, this line should pass through the midline of the sacrum.

dramroo
Fig. 5. Coronal, sagittal, and axial illustrations of a scoliotic deformity. Note that this
single right thoracic curve has a significant rotatory component.
When there is a loss of coronal balance, as seen occasionally with scolio

sis, the head and shoulders are no longer centered over the pelvis. In order to
compensate, the patient often tries to bend the spine above or below the primary
curve in the opposite direction. Over time, a fixed, secondary scoliotic curve
can develop as a result of this bending. The apex of the secondary curve is
opposite that of the primary curve, and the magnitude of the secondary curve
is smaller than that of the primary curve.
Rotatory deformities result in loss of axial balance and often complicate
scoliotic curvatures. They transmit loads unevenly between successive vertebral
segments (fig. 5). Rotatory deformity can be evaluated on the P/A x-rays by
assessing the symmetry in the pedicle positions at each level. Nash and Moe [4]
classified the rotatory deformity by grading it from zero (no rotation) to grade
IV (only one pedicle shadow is visible and it is rotated past the center of the
vertebral body).
dramroo
Convex side Concave side
T11
Wide
En
dv
Wide er
T1 tebra
2
L1
Wide Cobb
angle
L2
Wide Cobb
angle
L3
Wide
L4
bra
verte
End Wide
L5
Fig. 6. Measurement of the Cobb angle. End vertebrae are the last levels that are tilted
into the curve concavity. A line is drawn from the superior endplate of the superior end ver
tebral body, and another line is drawn from the inferior endplate of the inferior vertebral
body. The angle of the intersection of these lines is the Cobb angle.
Curve Measurement
Scoliosis x-rays are evaluated to establish the number of curves, the location
of the curve(s), the direction of the curve(s), and the magnitude of the curve(s).
When multiple scoliotic curves are present, the major curve is defined as the
largest, rigid curve. Minor curves are compensatory curves that are created to
try to return to overall spinal balance. Minor curves are smaller than the major
curve and usually are flexible on side-bending x-rays.
The magnitude of each coronal plane curve is quantified through the Cobb
angle measurement. To measure the Cobb angle, the physician identifies the
end vertebrae of the curve on the P/A x-ray. End vertebrae are the last levels
that are tilted into the curve concavity. A line is drawn from the superior end-
plate of the superior end vertebral body, and another line is drawn from the infe
rior endplate of the inferior vertebral body. The angle of the intersection of
these lines is the Cobb angle (fig. 6).

dramroo
Indications for Treatment and Nonoperative Treatment Modalities
Adolescent Idiopathic Scoliosis

Adolescent idiopathic scoliosis is the most common form of scoliosis seen
in the United States. School screening programs are effectively screening for
this disorder, and the prevalence of scoliosis in the school age population is
estimated to be 13% [5, 6]. Most patients with adolescent idiopathic scoliosis
are female with a right thoracic curve. Patients with atypical thoracic curve
patterns (i.e. left thoracic curve) also should be evaluated with spinal axis MRI
to rule out other causes of adolescent scoliosis (i.e. tumor).
Indications for treatment are dependent on the initial Cobb angle, the
skeletal maturity of the patient, and the progression of the curve. Skeletally
immature patients (Risser score of 03) with curves less than 25 usually are
not treated and are followed with serial x-rays at 4- to 6-month intervals to
check for curve progression.
If the curve progresses to greater than 30 (but less than 40), then nonop
erative treatment with bracing is initiated [7]. For curves with their apex above
T6, a cervical extension may be needed. Otherwise, underarm, rigid, thora
columbar braces are satisfactory. Bracing requires cooperation and compliance
from the patient and the family. Braces should be worn daily for over 18 h per
day until skeletal maturity.
For braced patients, follow-up radiographs are obtained at 4- to 6-month
intervals until skeletal maturity is reached. If the curve remains under 40 and
the patient reaches the end of growth (Risser 4 status and at least 18 months
since menarche), then the brace is discontinued. If the curve progresses beyond
40, then surgical treatment is indicated [8, 9].
Adult Scoliosis
Adult scoliosis has two primary etiologies: adolescent idiopathic scoliosis
(which has progressed after skeletal maturity) and degenerative adult-onset
(de novo) scoliosis.
Adult Patients with Progressive Adolescent Idiopathic Scoliosis

Studies have shown that adolescent idiopathic scoliosis can progress
after skeletal maturity, especially if the magnitude of the curve is greater than
30 [8, 10]. Consequently, it is important to continue to periodically follow
patients with adolescent idiopathic scoliosis with curves greater than 30 at
skeletal maturity. Patients with adolescent idiopathic scoliosis suspected of
progressing during adulthood should be evaluated with serial scoliosis x-rays at
6-month intervals to document progression.
dramroo
Adult patients with progressive adolescent idiopathic scoliosis typically
become symptomatic in the fourth and fifth decades of life. These patients are
usually female with thoracic or thoracolumbar junction curves. The most fre
quent presenting complaint in this patient population is back pain (lumbar more
common than thoracic), and the etiology of the back pain may be multifactor
ial. Causes of back pain in this population include muscle fatigue, facet
arthropathy, radiculopathy from foraminal compression, and lumbar degenera
tive disc disease. Other presenting complaints in this population include
cosmetic deformity or, rarely, cardiopulmonary dysfunction.
For patients presenting with pain, the surgeon must deduce the cause of the
pain via the patients history, exam, and/or diagnostic testing. Fatigue-related back
pain, for example, usually occurs on the convex side of the curve, is absent in the
morning, worsens as the day progresses, and resolves with rest. It is often allevi
ated with physical therapy (strengthening of the back and abdominal muscles).
If patients fail physical therapy, then bracing may be tried to reduce the workload
on the muscles (bracing in the adult does not prevent curve progression).
Patients with facet arthropathy, on the other hand, often have pain on the
concave side of their curve or in the lower lumbar spine. This pain is also activ
ity related. Facet arthropathy can be confirmed by, and effectively treated by,
facet joint injections.
Patients with thoracic or lumbar radiculopathy can be diagnosed by their
history of radiating pain. The diagnosis can be confirmed by selective nerve
root sleeve injection. The selective nerve root sleeve injections can be periodi
cally repeated to give long-term pain relief.
Patients with degenerative disc disease causing mechanical low back pain
should undergo a trial of physical therapy to strengthen their low back and
abdominal muscles. This often alleviates their pain because these muscle
groups can then function to decrease the workload of the lumbar discs. If phys
ical therapy is ineffective, then further evaluation can be done with discography.
If the discogram correlates with their pain and with disc degeneration seen on
MRI, then lumbar fusion may be considered (if the patient is not osteoporotic).
For scoliotic patients presenting with pulmonary limitations, one key
treatment option is cessation of smoking.
In general, initial management of adults with progressive idiopathic scol
iosis should be nonoperative. Most patients will respond to the measures above
in combination with oral nonsteroidal anti-inflammatory medications.
Degenerative Adult-Onset Scoliosis

Patients with degenerative adult-onset scoliosis typically present during
the sixth or seventh decade of life. These patients have primarily lumbar defor
mities, and the male to female ratio is equal. The presenting complaints are

dramroo
usually related to lumbar stenosis or extraforaminal stenosis with radicular pain
and neurogenic claudication. Extraforaminal stenosis usually occurs at the apex
of the scoliosis on the concave side as a result of the transverse processes
abutting each other.
Low back pain may be due to facet arthropathy or degenerative disc disease.
Discogenic pain commonly occurs below the scoliotic deformity in the normal
region of the lumbar spine because this region is compensating for the scoliotic
curvature. Loss of lumbar lordosis (flat back) is common. The etiology of the
low back pain must be ascertained. On physical exam, the facet joints can be
loaded by asking the patient to extend the lumbar spine posteriorly and
laterally; pain caused with this maneuver is indicative of ipsilateral lumbar
facet arthropathy. Facet arthropathy can be confirmed by and treated with facet
joint injections.
The workup and treatment for degenerative disc disease has been previ
ously mentioned and will not be repeated here.
For symptoms of stenosis and radiculopathy, epidural steroid injections
and selective nerve root sleeve injections may assist with the diagnosis and
provide temporary relief from symptoms.
Kyphosis
Normal thoracic kyphosis ranges from 20 to 40. Hyperkyphosis is a sagit
tal plane deformity with excessive flexion of the thoracic spine. Hyperkyphosis
is subclassified as either postural or structural (fig. 3).
Postural kyphosis is due to poor posture, and the patient can consciously
correct the curve by standing up straight. Postural kyphosis is characterized
by a smooth, rounding pattern in the thoracic spine. There is no gibbus on
forward flexion. For patients with cosmetic issues related to their postural
kyphosis, we recommend physical therapy with back strengthening exercises to
improve posture. No operative intervention is indicated.
Structural kyphosis, on the other hand, cannot be consciously corrected by
the patient. Often, a gibbus is seen when the patient is asked to flex forward, and
there is often a sharp angular pattern to the kyphosis on x-ray. Structural kypho
sis can either be primary or secondary. The most common cause of primary struc
tural kyphosis is Scheuermanns disease (juvenile kyphosis). Scheuermanns
disease is defined as three consecutive levels of at least 5 of segmental kyphosis
(anterior wedging) at each level [11]. The etiology of Scheuermanns disease is
unknown; however, there does appear to be a familial occurrence. Nonoperative
treatment for Scheuermanns disease is indicated for progression of the kyphosis
beyond 40 (but under 70) in a skeletally immature patient (Risser 14). For
these patients, an underarm, hyperextension, thoracolumbar orthosis is often
satisfactory to halt progression of the kyphosis [12, 13].
dramroo
Secondary structural kyphosis, on the other hand, has an underlying
pathological process. The most common etiologies include multiple level degen
erative disc disease, vertebral body fracture or tumor, or prior multilevel laminec
tomy. Multilevel thoracic degenerative disc disease can result in thoracic
hyperkyphosis, whereas multilevel lumbar degenerative disc disease is the most
common cause of lumbar flat backs. Normally, lumbar discs are taller anteri
orly than posteriorly and thus create lumbar lordosis. As the lumbar discs
dehydrate, the lumbar spine loses this normal lordosis and becomes straight.
The patient often loses overall spinal sagittal balance as a result and tries to
compensate by flexing at the hips and knees. Patients with secondary structural
kyphosis are often older adults, and the typical presenting complaint is back pain.
The physician must, once again, determine the cause of the pain. The pain may be
due to muscle fatigue, which typically worsens as the day progresses. Pain related
to muscle fatigue can often be effectively treated by back strengthening physical
therapy. Another treatment modality is daytime bracing with an underarm thora
columbar orthosis in those unable to perform back strengthening exercises. The
orthosis can share the load with the posterior spinal musculature and reduce mus
cle fatigue and pain. However, in the skeletally mature patient, an orthosis will not
prevent progression of the kyphosis. In addition, the orthosis can contribute to fur
ther weakening and atrophy of the lumbar musculature by shielding these muscles
from spinal loads.
Pain may be due to degenerative disc disease (the workup and treatment
have been previously mentioned and will not be repeated here). Pain may be
due to facet arthropathy. This pain also tends to worsen as the day progresses.
Facet joint injections can be both diagnostic and therapeutic for this problem.
Conclusion
For the majority of patients with mild or moderate spinal deformities, ini
tial evaluation and a course (or several courses) of the appropriate nonsurgical
treatment are often satisfactory to alleviate symptoms.
For those patients who fail to have relief of their symptoms with conserva
tive treatment, and for those patients who have severe deformities, surgical
intervention is often necessary. Part 2 will focus on the indications for surgical
treatment and the options for surgical treatment of spinal deformity [14].
Acknowledgments
We are grateful to Drew Imhulse and Tom Fletcher for assistance with the radiographic
images in the figures. We are grateful to Sherry Ballenger for editorial assistance.

dramroo
The following authors are consultants for Medtronic Sofamor Danek: Stephen L. Ondra,
MD and Rick C. Sasso, MD.
References
1 Ogilvie JW: Historical aspects of scoliosis; in Lonstein JE, Bradford DS, Winter RB, Ogilvie JW
(eds): Moes Textbook of Scoliosis and Other Spinal Deformities. Philadelphia, Saunders, 1995,
pp 13.
2 Huebert HT: Scoliosis. A brief history. Manit Med Rev 1967;47:452456.
3 Risser JC: The iliac apophysis: An invaluable sign in the management of scoliosis. Clin Orthop
1958;11:111.
4 Nash CL Jr, Moe JH: A study of vertebral rotation. J Bone Joint Surg Am 1969;51:223229.
5 Lonstein JE, Carlson JM: The prediction of curve progression in untreated idiopathic scoliosis
during growth. J Bone Joint Surg Am 1984;66:10611071.
6 Lonstein JE, Bjorklund S, Wanninger MH, Nelson RP: Voluntary school screening for scoliosis in
Minnesota. J Bone Joint Surg Am 1982;64:481488.
7 Nachemson A, Peterson L: Effectiveness of treatment with a brace in girls who have adolescent
idiopathic scoliosis. J Bone Joint Surg Am 1995;177:815822.
8 Weinstein SL, Ponseti IV: Curve progression in idiopathic scoliosis: Long-term follow-up. J Bone
Joint Surg Am 1983;65:447.
9 Weinstein SL: Natural history. Spine 1999;24:2592.
10 Ascani E, Bartolozzi P, Logroscino CA, Marchetti PG, Ponte A, Savini R, Travaglini F, Binazzi R,
Di Silvestre M: Natural history of untreated idiopathic scoliosis after skeletal maturity. Spine
1986;11:784789.
11 Sorensen KH: Scheuermanns Juvenile Kyphosis. Copenhagen, Munksgaard, 1964.
12 Sachs BL, Bradford DS, Winter RB, Lonstein J, Moe J, Willson S: Scheuermann kyphosis. Follow-up
of Milwaukee-brace treatment. J Bone Joint Surg Am 1987;69/1:5057.
13 Murray PM, Weinstein SL, Spratt KF: The natural history and long-term follow-up of Scheuermanns
kyphosis. J Bone Joint Surg Am 1993;75/2:236.
14 Mummaneni PV, Ondra SL, Sasso RC: Thoracolumbar deformity advances. 2. Operative treatment
of thoracolumbar deformity. Prog Neurol Surg. Basel, Karger, 2003, pp 225239.
Assistant Professor, Department of Neurosurgery, The Emory Clinic
550 Peachtree St., Suite 806, Atlanta, GA 30308 (USA)
dramroo
Thoracolumbar Deformity Advances

2. Operative Treatment of Thoracolumbar Deformity
Praveen V. Mummanenia, Stephen L. Ondrab, Rick C. Sassoc

a
b
Department of Neurosurgery, Northwestern University, Chicago, Ill., and
c
Surgical fusions to treat spinal deformity were first developed in the

1900s. In 1911, Hibbs [1] performed a noninstrumented, posterior spinal fusion
for deformity related to tuberculous spondylitis. In 1962, Harrington [2]
revolutionized spinal fusion for deformity with the use of his distraction rod
instrumentation system.
In the past 50 years, there has been a revolution in the use of instrumenta
tion for the treatment of deformity, and new techniques and new constructs
have improved the surgical outcomes for deformity patients.
Indications for Surgery and Surgical Techniques
Adolescent Idiopathic Scoliosis

Surgical treatment is indicated for adolescent idiopathic scoliosis (AIS)
that progresses beyond 4050 in the growing child. Typically, curves over 40
in a growing child will continue to progress if not treated surgically. The pri
mary reasons to perform surgical correction of these curves is to halt and
reverse curve progression as well as to improve cosmesis. Curves between 40
and 90 usually do not cause cardiopulmonary difficulties. However, curves
greater than 90 may cause cardiopulmonary compromise, and this is an addi
tional indication to surgically correct these curves. Preoperative planning for
these patients should include a radiographic assessment (with 36-inch standing
x-rays) of curve magnitude, pedicle rotation, location of curve apex, vertebral
levels marking the ends of the curve, and overall sagittal and coronal alignment.
dramroo
The ends of the curve are defined as the vertebrae that are both stable and
neutral. The stable vertebra is the end vertebra that is most closely bisected by
the center sacral line. The neutral vertebra is the end vertebra that is the least
rotated. Usually the stable vertebra is also the neutral vertebra, but if they do
not correspond then fusion to the stable vertebra (not the neutral vertebra) is
recommended [3].
This will give the surgeon a good idea of the amount of correction needed
to reestablish spinal balance. Lateral bending films are useful to identify struc
tural (usually rigid thoracic curves) and compensatory curves (usually flexible,
lower magnitude, lumbar curves without pedicle rotation). In addition, the size
of the rib hump should also be measured, and the surgeon can consider thora
coplasty for significant rib hump deformity.
Curve Classification
Classification of AIS curves is important because the selection of fusion
levels is dependent on the classification. Over the past few decades, the King
classification system has been the most commonly used system to classify AIS
curves. Recently, however, Lenke et al. [4] described a more comprehensive and
descriptive classification system for AIS.
We prefer to use the Lenke classification system for several reasons.
The King system only describes curves in the coronal plane and does not take
sagittal balance into account. In addition, interobserver variability in classifi
cation with the King system is high. The Lenke system, on the other hand, takes
into account sagittal alignment with a sagittal modifier. In addition, the Lenke
system is highly descriptive and uses a three-tiered classification scheme
(6 curve types, lumbar spine modifier, and sagittal thoracic modifier) (fig. 1) [4].
The most frequent curve types with the Lenke system were found to be type 1
(main thoracic curve) and type 2 (double thoracic curve) with the most frequent
curve classifications found to be types 1AN, 1BN, and 2AN [4].
Surgical Correction
The Lenke classification system can be used to predict which levels should
be instrumented to correct AIS [5]. For type 1 curves, only the main thoracic
curve should be instrumented. For type 2 curves, both the proximal and main
thoracic curves should be fused. For type 3 curves (double major), both the
main thoracic and the thoracolumbar/lumbar curves should be fused. For type 4
curves (triple major), all three curves should be instrumented.
Surgical treatment for AIS is usually reserved until the patient is close to
skeletal maturity. If the patient is not near skeletal maturity, then the patient
may be at risk for a crankshaft phenomenon in the years following surgery.
The crankshaft is due to continued growth from the vertebral endplates while
dramroo
No to minimal lumbar curve
1A 2A 3A 4A
Lumbar spine
modifier
Moderate lumbar curve
1B 2B 3B 4B
Large lumbar curve
1C 2C 3C 4C 5C 6C
Type 1 Type 2 Type 3 Type 4 Type 5 Type 6

Main Double Double Triple TL/L TL/L-MT
thoracic thoracic major major
Fig. 1. Artists illustration of the Lenke classification system for scoliosis. The Lenke
system is highly descriptive and uses a three-tiered classification scheme (6 curve types,
a lumbar spine modifier, and a sagittal thoracic modifier). TL Thoracolumbar; L lumbar;
MT main thoracic.
the posterior tension band is secured by instrumentation; the spine then crank
shafts from side to side to accommodate the vertebral body growth.
In the past few decades, hook-rod instrumentation has been used success
fully for the correction of AIS. In the 1990s, however, pedicle screw-rod
Thoracolumbar Deformity Advances 227

dramroo
instrumentation has gained in popularity. Pedicle screws are stronger anchors
than hooks. In addition, unlike laminar hooks, pedicle screws do not occupy the
spinal canal nor do they disrupt the ligamentum flavum.
Structural Right Thoracic Curve (Lenke Type 1)

The structural right main thoracic curve with a compensatory (flexible,
low-magnitude, nonrotated) lumbar curve is classified as a King type II or a
Lenke type 1 curve. It is the most common curve type in AIS, and we will
discuss its treatment further.
Posterior Approach
The classic operation performed for a structural right thoracic curve in an
adolescent is a selective thoracic fusion. Classically, a selective thoracic fusion
is maintained by distraction with hook-rod instrumentation on the concave side
of the curve and compression with instrumentation on the convex side of the
curve [6, 7].
Initially, the thoracic curve is instrumented with hooks on the concave
side; a rod is attached to the concave hooks and a rod rotation maneuver is
performed to correct the curve (i.e. attach a concave rod to the hooks and then
rotate the rod until it is straight in the sagittal plane). On the concave side,
up-going pedicle hooks are typically placed at the superior end of the curve, and
down-going laminar hooks are typically placed at the inferior end of the curve.
This hook pattern allows for distraction on the rod, which serves to maintain the
correction after a rod rotation maneuver. On the convex side, a claw is placed
at the superior end of the curve (the claw consists of down-going hooks above
and up-going hooks at the next level below), an up-going hook is placed at the
apex of the curve, and down-going hooks are placed inferiorly. This hook pat
tern allows for compression on the rod in order to maintain curve correction on
the convex side of the curve. The compensatory lumbar curve is not instru
mented and will typically correct to normal when the structural thoracic curve
is straightened [8].
Alternative posterior fusion options include hybrid hook-screw-rod
systems, hook-endplate screw systems, or pure pedicle screw-rod systems.
In the hybrid system, hooks are placed superiorly (where the pedicle diameters
are smaller) and screws are placed inferiorly (where the pedicle diameters are
larger). In the pedicle hook-endplate screw system, a pedicle hook is placed and
then anchored to an endplate screw; this rigid fixation technique is similar to
pedicle screws in pullout strength. In pure pedicle screw systems, no hooks
are used; instead, pedicle screws are placed in lieu of the hooks. In all three
systems, a rod rotation maneuver is initially done on the concave side of the
curve to correct the curve.
dramroo
When we place thoracic pedicle screws, we have found the following sur
gical nuances to be helpful. We locate the entry point for thoracic pedicle
screws by using a high-speed burr to decorticate the surface of the thoracic
facet overlying the area of the pedicle. Once this mild decortication has been
done, a blush of blood is usually noticeable, and this blush marks the entry to
the pedicle. This blush of red arises from the cancellous bone at the center of
the pedicle, which tends to bleed when uncovered. The surrounding facet area
does not tend to blush with this maneuver. The surgeon can reliably establish
the pedicle entry point through this technique.
The typical location of the pedicle with respect to the thoracic transverse
process and facet has been elucidated by Lenke et al. [9].
One word of caution is in order, however. If the central portion of the pedi
cle is corticated, then this blush technique cannot be used, as there is no can
cellous bleeding bone. In this case, we recommend the surgeon perform a
laminoforaminotomy to feel the medial, superior, and inferior walls of the pedi
cle to guide the screw placement [10].
Once the entry point has been established, there are three options we use to
establish a pathway through the pedicle for the pedicle screw. The first method we
use entails tapping the pedicle from the entry point with a 3.5-mm tap from the
VERTEX posterior cervical set (Medtronic Sofamor Danek, Memphis, Tenn.,
USA). This tap has fine cutting flutes and with minimal downward pressure, the
tap has a tendency to find the appropriate trajectory for screw placement by cut
ting through the cancellous portion of the pedicle without violating the cortical
pedicle walls. The screw can then be placed through this trajectory.
The second option is to use the thoracic pedicle probe recently created by
Lenke et al. [9]. This small tap is ideal for the thoracic pedicles. It has a small
angled end, and this end is initially placed through the pedicle entry point facing
laterally (in order to avoid medial cortical wall violation and spinal canal penetra
tion). After a depth of 15 mm (marked on the tap itself), the probe is then turned
medially to enter into the vertebral body and to avoid lateral wall breakout from
the vertebral body. It is important to note that after the first 15 mm, the probe is
beyond the spinal canal in most patients, and the risk of spinal cord injury is
minimized during the first 15 mm by the laterally directed angle of the probe [9].
We avoid drilling down the pedicle as cortical breakthrough can occur with
the drill. However, when the pedicle is corticated, then drilling becomes a
necessity (the third option). We guide our drill trajectory by feeling the outer
walls of the pedicle after performing a laminoforaminotomy. When the pedicles
are too small to accommodate 4.5-mm screws, several options should be con
sidered. First, by using a 3.5-mm tap (we prefer to use the one available on
the VERTEX posterior cervical system), we establish a pathway through the
pedicle into the vertebral body. Then we use a 4.5-mm tap to dilate the pedicle.

dramroo
AT
ST
Fig. 2. Artists illustration of the difference between the anatomic (AT) and the
straightforward (ST) trajectories for thoracic pedicle screw placement. The thoracic pedicle
screws can either be placed parallel to the endplates (straightforward trajectory), or they can
be placed parallel to the axis of the pedicle itself (anatomic trajectory).
In adolescents, the pedicle often dilates with this successive tapping, and this
allows for placement of a 4.5-mm screw.
Another option for dealing with a small pedicle is to use the in-out-in
technique for screw placement [11]. In this technique, the screw enters the pedicle
and then is purposely placed through the lateral wall of the pedicle and engages
the bone of the transverse process and rib head. The screw subsequently reen
ters the pedicle more anteriorly and ends in the vertebral body [12, 13].
The surgeon should keep in mind that the pedicle on the concave side of
the curve may be smaller than the pedicle on the convex side. Consequently,
different screw widths may be needed for the contralateral pedicles of the same
level in scoliotic patients [14].
Use of fluoroscopy can be helpful in establishing the appropriate sagittal
plane angulation of the pedicle screw. Thoracic pedicle screws can either be
placed parallel to the endplates (straightforward trajectory), or they can be
placed parallel to the axis of the pedicle itself (anatomic trajectory) (fig. 2).
When placed parallel to the axis of the pedicle (anatomic trajectory), a longer
screw can be used, and the tip of the screw is directed at the anterior, inferior
vertebral endplate [1517].
Anterior Approach
Another alternative is to perform the correction through an anterior
approach by releasing the anterior longitudinal ligament, debriding the discs
anteriorly, packing the disc spaces with bone graft harvested from a rib, placing
dramroo
vertebral body screws, and derotating the spine with a connected rod. This can
be done either through a video-assisted thoracoscopic surgery (VATS), via an
open thoracotomy, or via a mini-open thoracotomy, which is a hybrid tech
nique utilizing a small thoracotomy inferiorly and VATS superiorly [18, 19].
In patients who are not near skeletal maturity, an anterior operation may be
required to release the spine anteriorly and remove the growing vertebral end-
plates. The surgeon can choose to perform only an anterior release and epiph
ysiodesis with a subsequent posterior instrumented procedure, or he can
perform an anterior release, derotation, and fusion through a single anterior
approach.
The anterior approach for derotation and fusion has advantages and disad
vantages. VATS offers the advantages of decreased blood loss, sparing of the
posterior muscular tension band, and a smaller surgical scar. However, VATS
requires a lung takedown (with subsequent chest tube placement) and entails a
longer operative time (often double the time of a posterior operation).
Anterior derotation and fusion procedures typically do not achieve the
same degree of curve correction that posterior procedures do; this is likely due
to the limitations in the number of segments that can be accessed via an ante
rior approach. When patients are hyperkyphotic the anterior approach is not
possible, as access to the anterior vertebral bodies is limited. In addition, long-
term follow-up of anterior derotation and fusion procedures is currently lacking.
Consequently, a posterior derotation and fusion operation is still the gold-
standard operation for correction of AIS.
Double Thoracic Curves (Lenke Type 2)

In progressive double thoracic curves, both curves are structural.
Consequently, the fusion will need to extend to the stable vertebrae above and
below both the curves. The preferred approach for the correction of double tho
racic curves is the posterior approach.
An anterior approach for these curves may be necessary as an adjunct to a
posterior correction and fusion (for an anterior release and epiphysiodesis), but
instrumentation and curve correction via an anterior approach are very diffi
cult, if not impossible due to the difficulty of accessing and instrumenting
numerous thoracic levels.
Double Major Curves (Lenke Type 3)

Double major curves have a structural thoracic and a structural thoraco
lumbar or lumbar curve. In general, the instrumentation should extend to the
stable vertebra above the thoracic curve and the stable vertebra below the tho
racolumbar (or lumbar) structural curve.

dramroo
An area of controversy in treating progressive structural curves that extend
to the low lumbar spine is whether or not to extend the fusion to the sacrum.
In general, in the treatment of AIS, the instrumentation should extend to the
stable vertebra at the inferior end of the lowest structural curve (which is rarely
at L4 or below) and fusion to the sacrum should be avoided unless the low
lumbar discs are severely degenerated.
Extension of fusion to the sacrum has several disadvantages. First of all,
the risk of pseudoarthrosis increases when a long construct is taken down to the
sacrum [20]. Also, long constructs can fail due to screw pullout from the
sacrum. In addition, extension of the fusion to the sacrum increases the risk of
not completely correcting a patients spinal imbalance. Patients often rebalance
themselves following scoliosis corrections by accommodating any offset from
the balanced state through alterations in the position of the lower lumbar spine.
When the lumbosacral junction is fused, this accommodation potential is lost.
Finally, extension to the sacrum runs the risk of creating flat-back syndrome if
an appropriate amount of thoracic kyphosis and lumbar lordosis is not placed
in the rods.
Adult Scoliosis
Adult Patients with Progressive AIS
The indications for surgery in this patient population are curve progression
greater than 5 in a single year, back pain that is not relieved by conservative
measures, pulmonary compromise from severe curvature, and unacceptable
cosmetic deformity. Of these, the two most common indications are unrelent
ing back pain and cosmetic deformity.
Surgical Correction
The principles for surgical correction of adult patients with AIS are simi
lar to those for adolescents with AIS. The curve can be evaluated and classified
by the Lenke system. The instrumentation pattern is dependent upon the curve
classification.
There are, however, two main differences between treating AIS in adults
versus adolescents. First, the rate of pseudoarthrosis is higher for adults than for
adolescents. Consequently, adults may need both anterior and posterior surgery
to achieve a solid fusion. The anterior approach is often required in long curves
to release the anterior longitudinal ligament and incise disc spaces to allow for
a rigid curve to be corrected. The disc spaces can then be packed with structural
graft to promote fusion, and a subsequent posterior procedure can be performed
to instrument and correct the curvature.
The second major difference between adults and adolescents with AIS is
that adults often have significant disc degeneration and spondylosis in the
dramroo
lumbar spine. Stopping a long segment fusion at a stable and neutral vertebra
in the mid lumbar spine may not be appropriate in the face of significant lum
bar spondylosis and degenerative disc disease (DDD). Patients with significant
lumbar spondylosis and DDD may have progression of the spondylosis or sig
nificant low back pain from accelerated disc degeneration due to the greater
loads on these segments following a long segment fusion above. Consequently,
the need for extension of the fusion to the sacrum is greater in adults than it is
in adolescents. However, if there is no significant lumbar spondylosis or disc
degeneration, then stopping the fusion in the mid lumbar spine is appropriate.
Degenerative Adult-Onset Scoliosis

Degenerative adult-onset scoliotic curves are often rigid, lumbar curves.
The indications for operative intervention are curve progression and unrelent
ing low back and radicular pain (not responsive to the conservative measures
previously discussed). Often patients have significant radicular pain on the con
cave side of the curve due to severe foraminal compromise. Patients with these
curves may be in overall spinal balance as many of them compensate at the tho
racolumbar and lumbosacral junctions. Consequently, evaluation of overall
spinal balance with 36-inch scoliosis x-rays is important. If a patient is a surgi
cal candidate but is in overall spinal balance, then correction of the curve is not
necessary.
Surgical Correction
In the subgroup of patients who are in overall spinal balance, there are two
surgical treatment options. First, if the patient suffers from primarily radicular
symptoms, then a unilateral Wiltse approach and decompression (posterolateral
approach between the multifidus and longissimus muscles to perform an
extraforaminal nerve root decompression) on the concave side of the curve are
often sufficient to provide relief from radicular pain [21]. The advantage of this
option is that the posterior ligamentous tension band is left intact and bony
removal of the posterior column is kept to a minimum. Only the lateral portion
of the facet joints is removed and the roof of the neural foramen is opened from
lateral to medial.
If the patient, who is in overall spinal balance, is suffering from both low
back and radicular pain, then another surgical treatment option is needed. For
these patients, we recommend foraminal decompression on the concave side of
the curve with an in situ instrumented fusion with pedicle screw-rod instru
mentation. The foraminal decompression from a standard midline posterior
approach will address the radicular symptoms. The fusion will relieve the low
back pain (which is likely secondary to DDD and lumbar spondylosis). We
always use autograft bone harvested from the iliac crest to establish the fusion.

dramroo
a b
Fig. 3. a Preoperative 36-inch x-ray of a fixed thoracolumbar scoliotic deformity in an

adult. b Postoperative 36-inch anterior-posterior x-ray of the same patient following an ante
rior lumbar release with interbody fusion (Pyramesh, Medtronic Sofamor Danek) and a sub
sequent posterior multisegmental instrumented correction with thoracic hooks superiorly and
thoracic and lumbar pedicle screws inferiorly (M-10 system, Medtronic Sofamor Danek).
When patients with degenerative adult-onset scoliosis are not in overall

spinal balance and when their symptoms cannot be controlled by conservative
measures, then correction of the scoliotic curvature is needed (fig. 3a). This is
best accomplished by a combined anterior and posterior approach. An anterior
approach is often helpful to release the rigid curve by incising the anterior lon
gitudinal ligament and the anterior disc spaces. Autograft-filled Pyramesh
cages (Medtronic Sofamor Danek) can be placed in the anterior interbody space
to promote fusion (fig. 3b). If no anterior column support with autograft is
dramroo
performed, then the risk of pseudoarthrosis is high in this elderly patient
population.
A subsequent posterior approach with a coronal plane wedge osteotomy of
the convex facets is performed (i.e. the convex facets are removed from the pedi
cle above the neural foramen to the pedicle below the neural foramen). The
wedge-shaped osteotomy of the convex side of the curve prevents compression of
the exiting nerve roots on the convex side. This osteotomy is extended beyond the
midline and includes the convex side of the laminae in order to avoid thecal sac
compression when the convex side is compressed. Pedicle screws are then placed
at all levels from the superior to the inferior stable vertebrae of the lumbar scoli
otic curve. Rods are connected to the screws and the convex side of the curve is
compressed while the concave side is distracted, and the curve is straightened.
If the surgeon wishes to perform the entire operation from a posterior
approach, then multilevel transforaminal lumbar interbody fusions can be per
formed and Pyramesh cages filled with autograft can be placed into the inter-
body space through this approach. Fusion to the sacrum is avoided if the
inferior stable vertebra is at L4 or above and if the L4/5 and L5/S1 discs are in
good condition. If the lumbosacral junction is not fused, then additional curve
compensation is possible for the patient. However, if the curve extends to L5 or
if the low lumbar discs are significantly degenerated, then fusion to the sacrum
is often needed. There are several options for extension of the fusion to the
sacrum. First of all, S1 (and S2) pedicle screws can be placed. However, the
sacral pedicles have large cancellous centers, and they often do not allow for
solid screw purchase. Sacral pedicle screws have greater pullout strengths if
they are bicortical, and lateral fluoroscopy can assist with bicortical placement.
Long segment fusions to the sacrum are subject to large cantilever loads,
which can lead to pullout of sacral pedicle screws, sacral fractures, and lum
bosacral pseudoarthrosis. Consequently, some surgeons elect to supplement
sacral pedicle screws with intrasacral rods, transacral rods, and/or iliac screw
fixation [22, 23]. Iliac screw fixation can be hindered by harvesting posterior
iliac crest bone graft. However, the harvest of posterior iliac crest bone graft
does not completely preclude the placement of iliac crest screws [24].
Kyphosis
Surgical treatment of kyphosis can be considered if there is a rigid thoracic
or thoracolumbar kyphotic curve greater than 6070 in either the adolescent or
the adult. Curves of 6070 are often cosmetically deforming, but they usually
do not cause significant cardiopulmonary symptoms. However, these curves
can cause intractable back pain that is not responsive to conservative treatment
measures. Curves over 8090, on the other hand, can cause cardiopulmonary
compromise and surgical correction is definitely indicated.

dramroo
Preoperative Postoperative
Bony resection
Winn after Bridwell
Fig. 4. Artists illustration of a three-column PSO. The technique involves removing a

large wedge of the posterior and middle columns (including the pedicle) of the vertebral
body. The remainder of the posterior and middle columns is then approximated by pulling
together pedicle screws above and below the osteotomy level. The anterior column of the
PSO level acts as the fulcrum for this correction.
Surgical Correction
Surgical correction of large, structural thoracic or thoracolumbar curves
can be performed through a posterior approach or a combined anterior and
posterior approach. Correction through an anterior only approach is difficult
because the severe thoracic kyphosis limits the surgeons access to the anterior
thoracic spine.
Curves of 6070 can be treated via a posterior only approach, especially if
the curves are flexible. The treatment strategy for these curves is to expose the
ends of the curve, and then to either perform a pedicle subtraction osteotomy
(PSO) or multilevel posterior closing wedge osteotomies (multiple Smith
Petersen osteotomies). PSO is useful for a focal thoracic kyphosis (i.e. thoracic
wedge compression fracture) because with PSO, the majority of the correction is
performed over a single segment. The technique involves removing a large wedge
of the posterior and middle columns (including the pedicle) of the vertebral body.
The remainder of the posterior and middle columns is then approximated by
pulling together pedicle screws above and below the osteotomy level. The ante
rior column of the PSO level acts as the fulcrum for this correction (fig. 4), and
up to 30 of correction can be achieved with a single level of PSO [25].
Prior to approximating the remainder of the posterior and middle columns
in PSO, a Z-plasty of the underlying dura may be needed to avoid buckling of
the dura into the spinal canal. We prefer to perform PSO below the conus to
avoid injury to the spinal cord.
dramroo
Preoperative Postoperative
Bony resection
Winn after Bridwell
Fig. 5. Artists illustration of a SmithPetersen osteotomy. The technique involves

removing multiple wedges of the facets bilaterally over several levels. The remaining poste
rior columns are then pulled together (by pedicle screw-rod instrumentation) while the neural
foramina are monitored closely to ensure that excessive foraminal narrowing does not occur.
The fulcrum of the correction is the anterior disc space, and, consequently, in rigid curves,
an anterior procedure may be needed to first incise the anterior longitudinal ligament.
Multiple SmithPetersen osteotomies, on the other hand, are much safer

because the correction is spread out over multiple levels. Only 1 of correction
can be expected for every 1 mm of bone removed with this technique.
Consequently, multiple SmithPetersen osteotomies are needed to correct a sig
nificant kyphosis. The technique involves removing multiple wedges of the facets
bilaterally over several levels (fig. 5) [26]. The remaining posterior columns are
then pulled together (by pedicle screw-rod instrumentation) while the neural
foramina are monitored closely to ensure that excessive foraminal narrowing
does not occur. The fulcrum of the correction is the anterior disc space, and,
consequently, in rigid curves, an anterior procedure may be needed to first
incise the anterior longitudinal ligament and the anterior disc space [27].
Conclusion
Surgical correction of spinal deformity has advanced rapidly over the past
decade. Correction of severe scoliotic and kyphotic deformities can now be
performed with low mortality rates. However, the morbidity rates for major
deformity corrections remain very high. New advances will likely blend more
minimally invasive techniques with the current instrumentation systems in an

dramroo
attempt to reduce the surgical morbidity associated with multisegmental
correction and instrumented fusion.
Acknowledgments
We are grateful to Drew Imhulse and Tom Fletcher for assistance with the radiographic
images in the figures, to Sherry Ballenger for editorial assistance, and to Bill Winn for assis
tance with the illustrations. We thank Medtronic Sofamor Danek for permission to use
several of the illustrations.
The following authors are consultants for Medtronic Sofamor Danek: Stephen L.
Ondra, MD and Rick C. Sasso, MD.
References
1 Hibbs RA: An operation for progressive spinal deformities. NY Med J 1911;93:10131016.

2 Harrington PR: Treatment of scoliosis: Correction and internal fixation by spine instrumentation.
3 King HA, Moe JH, Bradford DS, Winter RB: The selection of fusion levels in thoracic idiopathic
scoliosis. J Bone Joint Surg Am 1983;65:13021313.
4 Lenke LG, Betz R, Harms J, Bridwell KH, Clements DH, Lowe TG, Blanke KM: Adolescent idio
pathic scoliosis: A new classification to determine extent of spinal arthrodesis. J Bone Joint Surg
Am 2001;83:11691181.
5 Lenke LG, Betz RR, Clements D, Merola A, Haher T, Lowe T, Newton P, Bridwell KH, Blanke K:
Curve prevalence of a new classification of operative adolescent idiopathic scoliosis: Does
classification correlate with treatment? Spine 2002;27:604611.
6 Cotrel Y, Dubousset J: New segmental posterior instrumentation of the spine. Orthop Trans
1985;9:118.
7 Lenke LG, Bridwell KH, Baldus C, Blanke K, Schoenecker PL: Cotrel-Dubousset instrumenta
tion for adolescent idiopathic scoliosis. J Bone Joint Surg 1992;74:10561067.
8 Lenke LG, Betz R, Bridwell KH, Harms J, Clements D, Lowe T: Spontaneous lumbar curve
coronal correction after selective anterior or posterior thoracic fusion in adolescent idiopathic
scoliosis. Spine 1999;24:16631672.
9 Lenke LG, Rinella A, Kim YJ: Free hand thoracic pedicle screw placement. Semin Spine Surg
2002;14/1:4857.
10 Mummaneni PV, Traynelis VC, Sasso RC: Cervical lateral mass advances. Prog Neurol Surg.
Basel, Karger, 2003, pp 165175.
11 Belmont PJ Jr, Klemme WR, Dhawan A, Polly DW Jr: In vivo accuracy of thoracic pedicle screws.
Spine 2002;26:23402346.
12 Dvorak M, MacDonald S, Gurr KR, Bailey SI, Haddad RG: An anatomic, radiographic, and
biomechanical assessment of extrapedicular screw fixation in the thoracic spine. Spine 1993;18:
16891694.
dramroo
14 Liljenqvist UR, Link TM, Halm HFH: Morphometric analysis of thoracic and lumbar vertebrae in
15 Klemme WR, Belmont PJ, Polly DW: Transpedicular thoracic screws: General concepts and inser
tion technique using fluoroscopic guidance. Semin Spine Surg 2002;14/1:4347.
16 McCormack BM, Benzel EC, Adams MS, Baldwin NG, Rupp FW, Maher DJ: Anatomy of the
thoracic pedicle. Neurosurgery 1995;37:303308.
17 Kuklo TR, Polly DW: Surgical anatomy of the thoracic pedicle. Semin Spine Surg 2002;14/1:37.
18 Newton PO, Wenger DR, Mubarak SJ, Meyer RS: Anterior release and fusion in pediatric spinal
deformity: A comparison of early outcome and cost of thoracoscopic and open thoracotomy
approaches. Spine 1997;22:13981406.
19 Lenke LG, Rhee J: Adolescent scoliosis: Anterior surgical techniques for adolescent idiopathic
scoliosis. Curr Opin Orthop 2001;12/3:199205.
20 Eck KR, Bridwell KH, Ungacta FF, Riew KD, Lapp MA, Lenke LG, Baldus C, Blanke K:
Complications and results of long adult deformity fusions down to L4, L5, and the sacrum. Spine
2001;26:E182E192.
21 Wiltse LL: The paraspinal sacrospinalis-splitting approach to the lumbar spine. Clin Orthop 1973;
91:4857.
22 Mazda K, Khairouni A, Pennecot GF, Bloch J: The ideal position of sacral transpedicular endplate
screws in Jacksons intrasacral fixation. An anatomic study of 50 sacral specimens. Spine 1998;
23:21232126.
23 Emami A, Deviren V, Berven S, Smith JA, Hu SS, Bradford DS: Outcome and complications of
long fusions to the sacrum in adult spine deformity: Luque-Galveston, combined iliac and sacral
screws, and sacral fixation. Spine 2002;27:776786.
24 Kuklo TR, Bridwell KH, Lewis SJ, Baldus C, Blanke K, Iffrig TM, Lenke LG: Minimum 2-year
analysis of sacropelvic fixation and L5S1 fusion using S1 and iliac screws. Spine 2001;26:
19761983.
25 Kim KT, Suk KS, Cho YJ, Hong GP, Park BJ: Clinical outcome results of pedicle subtraction
osteotomy in ankylosing spondylitis with kyphotic deformity. Spine 2002;27:612618.
26 Smith-Petersen MN, Larson CB, Aufranc OE: Osteotomy of the spine for correction of flexion
deformity in rheumatoid arthritis. Clin Orthop 1969;66:69.
27 Booth KC, Bridwell KH, Lenke LG, Baldus CR, Blanke KM: Complications and predictive
factors for the successful treatment of flatback deformity (fixed sagittal imbalance). Spine 1999;
24:17121720.
Assistant Professor, Department of Neurosurgery, The Emory Clinic
550 Peachtree St., Suite 806, Atlanta, GA 30308 (USA)

dramroo
Vertebral Augmentation for
Osteoporotic and Osteolytic
Vertebral Compression Fractures:
Vertebroplasty and Kyphoplasty
Isador Lieberman
The Cleveland Clinic Foundation, Cleveland, Ohio, USA
Osteoporosis is a systemic disease currently afflicting more than 40 mil

lion Americans; this figure is likely to rise further as the population ages. It
results in progressive bone mineral loss and concurrent changes in bony archi
tecture that leave bone vulnerable to fracture, often after minimal or no trauma.
Osteolysis secondary to metastatic disease or multiple myeloma affects up to
70% of patients on initial presentation. The spine is the most common site of
osteoporotic or osteolytic fracture, with vertebral compression fracture (VCF)
occurring in up to 50% of women 80 years and older and 25% of women 70
years and older [1, 2]. Overall, 700,000 people per year in the US suffer a VCF
secondary to osteoporosis, exceeding even the frequency of hip fractures [11].
Osteoporotic VCFs have been shown to be associated with up to a 30% age-
adjusted increase in mortality [3]. The cost to society of managing osteoporotic
VCF patients in the United States in 1995 was USD 746 million [4]. Possible
acute complications of either osteoporotic or osteolytic vertebral fracture
include cord compression, urinary retention, and ileus [10]. Long-term conse
quences include considerable pain (reported in 35% of detectable VCFs) [12],
as well as pulmonary compromise (a 9% loss in predicted forced vital capacity
with each vertebral fracture) [13]. Other chronic sequelae include decondition
ing, deformity, insomnia, and depression, resulting in substantial physical,
functional, and psychosocial impairment [13, 14].
Nonoperative Management of VCFs

Two thirds of patients with acute, painful osteoporotic VCFs improve
regardless of the treatment applied. Traditional, nonoperative management
dramroo
includes bed rest, analgesics, and bracing. This type of medical management,
however, fails to restore spinal alignment, and the lack of mobility itself can
result in secondary complications, including worsening osteoporosis, atelectasis,
pneumonia, deep vein thrombosis, decubitus ulcer, and pulmonary embolism.
An alternative approach is supervised ambulatory mobility by a physiotherapist
plus hydrotherapy [15]. In one third of patients, severe pain, limited mobility,
and poor quality of life persist despite appropriate nonoperative management.
No patient spontaneously achieves a realigned spine, corrected sagittal contour,
or restoration of vertebral height.
Half the patients with metastases to the spine report pain relief after exter
nal beam radiation [31]. Patients with radiosensitive tumors (breast, prostate,
myeloma) typically do well, however radiotherapy does not protect the spine
from progressive osteolytic collapse and presents the treating surgeon with
major concerns regarding postoperative wound healing and bone fusion should
surgery be indicated [32]. Similar concerns exist when considering the use of
chemotherapy to treat various metastases to the spine, although there are new
investigational bisphosphonates which are promising reversal of bone loss [33].
Operative Management of VCFs

Historically, the only alternative to nonoperative management for sympto
matic osteoporotic or osteolytic vertebral fractures was open surgical decom
pression (anterior or posterior decompression and stabilization via internal
fixation hardware and bone grafting), and this was usually reserved for those
patients with gross spinal deformity or neurologic impairment (0.5%) [10, 32].
The reason for this surgical caution was due to the adverse risk/benefit ratio in
this elderly or cancer population with poor bone quality and multiple comorbid
conditions.
Percutaneous vertebroplasty (PVP) is a minimally invasive method that
involves the percutaneous injection of polymethyl methacrylate (PMMA) into
a collapsed vertebral body to stabilize the vertebra. Originally developed for
osteolytic metastasis, myeloma, and hemangioma, the procedure resulted in
quick, effective pain relief and a low complication rate [57]. PVP is now also
increasingly used for the treatment of osteoporotic vertebral fractures [8].
However, PVP does not expand the collapsed vertebra, potentially locking the
spine in a kyphotic posture. In addition, the PMMA bone filler has associated
problems (epidural leakage, thermal necrosis, inability to integrate with bone,
handling difficulties, toxicity to patient and operator) [2, 9].
Kyphoplasty is a newer minimally invasive technique with a number of
potential advantages over PVP, including lower risk of cement extravasation and
better restoration of vertebral body height [29]. A cannula is introduced into the
vertebral body, followed by insertion of an inflatable bone tamp (IBT), which
Vertebroplasty and Kyphoplasty 241

dramroo
when deployed, reduces the compression fracture and restores the vertebral
body toward its original height, while creating a cavity to be filled with bone
cement. The cement augmentation is therefore done with more control into the
low-pressure environment of the preformed cavity with viscous, partially cured
cement.
Percutaneous Vertebroplasty
Background
Percutaneous vertebral augmentation (vertebroplasty, PVP) was first
devised by Galibert et al. [16] in 1987, and initially involved the augmenta
tion of the vertebral body with PMMA during open procedures to allow stable
fixation of internal hardware. PVP was first described in the French literature in
1987 [16] but was not performed in the United States until 1994. Originally
targeted for osteolytic metastasis, myeloma, and hemangioma, PVP resulted in
early appreciable pain relief and a low complication rate [7, 16]. Its indications
now include osteoporotic vertebral collapse with chronic pain, expanding further
to include treatment of asymptomatic vertebral collapse and even prophylactic
intervention for at risk vertebral bodies [17]. Nevertheless, the treatment of acute
fractures in ambulatory patients and prophylactic treatment remains controver
sial [18]. In fact, vertebral augmentation itself is somewhat controversial, with
questions concerning a lack of defined indications, expected complications,
outcome measures, and the need for long-term follow-up data [2].
An open question in PVP is the mechanism of pain relief. The most
intuitive explanation involves simple mechanical stabilization of the fracture;
the cement stabilizes vertebral bodies and offloads the facet joints. However,
another possibility is that analgesia results from local chemical, vascular, or
thermal effects of PMMA on nerve endings in surrounding tissue [8, 19].
Supporting this concept is the lack of correlation between cement volume and
pain relief [20, 21]. Further evidence against an effect resulting solely from
mechanical stabilization is the fact that PVP typically does not restore lost ver
tebral body height and therefore does not correct altered biomechanics [1, 18].
Technique
Injection of opacified PMMA is performed via a transpedicular or para
vertebral approach under continuous fluoroscopic guidance to obtain adequate
filling and to avoid PMMA leakage. For complex or high-risk cases, CT and
fluoroscopic guidance are sometimes combined [5, 18]. In routine cases, PVP
can be performed under local anesthesia with slight sedation in less than an
hour [1], although general anesthesia is sometimes required because pain may
Lieberman 242
dramroo
intensify during cement injection [8]. Preceding PMMA injection, intraosseous
venography is often used to determine the filling pattern and identify sites of
potential PMMA leakage (outline the venous drainage pattern, confirm needle
placement within the bony trabeculae, and delineate fractures in the bony
cortex). However, others have dispensed with routine venography [1].
Contraindications to vertebroplasty include coagulopathy, absence of facil
ities to perform emergency decompressive surgery in the event of a complica
tion, and extreme vertebral collapse (6570% reduction in vertebral height)
[8]. However, this last contraindication has been questioned recently [18, 22].
Results
The available clinical studies, mostly European, report pain relief in about
90% of cases treated for osteoporotic fracture, with only infrequent clinically
significant complications (010%), most of them minor [7, 18]. In a series of 80
patients treated with PVP for osteoporotic vertebral collapse, 90% gained imme
diate pain relief [5]. During a follow-up of 1 month to 10 years, only one com
plication was reported: intercostal neuralgia treated by local anesthetic
infiltration. In another prospective study of 45 vertebral body augmentations in
17 osteoporotic patients led to significant and lasting pain reduction during the
1-year follow-up [1]. Although cement leakage occurred in 20% of vertebral bod
ies, none had clinical sequelae. In a retrospective review of 70 augmented verte
brae in 38 consecutive patients with osteoporosis, treatment resulted in pain relief
within 48 h in 36 patients (95%). The pain relief was durable in 34 of the patients
(89%) during a follow-up averaging 18 months. Twenty-four patients (63%) expe
rienced marked to complete pain relief, 12 (32%) moderate relief, and 2 (5%) no
significant change [18]. Of 8 patients suffering malignant neoplasm of the spinal
column and treated with vertebroplasty, 4 (50%) found pain relief in this series
[18]. Cortet et al. [23] studied 16 patients with 20 osteoporotic vertebral com
pression fractures who underwent PVP. This study found a statistically significant
decrease in pain with several standardized scoring systems at all observed time
points during the 6-month follow-up, along with concurrent, significant improve
ment in overall health status. No adverse events and no vertebral fractures
occurred during the follow-up period. Another study of 29 patients with 47
painful osteoporotic vertebral fractures reported a 90% success rate in terms of
significant pain relief immediately after treatment [7]. Two patients sustained rib
fractures during the procedure resulting in pain that subsequently resolved;
otherwise, no clinically significant complications were noted.
Complications
The principal risk of PVP, which involves the forced injection of low-
viscosity PMMA cement into the closed space of the collapsed vertebral body,

dramroo
is cement extravasation. Extravasation rates are as high as 40% when used to
treat osteoporotic fractures [7]. The likelihood is greater when using cement
with a liquid rather than paste consistency or with higher PMMA volume [24].
However, in most settings, the majority of extravasations have no clinical rele
vance, at least in the short term [1].
The consequence of an extravasation depends on its location. In epidural or
foraminal extravasation, nerve root compression and radiculopathy are the major
risk, occurring in 11 of 274 patients (4%) treated by Deramond et al. [5]. Three
patients required surgical nerve root decompression, as has been described by
others as well [10, 20]. Extravasation into perivertebral veins can cause cement
embolism to the lungs; deaths attributed to cement embolism have been docu
mented. However, the 2 deaths attributed to pulmonary embolism were reported
to be unrelated to the procedure; no cement material was detected by chest x-ray
of the first patient [8, 25], and the second pulmonary embolism arose from deep
venous lower extremity thrombosis [5]. On the other hand, extravasation into
adjacent disks or paravertebral tissue, although common, generally produces no
patient symptoms and carries little clinical significance; many such extravasa
tions can be avoided by careful needle positioning [5].
Other operative and long-term complications of PVP are specific to PMMA
as a filler [1, 17, 26]. The physician may work with PMMA in large batches in
order to keep it liquid and to extend the working time for vertebroplasty. However,
its high polymerization temperature (86107C within cement core) [27] can
damage adjacent tissue, including the spinal cord and nerve roots [9], leading to
an inflammatory reaction and transitory exacerbation of pain [8]. When injecting
PMMA monomer, vigilance and caution of the physician are required.
Absorption of PMMA during the injection can induce hypotension by virtue of
its cardiotoxic and arrhythmogenic properties [28]. Keeping in mind that placing
a material in the spine affords proximity and access to the chest and the heart,
vertebral augmentation with PMMA demands meticulous attention to technique.
Overall, the risk of complications that carry clinical significance follow
ing PVP for osteoporotic vertebral fracture is 13%, and most potential com
plications can be avoided with a good technique [5].
Kyphoplasty
Background
Kyphoplasty is a new technique evolved from a marriage of vertebroplasty
with balloon angioplasty. It has a number of potential advantages, including
lower risk of cement extravasation and better restoration of vertebral body
height. A cannula is introduced into the vertebral body, via a transpedicular or
Lieberman 244
dramroo
a b c
Fig. 1. ac Kyphoplasty example, IBT inflation and PMMA augmentation.
extrapedicular route, followed by insertion of an IBT, which when deployed,

reduces the compression fracture and restores the vertebral body toward its
original height. This then creates a cavity to be filled with bone cement. The
cement augmentation can now be completed with more control into the low-
pressure environment of the preformed cavity with viscous, partially cured
cement. Using a cannula for bone filler with a steel stylet as a plunger enables
the operator to apply cement at considerably higher viscosity than is possible
with injection through a 5-ml syringe and 11-gauge needle. Both the higher
cement viscosity and lower-pressure injection reduce the risk of cement
extravasation. Filling is performed under continuous lateral fluoroscopic guid
ance similar to vertebroplasty. The procedure can be performed under general
anesthesia or local anesthesia with intravenous sedation; some patients are able
to return home the same day of procedure.
Technique
With the patient under general or local anesthesia in prone position on a
radiolucent spinal frame, two C-arms are positioned for anteroposterior and lat
eral fluoroscopic images. Once positioned the C-arms or patient are not moved
to ensure repeatable images throughout the case. Two 3-mm incisions are made
at the vertebral level, parallel to the pedicles in both planes. Then a guide wire
or biopsy needle is advanced into the vertebral body via a transpedicular or
extrapedicular approach, depending on fracture configuration and patients
anatomy. The guide wire is then exchanged for the working cannula using a
series of obturators. Once the working cannula is positioned the surgeon reams
out a corridor to accommodate the IBT and positions IBT under the collapsed
endplate. To deploy the IBT, inflation proceeds slowly under fluoroscopy until
maximum fracture reduction is achieved or the balloon reaches a cortical wall
(see fig. 1). At this point the surgeon deflates and removes the IBT, mixes the
cement, prefills the cement cannulae and allows the cement to partially cure in

dramroo
the cement cannulae. Once partially cured PMMA is slowly extruded into the
vertebral body through each pedicle under continuous lateral fluoroscopic guid
ance. This technique permits a low-pressure fill. In most instances, the volume
of cement can slightly exceed that of the bone cavity to interdigitate filler from
the central bolus with the surrounding bone. Once filling is complete and the
cement has hardened the surgeon removes the cannula and closes the 3-mm
incisions.
Results
Kyphoplasty has been slowly advanced through a few surgical centers par
ticipating in a multicenter study begun in 1999. A phase I efficacy study of 70
consecutive kyphoplasty procedures in 30 patients with painful, progressive
osteoporotic/osteolytic VCFs was recently completed [29]. Mean duration of
symptoms was 5.9 months. Symptomatic levels were identified by correlating
the clinical data with MRI findings. Preoperative and postoperative x-rays were
compared to calculate the percentage height restored. Outcome was further
assessed by comparing the preoperative and latest postoperative survey of
patients self-reported health status using the 36-item Short Form Health Survey
(SF-36) [30]. In 70% of the vertebral bodies, kyphoplasty restored on average
47% of the lost vertebral height (p 0.001). SF-36 scores for bodily pain
(p 0.0001), physical function (p 0.002), and vitality (p 0.001) were
among the subscales that showed substantial and significant improvement.
Complications were infrequent. One patient experienced perioperative pul
monary edema and a myocardial infarction secondary to intraoperative fluid
overload; 2 patients suffered rib fractures due to positioning during the proce
dure. Cement leakage occurred at 6 of 70 treated levels (8.6%); however, there
were no complications that related directly to selection of this technique or to
use of the IBT. In an ongoing evaluation the results of this initial series have
been maintained in the most recent follow-up of over 70 consecutive patients
up to 14 months.
In a second prospective evaluation the safety and efficacy of kyphoplasty in
the treatment of osteolytic vertebral compression fractures due to multiple
myeloma found similar satisfying results [34]. Fifty-five consecutive kypho
plasty procedures were performed in over 27 sessions in 18 patients. The mean
age of patients was 63.5 years (4879), the mean duration of symptoms was 11
months, and the mean follow-up 7.4 months. The range of levels treated were
from T6 to L5 (T11 9, T12 7, L1 8, L2 7). There were no major com
plications related directly to the use of this technique. On average, 34% of height
lost at the time of fracture was restored. After stratifying for those where height
was not restored the remaining vertebral bodies showed an average of 56%
height restoration. Asymptomatic cement leakage occurred at 2/55 levels (4%).
Lieberman 246
dramroo
Significant improvement in SF-36 scores occurred for bodily pain: 23.2 to 55.4
(p 0.0008) and physical function: 21.3 to 50.6 (p 0.0010), vitality: 31.3 to
47.5 (p 0.010), and social functioning: 40.6 to 64.8 (p 0.014). The authors
concluded that the kyphoplasty technique was efficacious in the treatment of
osteolytic vertebral compression fractures due to multiple myeloma and associ
ated with early clinical improvement of pain and function as well as some
restoration of vertebral body height in these patients.
Vertebroplasty versus Kyphoplasty
Although both vertebroplasty and kyphoplasty provide excellent pain

relief, kyphoplasty has the potential to improve spine biomechanics and
decrease the risk of cement extravasation. PVP usually will not expand the ver
tebral body or regain normal spine alignment. Preliminary data indicate that
kyphoplasty may restore near-normal height, preventing kyphosis that leads to
respiratory and digestive problems. Restoration of height and sagittal alignment
may also work to protect vulnerable vertebral levels above or below the site(s)
treated by minimizing force transfer.
PVP is much more prone to cement leaks since the PMMA is injected in
a liquid state and will take the path of least resistance through any cracks in
surrounding bone. In administering vertebroplasty the operator injects the
liquid cement, typically pausing or stopping once a leak becomes evident. On
the other hand, in kyphoplasty, the expanded balloon creates a cavity and
pushes bone to the edges of the cavity, thus sealing off potential fissures and
cracks. Greater placement control is possible in a kyphoplasty, in which the
operator can fill the cavity with a more viscous cement to the point at which
the cement bolus reaches and interdigitates with the bony margins. The initial
kyphoplasty findings show lower rates of cement extravasation compared with
published vertebroplasty series, supporting the hypothesis that injection of
high-viscosity cement into a previously formed cavity may be an improvement
over the injection of low-viscosity liquid cement into the unreduced vertebral
body.
Conclusion
Osteoporotic and osteolytic VCFs pose a significant clinical problem

including spinal deformity, pain, reduced pulmonary function and mobility, as
well as an overall increase in mortality in the elderly. Traditional medical and
surgical options in many cases prove inadequate.

dramroo
a b
Fig. 2. Kyphoplasty example, height and kyphosis restoration. a Postheight 37 mm,

midheight 12 mm, kyphosis 25. b Postheight 37 mm, midheight 28 mm, kypho
sis 10.
PVP is a relatively noninvasive technique that has gained increased accept

ance over the last decade in the treatment of symptomatic osteoporotic verte
bral fractures. The available clinical studies describe pain relief achieved in
greater than 90% of symptomatic osteoporotic fractures, with only infrequent,
mostly minor, complications. Some of the drawbacks of PVP stem from the use
of PMMA, because of its toxicity and poor handling characteristics, rather than
from the procedure itself.
Kyphoplasty is a modification of PVP that is still in evolution. It may add
a margin of safety by virtue of a lower observed incidence of cement leakage.
Kyphoplasty may be shown to be worthwhile in acute vertebral fracture, in
high-risk patients to predictably restore vertebral height (see fig. 2) and to facil
itate a low-pressure fill procedure. Favorable outcomes in early trials appear to
permit early mobilization, which has the potential to decrease mortality.
Considering the greater mortality that is associated with osteoporotic compres
sion fractures, early mobilization in these patients is of prime importance.
The next logical step beyond treatment of evident vertebral fractures is
prophylactic augmentation. Prevention of osteoporotic vertebral fractures with
Lieberman 248
dramroo
a combination of pharmacologics and timely reinforcement of at risk osteo
porotic vertebrae is the ultimate goal aside from prevention of osteoporosis
itself. It is here that new osteoconductive synthetic composites will figure more
prominently as an emerging alternative to cement. Advances in minimally
invasive surgical techniques, imaging, and synthetic engineering are rapidly
changing the treatment protocols available for osteoporotic compression
fracture.
References
1 Heini PF, Walchli B, Berlemann U: Percutaneous transpedicular vertebroplasty with PMMA:

Operative technique and early results. A prospective study for the treatment of osteoporotic
compression fractures. Eur Spine J 2000;9:445450.
2 Einhorn TA: Vertebroplasty: An opportunity to do something really good for patients. Spine 2000;
25:10511052.
3 Kado DM, Browner WS, Palermo L, et al: Vertebral fractures and mortality in older women: A
prospective study. Study of Osteoporotic Fractures Research Group. Arch Intern Med 1999;159:
12151220.
4 Ray NF, Chan JK, Thamer M, Melton LJ 3rd: Medical expenditures for the treatment of osteo
porotic fractures in the United States in 1995: Report from the National Osteoporosis Foundation.
J Bone Miner Res 1997;12:2435.
5 Deramond H, Depriester C, Galibert P, Le Gars D: Percutaneous vertebroplasty with polymethyl-
methacrylate. Technique, indications, and results. Radiol Clin North Am 1998;36:533546.
6 Chiras J, Depriester C, Weill A, Sola-Martinez MT, Deramond H: Percutaneous vertebral surgery.
Techniques and indications. J Neuroradiol 1997;24:4559.

7 Jensen ME, Evans AJ, Mathis JM, Kallmes DF, Cloft HJ, Dion JE: Percutaneous polymethyl-
methacrylate vertebroplasty in the treatment of osteoporotic vertebral body compression frac
tures: Technical aspects. Am J Neuroradiol 1997;18:18971904.
8 Cotten A, Boutry N, Cortet B, et al: Percutaneous vertebroplasty: State of the art. Radiographics
1998;18:311320.
9 Deramond H, Wright NT, Belkoff SM: Temperature elevation caused by bone cement polymer
ization during vertebroplasty. Bone 1999;25(suppl 2):17S21S.
10 Bostrom MP, Lane JM: Future directions. Augmentation of osteoporotic vertebral bodies. Spine
1997;15(suppl 24):38S42S.
11 Riggs BL, Melton LJ 3rd: The worldwide problem of osteoporosis: Insights afforded by epidemi
ology. Bone 1995;17(suppl 5):505S511S.
12 Cooper C, Atkinson EJ, OFallon WM, Melton LJ 3rd: Incidence of clinically diagnosed vertebral
fractures: A population-based study in Rochester, Minnesota, 19851989. J Bone Miner Res
1992;7:221227.
13 Leech JA, Dulberg C, Kellie S, Pattee L, Gay J: Relationship of lung function to severity of osteo
porosis in women. Am Rev Respir Dis 1990;141:6871.
14 Lyles KW, Gold DT, Shipp KM, Pieper CF, Martinez S, Mulhausen PL: Association of osteo
porotic vertebral compression fractures with impaired functional status. Am J Med 1993;94:
595601.
15 Patel U, Skingle S, Campbell GA, Crisp AJ, Boyle IT: Clinical profile of acute vertebral
compression fractures in osteoporosis. Br J Rheumatol 1991;30:418421.
16 Galibert P, Deramond H, Rosat P, Le Gars D: Preliminary note on the treatment of vertebral
angioma by percutaneous acrylic vertebroplasty. Neurochirurgie 1987;33:166168.
17 Bai B, Jazrawi LM, Kummer FJ, Spivak JM: The use of an injectable, biodegradable calcium
phosphate bone substitute for the prophylactic augmentation of osteoporotic vertebrae and the
management of vertebral compression fractures. Spine 1999;24:15211526.

dramroo
18 Barr JD, Barr MS, Lemley TJ, McCann RM: Percutaneous vertebroplasty for pain relief and spinal
stabilization. Spine 2000;25:923928.
19 Mathis JM, Petri M, Naff N: Percutaneous vertebroplasty treatment of steroid-induced osteo
porotic compression fractures. Arthritis Rheum 1998;41:171175.
20 Cotten A, Dewatre F, Cortet B, et al: Percutaneous vertebroplasty for osteolytic metastases and
myeloma: Effects of the percentage of lesion filling and the leakage of methyl methacrylate at
clinical follow-up. Radiology 1996;200:525530.
21 Dean JR, Ison KT, Gishen P: The strengthening effect of percutaneous vertebroplasty. Clin Radiol
2000;55:471476.
22 OBrien JP, Sims JT, Evans AJ: Vertebroplasty in patients with severe vertebral compression
fractures: A technical report. Am J Neuroradiol 2000;21:15551558.
23 Cortet B, Cotten A, Boutry N, et al: Percutaneous vertebroplasty in the treatment of osteoporotic
vertebral compression fractures: An open prospective study. J Rheumatol 1999;26:22222228.
24 Martin JB, Jean B, Sugiu K, et al: Vertebroplasty: Clinical experience and follow-up results. Bone
1999;25(suppl 2):11S15S.
25 Padovani B, Kasriel O, Brunner P, Peretti-Viton P: Pulmonary embolism caused by acrylic
cement: A rare complication of percutaneous vertebroplasty. Am J Neuroradiol 1999;20:375377.
26 Cunin G, Boissonnet H, Petite H, Blanchat C, Guillemin G: Experimental vertebroplasty using
osteoconductive granular material. Spine 2000;25:10701076.
27 Leeson MC, Lippitt SB: Thermal aspects of the use of polymethylmethacrylate in large meta
physeal defects in bone: A clinical review and laboratory study. Clin Orthop 1993;295:239245.
28 Phillips H, Cole PV, Lettin AW: Cardiovascular effects of implanted acrylic bone cement. Br Med J
1971;iii:460461.
29 Lieberman IH, Dudeney S, Reinhardt M-K, Bell G: Initial outcome and efficacy of kyphoplasty
in the treatment of painful osteoporotic vertebral compression fractures. Spine 2001;26:
16311638.
30 Ware JE Jr, Gandek B: Overview of the SF-36 Health Survey and the International Quality of Life
Assessment (IQOLA) Project. J Clin Epidemiol 1998;51:903912.
31 Sundaresan N, Krol G, Digiacinto G, Hughes J: Metastatic tumors of the spine; in Sundaresan N,
Schmidek HH, Schiller AL, Rosenthal D (eds): Tumors of the Spine; Diagnosis and Clinical
Management. London, Saunders, 1990, chap 29, pp 279304.
32 McLain RF, Weinstein JN: Tumors of the spine; in Herkowitz HN, Garfin SR, Balderston RA,
Eismont FJ, Bell GR, Wiesel SW (eds): Rothman-Simeone, the Spine. London, Saunders, 1999,
chap 38, pp 11711206.
33 Rosen LS, Gordon D, Kaminski M, et al: Zoledronic acid versus pamidronate in the treatment of
skeletal metastases in patients with breast cancer or osteolytic lesions of multiple myeloma:
A phase III, double-blind, comparative trial. Cancer J, in press.
34 Dudeney S, Lieberman IH, Reinhardt MK, Hussein M: Kyphoplasty in the treatment of osteolytic
vertebral compression fractures as a result of multiple myeloma. J Clin Oncol 2002;20:
23822387.
Isador H. Lieberman, B.Sc. MD, FRCS(C)
Department of Orthopaedic Surgery, A-41
The Cleveland Clinic Foundation, 9500 Euclid Ave., Desk A-41
Cleveland, OH 44195 (USA)
Lieberman 250
dramroo
Minimally Invasive Update
Endoscopic Posterior
Cervical Foraminotomy and
Microdiscectomy
Mick J. Perez-Cruet, Richard G. Fessler

Institute for Spine Care, Chicago Institute of Neurosurgery and Neuroresearch,
Rush-Presbyterian-St. Lukes Medical Center, Chicago, Ill., USA
Recently microendoscopic discectomy (MED) techniques have been

applied to perform laminoforaminotomy and discectomy for unilateral cervi
cal radiculopathy [1, 11] utilizing the MED technique and instrumentation
(fig. 1) developed by Smith and Foley in 1997 for lumbar disc disease [4, 6].
The muscle-splitting approach used in this technique is effective in limiting
postoperative pain and muscle spasms while maintaining the integrity of
midline posterior muscular and ligamentous attachments to the spine.
Minimally invasive posterior cervical microendoscopic discectomy and
laminoforaminotomy (CMED) can provide a number of advantages including
reduced approach-related morbidity, preservation of the motion segment,
reduced patient postoperative pain and discomfort, and quicker patient recov
ery. In avoiding an anterior approach to the spine and maintaining segmental
motion the incidence of hastened adjacent level disc degeneration may be
reduced [3, 9, 10]. The muscle splitting approach used in this technique as
apposed to the muscle stripping approach used in traditional posterior cervical
discectomy and laminoforaminotomy may help to reduce paraspinal muscle
denervation. Despite these many benefits there is a steep learning curve to
mastering this technique, which may require additional training under the
guidance of an experienced minimally invasive spine surgeon. Once trained in
these techniques the results can be very satisfying for both clinician and
patient. This chapter will focus on the technical considerations used in
performing CMED effectively and safely.
dramroo
a b
c d
Fig. 1. METRx instrumentation used for performing CMED with figures showing
K-wire, sequential dilators and tubular retractor (a), endoscopic assembly (b), long tapered
Drill (c), and specialized instruments (d).
Background
Posterior cervical foraminotomy and microdiscectomy is a proven effec

tive technique for the management of lateralized herniated disc and foraminal
stenosis [2, 5, 7, 8, 12, 1719]. The endoscopic approach is a further evolution
of this technique with a primary benefit of reduction in approach-related
morbidity. This is accomplished with a number of tubular dilators that allows
for muscle splitting of the longitudinal muscular fibers to reach the facet
Perez-Cruet/Fessler 252
dramroo
complex as opposed to striping the muscle from the spinous process and lamina
and aggressively retracting the muscles laterally to gain access to the facet
complex. The instrumentations used to perform this technique were originally
designed for the treatment of lumbar disc herniation (fig. 1). The ability to
perform CMED was first determined in cadaveric studies [16]. These studies
determined that the METRx system could be used effectively to treat degenera
tive cervical disease and then was applied safely in a clinical setting [1, 11].
Evaluation
Patient workups routinely include plain anteroposterior, lateral and oblique

x-ray views to determine spine alignment, disc space height and foraminal
encroachment. Additional radiographic evaluation of the cervical spine will
include either a magnetic resonance image (MRI) or myelogram, and/or com
puted tomographic myelogram (CTM) to visualize the area of neural compres
sion. CTM is particularly helpful in multilevel degenerative cervical disc
disease to determine the level of maximal neural compression since this study
clearly shows bone pathology, as well as foraminal stenosis and nerve root com
pression. In conjunction with thorough clinical history and physical examina
tion, radiographic evaluation will help to determine the operative level. Further
assessment can include selective nerve root blocks and or electromyelographic
studies. Ultimately, it is critical to correctly identify the anatomic level from
which the patients radiculopathy originates in order to achieve surgical
success.
Indications
Endoscopic posterior cervical foraminotomy is indicated for cases of

lateralized disc herniation (fig. 2), osteophyte compression, and foraminal
stenosis. Ideally, patients should present with painful cervical radiculopathy,
which correlates with neural compression seen on MRI, myelogram, and/or
CTM. This procedure is not recommended for cases of cervical myelopathy
from spondylosis or central disc herniation. Whereas traditional surgical teach
ing guides us to address pathology from an anterior approach when there is
straightening or kyphosis of the spine, the endoscopic posterior approach can
be used for cases of posterolateral disc herniation or foraminal stenosis with
mild to moderate kyphosis so long as there is no evidence of instability.
For these patients, particular caution must be taken to avoid resecting too much
of the medial facet complex during the decompression. Studies have shown that
Cervical Foraminotomy and Microdiscectomy 253

dramroo
Fig. 2. Lateralized left C6,7 disc herniation ideal for treatment using the CMED
approach.
up to 50% of the facet complex can be resected unilaterally without inducing

iatrogenic instability [15]. Since the 30 angled endoscope allows for more
extensive facet resection, care is taken during this procedure to avoid taking too
much facet [16].
It is important to realize that even with a good clinical history of radiculo
pathy and a seemingly appropriate radiographic finding, a potential for mis
diagnosis can exist. As the incidence of multilevel disc degeneration is high on
MRI and CTM for elderly patients, it is statistically not uncommon that some
degree of pathological change will be seen at any given level on imaging stud
ies. For such patients, a careful search to exclude other causes of nerve root pain
should be completed prior to operative intervention. The differential for nerve
root pain includes spinal canal tumors, trauma, inflammatory diseases,
demyelinating conditions, toxic and allergic conditions, hemorrhage, congeni
tal defects, metabolic diseases, neuropathies, thoracic outlet syndrome, rotator
cuff pathology, impingement syndromes, bursitis, arthritis of the shoulder, and
bicipital tendonitis. It is important for spinal surgeons to be familiar with these
disorders in order to properly exclude them from the list of possible etiologies
before subjecting the patient to unneeded surgery. Electromyographic (EMG)
and nerve conduction velocity studies done by a competent neurologist or
physiatrist may be particularly helpful in this regard. Lastly, it is prudent to
perform a psychological screening of all patients prior to performing this
dramroo
250.0
400.0
350.0
Average blood loss
200.0
Operative time
300.0
250.0 150.0
200.0
150.0 100.0
100.0 50.0
50.0
0.0 0.0
a All Sitting Prone b All Sitting Prone
Fig. 3. Operative blood loss (a) and time (b) when performing the CMED procedure
in all patients combined, in the semisitting position, and in the prone position.
procedure and to rule out secondary gain or psychological issues that may result
in a poor surgical outcome.
Anesthesia and Operative Setup
After appropriate preoperative evaluation and medical clearance, patients

were brought to the operative suite. Following the induction of general endo
tracheal anesthesia, adequate intravenous access was secured. An arterial line is
placed to enable continuous blood pressure monitoring and control. Adequate
blood pressure is maintained to assure spinal cord perfusion during this proce
dure. During the initial experience with the CMED technique, patients were
placed in the prone position. However, this led to bleeding that often obscured
the endoscopic image during the operative procedure and resulted in increased
operative times and blood loss (fig. 3). A change to the semisitting position
using a Mayfield head holder (fig. 4) has resulted in significantly improved
operative visualization, relieved epidural venous congestion, and decreased
operative blood loss and operative times (fig. 3). The advantage of this position
is that blood does not accumulate at the bottom of the tubular retractor. Blood
loss was found to be substantially less and the procedure can be preformed with
the tubular retractor positioned at a comfortable height for the surgeon (fig. 5).
Because of the slightly increased risk of air embolism, precordial Doppler as
well as end tidal pCO2 are measures. We do not routinely place a central venous
line. Adequate intraoperative measures taken in the event of an air embolism
should be familiar to the operative team prior to undertaking this procedure.
Before final patient positioning, utmost care is directed at ensuring that the
cervical spine and neck musculature are not kinked or held in an unfavorable

dramroo
a b
Fig. 4. Patient positioned in a semisitting position with head affixed in a Mayfield

head holder (a) with lateral C-arm in place (b).
Fig. 5. The position of the surgeon, while standing behind the patient, viewing the
endoscopic image on the operative monitor. Note the ergonomically favorable position of the
surgeons hands and body posture.
position. The neck is placed in a neutral or slightly flexed position to ensure

adequate jugular venous drainage and the head is secured in the Mayfield
holder to assure direct midline positioning. In most instances a Foley catheter
is not placed except in instances where more than one level is performed.
dramroo
Somatosensory evoked potentials, as well as EMG recordings, are measured to
further ensure the safety of the procedure. After the initial induction of anes
thesia, we have refrained from the use of neuromuscular paralytics to allow for
improved feedback from the nerve root(s) during the operation. A single
preoperative dose of either Ancef or vancomycin is used. We do not routinely
employ Solumedrol or other glucocorticoids for neural protection.
The fluoroscopic C-arm is brought into the surgical field and positioned
with the arc over the patient so that real-time lateral fluoroscopic images can
easily be obtained (fig. 6). Although we have not typically use anteroposterior
images, they can also be utilized to facilitate docking of the initial K-wire and
tubular retractors on the facet complex. The surgeon generally stands directly
behind the neck of the patient with the video and fluoroscopic monitors placed
within direct view of the surgeon to allow optimal ergonomic flow during the
procedure (fig. 5).
Surgical Technique
We employed the METRx system (Medtronic Sofamor-Danek, Memphis,

Tenn., USA) of endoscopic retractors, camera, and instruments for the proce
dures (fig. 1). As this system was initially designed for lumbar discectomy,
several modifications have been made to better optimize its use in the CMED
procedure. These include lengthening of the table-mounted retractor as well as
a better selection of smaller profile curettes and Kerrison punches for use in the
more delicate cervical spine. An initial stab incision was made approximately
12 cm off midline ipsilateral to and at the level of the pathology. Under
fluoroscopic guidance, a K-wire was inserted through the posterior cervical
musculature and fascia down to the facet complex or lateral mass of the opera
tive level. Particular caution was taken at this point to insure that the guidewire
was docked on bone to avoid inadvertent dural penetration medially or slipping
off the facet complex laterally (fig. 6). Once the guidewire is docked on the
facet complex in question, the skin incision was extended above and below the
K-wire for a total length of approximately 2.0 cm. The skin edges are retracted
with a small retractor and the cervical fascia incised using either a Metzenbaum
scissors or Bovey cautery. Care should be taken not to cut muscle fibers during
this procedure as this can cause unnecessary blood loss. Due to the relative
thickness of the posterior cervical fascia, this sharp opening of the fascia is
required to allow for easier and safe passage of the sequential dilating cannulas
with a minimum use of force. If a skin-covering barrier such as ioband was
placed, it should be circumferentially removed from around the skin incision to
prevent plastic sequestra that can occur during placement of the percutaneous

dramroo
a b c
d e f
Fig. 6. Lateral fluoroscopic images showing docking of the K-wire on the facet com
plex (a), first dilator (b), second dilator after K-wire is removed (c), third dilator (d), fourth
dilator (e), and tubular retractor in place with curette-identifying facet (f).
dramroo
soft tissue dilators. It is important to remove the K-wire after the initial dilator
is passed. A series of dilators are then sequentially inserted using a gentle
downward rotating maneuver through the posterior neck musculature, over
which an 18-mm tubular retractor was inserted. Real-time lateral fluoroscopic
images were obtained throughout the above procedure to insure proper docking
of the sequential dilators and tubular retractors on the facet complex (fig. 6af).
The working channel (tubular retractor) is then attached to a flexible arm
affixed to the operating table side rail and locked in position. The retractor arm
is positioned to avoid obscuring the lateral fluoroscopic image. The endoscope
is then attached to the tubular retractor via a circular plastic friction couple.
Additionally, the endoscopic camera used today has far superior resolution and
clarity than the cameras used during our previous cadaveric studies and initial
operative cases. Further modifications of the instruments are ongoing to
improve the safety, efficacy, and ease of this procedure.
Once the tubular retractor is set in the desired position, a Bovey cautery
with a long insulated tip is used to remove the remaining muscle and soft tissue
overlying the facet complex. A small straight or up-going curette in conjunction
with lateral fluoroscopic imaging can further define the bone anatomy and local
ization of surgical anatomy (fig. 6f). The dissection is initially started laterally
where the bone is easily palpated with the Bovey tip. Once the bone of the lat
eral facet complex in exposed, the dissection of muscle off the facet complex
continues medial to expose the laminofacet junction with care not to enter the
interlaminar space with the Bovey tip. To prevent Bovey cautery smoke from
obscuring the endoscopic image during this procedure a frequent on/off tech
nique is used that allows smoke to clear from the tube before proceeding.
A suction tubing is also attached to the endoscopic device. If the tubular retractor
is properly placed initially only a small piece of muscle tissue needs be removed
to expose the facet complex. Often the ligamentum flavum is thinned or
altogether absent near the lateral edge of the interlaminar space thereby placing
the dura and spinal cord at higher risk. With the bone well visualized the inferior
edge of the superior lamina and the medial edge of the lateral mass-facet
complex are identified with a small straight or up-going endoscopic curette. The
facet complex at the proper level is clearly identified before proceeding.
Bleeding from epidural veins is controlled using a long tipped endoscopic
bipolar cautery. For bleeding underneath the edge of the lamina, angled bipolar
forceps with a 45 angle are often useful. After the medial facet plane has been
clearly defined, a small angled 1- or 2-mm Kerrison rongeur is used to begin the
foraminotomy (fig. 7a). Periosteal and bone bleeding is addressed with bone
wax and cautery. In cases of marked facet arthropathy and enlargement, a drill
with a long endoscopic bit (e.g. AM-8 bit with Midas Rex or TAC bit with
MEDNext drill) can be used to further thin the medial facet and lateral mass.

dramroo
a b
c d
Fig. 7. Intraoperative images showing Kerrison punch initiating medial facet removal
(a), nerve probe passing out of the foramen once foraminotomy is performed with underly
ing nerve root exposed (b), lateral fluoroscopic image with down-going curette on disc under
nerve root (c), and endoscopic image showing removal of disc fragment (d).
Frequent dissection of the soft tissue off the bone with an angled curette facili
tates safe use of the Kerrison rongeur. In this fashion, the decompression is care
fully continued inferiorly and laterally along the course of the neural foramen.
The laminoforaminotomy is completed when the nerve root had been well
exposed along its proximal foraminal course. The adequacy of the decompres
sion should be confirmed by palpating the root along its course with a small
nerve hook (fig. 7b).
In cases where a herniated cervical disk or free disc fragment is present,
additional exposure is obtained by drilling a small portion of the superomedial
dramroo
pedicle directly below the exiting nerve root. As the nerve root lies directly
against this portion of the pedicle, removing a small portion of the pedicle will
create enough space for passing a down-going curette (fig. 7c) under the nerve
root to remove a disc fragment without traumatizing the nerve root (fig. 7d).
Additionally osteophytes encountered in this region can also be drilled or curet
ted as needed. Prior to closure the nerve root is palpated along its anterior
surface to ensure that no residual compression exists along its course.
Far lateral foraminal stenosis or disc herniations can also be decompressed
through this approach. As the nerve root often passes in close proximity to the
vertebral artery laterally, particular attention should be paid during decompres
sion in this area. Inadvertent passage of instruments beyond the bone defining
the posterior margin of the foramen transversarium should be avoided. Brisk
dark bleeding is often encountered from the rich venous plexus, which typically
surrounds the space around the vertebral artery. When encountered, such bleed
ing should serve as a useful warning to limit further dissection and thus prevent
inadvertent arterial injury. Unnecessary excessive decompression of the facet
should be avoided to prevent iatrogenic instability of the cervical motion
segment. Raynor et al. [15] concluded that the integrity of the majority of the
facet joint is essential for stability and that no more than 50% of the facet
should be removed to maintain its integrity.
Wound Closure
After inspection of the nerve root, meticulous hemostasis should be

obtained by a combination of bipolar cautery and gentle tamponade with
thrombin-soaked gel-foam pledgets. The area is then copiously irrigated with
lactated ringers impregnated with bacitracin antibiotics. Although optional, we
have usually placed a small piece of gel-foam soaked with Solumedrol gently
over the laminoforaminotomy defect. Use of epidural morphine paste or simi
lar cocktails is reasonable if there is no evidence of dural erosion or tear.
Alternatively, Marcaine (0.25%) can be injected around the incision. Such
agents may help to reduce postoperative pain and allow for more rapid recov
ery and ambulation. The tubular retractor and endoscope are then removed and
a routine closure of the fascia and skin performed. As the defect is typically
quite small, only a limited amount of closure need be performed and a drain is
not needed. A 0-Vicryl-type reabsorbable stitch is used to close the lumbodor
sal fascia in a figure of 8. Inverted 2-0 Vicryl stitches are used to reapproximate
the subcutaneous layer. Dermabond is used to reapproximate the skin edges and
no dressing is applied as the Dermabond will adhere to it. Dermabond is attrac
tive as it keeps the skin edges closely approximated for a 7- to 10-day period as

dramroo
well as providing a waterproof barrier. The patient can thus shower almost
immediately after surgery.
Postoperative Care
The patient was then awakened from anesthesia and taken to the postanes
thesia recovery unit. Most patients have the procedure on an ambulatory basis;
therefore, long-acting inhalational and intravenous agents should be avoided to
allow for rapid awakening of the patient postoperatively. Additionally, use of
only short-acting muscle relaxants for initial induction will allow for better
monitoring of nerve root function as well as quicker extubation of the patient
after surgery. This procedure performed on an outpatient basis requires
thorough perioperative patient education [14].
Once in the postanesthesia recovery area, the patients are allowed to
rapidly mobilize and ambulate as tolerated. Arterial and intravenous lines are
removed early on. If a Foley catheter was placed it is generally removed before
the patient leaves the operative suite. Done correctly, this procedure does not
result in either instability or fusion of the operated cervical motion segment.
Therefore no cervical collar is required. Soft collars and other comfortable
semirigid collars can be given to patients for their comfort if desired. It is
important to emphasize to patients, however, that chronic dependence on such
orthosis will only lead to further deconditioning of the cervical musculature.
Depending on their preoperative medications, patients are typically discharged
on a combination of muscle relaxants (e.g. baclofen or Flexeril), nonsteroidal
anti-inflammatories (e.g. Toradol, Vioxx, or Celebrex), and an oral opioid for
breakthrough pain (i.e. Vicodin or Darvocet). When we compared our patients
with patients treated via open cervical foraminotomy, we found that the
microendoscopic foraminotomy (MEF) group used significantly less pain med
ications postoperatively than did the open group [11]. Patients undergoing this
procedure typically recovered rapidly with only mild to moderate discomfort
upon discharge. Of our last 30 cases, the majority of patients were discharged
in 6 h or less.
Complication Avoidance
Complications can be avoided by having a thorough understanding of

cervical anatomy, proper training in endoscopic spinal techniques and a
knowledge of possible complications. Adamson [1] retrospectively reviewed 100
cases of patients undergoing endoscopic posterior cervical laminoforaminotomy
dramroo
(MEF) and reported complications in 3 patients; 2 cases of dural puncture
required no intervention other than gel-foam and 1 case of superficial wound
infection was reported. In a series by Khoo et al. [11] three complications
occurred in 25 patients and were attributable to surgical technique. These
included two small cerebral spinal fluid (CSF) leaks and 1 case of partial-
thickness dural violation. For the 2 cases where a CSF leak occurred, no direct
repair was required as the durotomy was very small. After 23 days of routine
lumbar drainage for patients with CSF leaks, none of these patients went on
to have long-term clinical sequelae of chronic CSF leak or symptomatic
pseudomeningocele. Thus we have routinely employed a lumbar drain for 23
days postoperatively to help closure of the small dural tear. Additional adjuncts
such as fibrin glue, fat or muscle grafts can also be used. Spinal headaches and
nausea associated with the lumbar drainage were treated symptomatically with
nonsteroidal anti-inflammatory medications and bed rest. If a large durotomy
occurs, direct dural repair can be attempted if specialized instruments are avail
able for use through the endoscopic tube. Castro-Viejo-type needle holders and
long forceps are particularly useful in this regard. In rare instances, conversion
to an open procedure may be necessary to close large dural violations. To date,
we have not had problems with delayed pseudomeningoceles or continued CSF
leaks. The risk of dural injury can be reduced with experience in performing
this technique with most durotomies occurring on initial patients undergoing
this procedure. Patients should be advised of this potential complication and
informed that with appropriate management durotomy results in no adverse
clinical result.
A potential complication, which has not been reported, is iatrogenic injury
during the surgical approach and muscle dilatation portion of this operation.
Unlike the thoracodorsal fascia, the posterior cervical fascia is very thick and
must be cut under direct visualization to prevent hyperextension of the neck
during insertion of the dilators. The initial K-wire or smaller dilators can be
inadvertently pushed between the cervical lamina resulting in nerve root or
spinal cord injury; therefore, this portion of the procedure is performed under
lateral fluoroscopic guidance. An anteroposterior fluoroscopic image can also
help in safely docking the K-wire and subsequent dilators on the facet complex.
Additionally, lateral displacement of the K-wire or dilators can result in nerve
root or vertebral artery injury. Brisk venous bleeding can also result if the
dilators inadvertently slip lateral to the facet complex during placement. This is
controlled with gentle gel-foam packing and/or bipolar cautery. The K-wire is
removed after the first dilator is passed and subsequent dilators securely docked
onto the laminar facet junction.
To help reduce intraoperative bleeding the CMED procedure is performed
in the semisitting position (fig. 4). A series of patients comparing cervical

dramroo
laminoforaminotomy performed in the open prone position, microendoscopic
prone position, or microendoscopic sitting position revealed reduced operative
times, estimated blood loss, postoperative length of stay, and pain medication
requirements in the microendoscopic sitting position group [11]. The sitting
position significantly reduces epidural venous engorgement, thus decreasing
blood loss. In addition, this position allows blood to flow out of the tubular
retractor, rather than accumulating and obscuring the endoscopic view of the
operative field. Although none of the sitting position patients experienced an air
embolism the potential increased risk for air embolism precludes the use of a
pericardial Doppler even though the incidence of clinically significant embolic
event remains extremely low.
Conclusion
The endoscopic posterior cervical laminoforaminotomy and discectomy

technique is a safe and effectively treatment of posterior lateral cervical disc
herniation and/or foraminal stenosis. The learning curve associated with this
technique requires additional training and we recommend that this training be
performed under the guidance of an experienced endoscopic minimally inva
sive spine surgeon. Cadaveric training can further familiarize the surgeon with
this technique. The clinical results are very satisfying since patients experience
less postoperative pain and can be discharged on the same operative day.
References
1 Adamson TE: Microendoscopic posterior cervical laminoforaminotomy for unilateral radiculo

pathy: Results of a new technique in 100 cases. J Neurosurg 2001;95(1 suppl):5157.
2 Aldrich F: Posterolateral microdiscectomy for cervical monoradiculopathy caused by postero
lateral soft cervical disc sequestration. J Neurosurg 1990;72:370377.
3 Baba G, Furusawa N, Imura S, et al: Late radiographic findings after anterior cervical fusion for
spondylotic myeloradiculopathy. Spine 1993;18:21672173.
4 Brayda-Bruno M, Cinnella P: Posterior endoscopic discectomy (and other procedures). Eur Spine
J 2000;9:S24S29.
5 Fager CA: Posterolateral approach to ruptured median and paramedian cervical disc. Surg Neurol
1983;20:443452.
6 Foley KT, Smith MM: Microendoscopic discectomy. Tech Neurosurg 1997;3:301307.
7 Henderson CM, Henney RG, Shuey HM, et al: Posterolateral foraminotomy as an exclusive
operative technique for cervical radiculopathy: A review of 846 consecutively operated cases.
8 Herkowitz HN, Kurz LT: Surgical management of cervical soft disc herniation: A comparison
between the anterior and posterior approach. Spine 1990;15:10261030.
9 Hilibrand AS, Yoo JU, Carlson GD, et al: The success of anterior cervical arthrodesis adjacent to
a previous fusion. Spine 1997;22:15741579.
dramroo
10 Hunter LY, Braunstein EM, Bailey RR: Radiographic changes following anterior cervical fusion.
Spine 1980;5:399401.
11 Khoo LT, Perez-Cruet MJ, Laich DT, Fessler RG: Posterior cervical microendoscopic foramino
tomy; in Perez-Cruet MJ, Fessler RG (eds): Outpatient Spinal Surgery. St. Louis, Quality Medical
Publishing, 2002, pp 7193.
12 Krupp W, Muke R: Clinical results of the foraminotomy as described by Fryholm for the treatment
of lateral cervical disc herniation. Acta Neurochir (Wien) 1990;107:2229.
13 Perez-Cruet MJ, Smith M, Foley K: Microendoscopic lumbar discectomy; in Perez-Cruet MJ,
Fessler RG (eds): Outpatient Spinal Surgery. St. Louis, Quality Medical Publishing, 2002,
pp 171183.
14 Perez-Cruet MJ, Rice-Wyllie L, Pieper DR: Patient education; in Perez-Cruet MJ, Fessler RG
(eds): Outpatient Spinal Surgery. St. Louis, Quality Medical Publishing, 2002, pp 3547.
15 Raynor RB, Pugh J, Shapiro I: Cervical facetectomy and its effect of spine strength. J Neurosurg
1985;63:278282.
16 Roh SW, Kim DH, Cardoso AC, Fessler RG: Endoscopic foraminotomy using microendoscopic
discectomy system in cadaveric specimens. Spine 2000;25:260264.
17 Simeone F, Dillin W: Treatment of cervical disc disease: Selection of operative approach.
Contemp Neurosurg 1986;8:16.
18 Tomaras CR, Blacklock JB, Parker WD, Harper RL: Outpatient surgical treatment of cervical
radiculopathy. J Neurosurg 1997;87:4143.
19 Williams RW: Microcervical foraminotomy. A surgical alternative for intractable radicular pain.
Spine 1983;8:708716.
Mick J. Perez-Cruet, MD, MS
Institute for Spine Care, Department of Neurosurgery
Chicago Institute of Neurosurgery and Neuroresearch,
Rush-Presbyterian-St. Lukes Medical Center, 1725 West Harrison Street
Suite 970, Chicago, IL 60612 (USA)

dramroo
An Update on Endoscopic Thoracic

Spinal Surgery: Thoracic
Microendoscopic Discectomy
Mick J. Perez-Cruet a, Bong-Soo Kimb, Faheem Sandhub,
Richard G. Fessler b
a
Directory of Minimally Invasive Spine Surgery, Department of Neurosurgery,
Chicago Institute of Neurosurgery and Neuroresearch, Rush-Presbyterian
St. Lukes Medical Center and
b
Division of Neurosurgery, University of Chicago, Chicago, Ill., USA
Clinically significant thoracic disc herniation is a relatively rare condition

with an estimated annual incidence of approximately one per million. Surgical
management of thoracic disc herniation comprises between 0.15 and 4% of all
operations for herniated discs including those of the cervical and lumbar spine
[3, 40, 41].
Several surgical techniques have been developed to approach the thoracic
spine including costotransversectomy, lateral extracavitary, lateral parascapular
extrapleural, transfacet, transpedicular, and transthoracic approaches [1, 17, 41].
Although all of these approaches have been successful in reaching thoracic
pathology, each procedure requires a relatively large skin incision and extensive
bony work that subsequently may lead to considerable postoperative morbidity
[4, 12, 37, 41].
Outcomes from the traditional posterior approach, thoracic laminectomy, to
treat thoracic disc herniation are especially poor with large percentages of
patients failing to improve or, more significantly, deteriorating further. Although
anterior approaches facilitate exposure to the ventral aspect of the thecal sac,
they require entry into the pleural cavity and place the intrathoracic contents at
risk. Posterolateral approaches, especially the lateral extracavitary one, provide
many of the benefits of an anterior approach without the additional risks of
entering the thoracic cavity. Unfortunately, the exposure for this approach
requires extensive muscle dissection that adds to the postoperative morbidity.
dramroo
Perot and Munro [33] and Ransohoff et al. [34] performed the initial
transthoracic approach to thoracic disc herniation after which the transthoracic
approach became a standard procedure to manage thoracic disc herniations. This
approach provides excellent visualization of the ventral aspect of the thoracic
spine without the risk of spinal cord manipulation. However, the thoracotomy
used in this approach requires a large skin incision, extensive lung and rib retrac
tion, and muscle dissection, all of which can contribute to postoperative
pulmonary dysfunction, pain and increased morbidity.
To reduce the morbidity associated with the thoracotomy technique and
other approaches previously mentioned, less invasive thoracoscopic techniques
have been developed and more recently refined for the treatment of herniated
thoracic disc [1, 8, 9, 35, 36]. Thoracoscopy is capable of producing the same
exposure as the transthoracic route without the need for a large thoracotomy
incision. Thoracoscopy was introduced by Jacobeus in 1910 [19]. However, this
technique was not widely adopted at that time due to poor illumination of
intrathoracic structures and other technical limitations. It was not until the
1970s that the modern era of surgical endoscopy began following a number of
technical advancements such as fiberoptics, and the operating endoscopes.
Introduction of the video camera into thoracoscopic surgery was the next major
advance leading to the introduction of the video-assisted thoracoscopic surgery
technique [14, 18, 19, 35]. Thoracoscopic discectomy is an effective alternative
to traditional open thoracotomy techniques that has been found to reduce the
incidence of pulmonary morbidity [11, 13], intercostal neuralgia, and shoulder
girdle dysfunction [13, 20]. Although the incidence thoracoscopic complica
tions may vary they are similar to those encountered with open thoracotomy
[16, 30].
Although the procedure-associated morbidity is much less than with
thoracotomy, the prevalence of pulmonary complications such as postoperative
atelectasis, pneumothorax, pleural effusion, and hemothorax is considerable
[7, 10, 22, 37, 39]. In addition, thoracoscopic surgery is technically demanding
and requires attainment of new surgical skills. Therefore, it is advised that
surgeons perform the thoracoscopic procedures only after pursuing appropriate
training that includes extensive practice of this skill in a surgical laboratory [4].
Even today the use of video-assisted thoracoscopic surgery techniques to treat
thoracic disc herniation have not been widely adopted due to the steep learning
curve which requires specialized training to master.
The primary indication for thoracoscopic spinal surgery is management of
herniated discs that compress the spinal cord and/or nerve root. Severe or
progressive myelopathy from spinal cord compression caused by disc herniation
is an absolute indication for surgery [19, 27]. In addition to thoracic discectomy,
current indications for thoracoscopic spine surgery include tissue biopsies,
Thoracic Microendoscopic Discectomy 267

dramroo
thoracic paravertebral abscess drainage and debridement, anterior spinal release
and/or fusion for spinal deformity, stabilization and fusion of thoracic and
thoracolumbar fractures, corpectomy for vertebral tumors and placement of
anterior spinal instrumentation for fusion [2, 58, 2326, 29, 36].
We have developed a novel posterolateral, minimally invasive thoracic
microendoscopic discectomy (TMED) technique that potentially provides a less
morbid approach to the treatment of thoracic disc herniation. The procedure is
performed via a posterolateral approach with the patient in the prone position
(fig. 1a) and therefore avoids entry into the thoracic cavity. A number of studies
have shown that minimally invasive spine surgery is associated with less disrup
tion of normal tissue, less blood loss, and less procedure time than traditional
open procedures, which translates to patients experiencing reduced postopera
tive pain, reduced hospitalization, and reduced recovery time [1, 4, 15, 32].
Based on our validation data from cadaveric studies we have demonstrated that
this technique could be safely implemented into clinical use. The technique is a
modification of the lateral extracavitary approach, which was a further modifi
cation of the lateral rachiotomy initially performed in 1961 for the treatment of
Potts disease by Larson et al. [21]. Unlike the traditional lateral extracavitary
approach which requires significant muscle, rib head and transverse process
excision to reach the thoracic disc, the TMED approach avoids much of this
dissection. It uses a series of muscle dilators, a tubular retractor and endoscope
that was initially developed to approach lumbar herniated disc pathology [31].
Herein we report this novel minimally invasive technique to treat herniated
thoracic disc pathology.
Indication
The clear indication to perform the TMED procedure is in patients with

thoracic disc herniation causing myelopathy in order to improve neurological
function and prevent further injury to the spinal cord. However, the longer the
myelopathy has been present the less likely is restoration of normal neurological
function. Severe and refractory radiculopathy is another indication for surgery.
Thoracic radicular pain from disc herniation is typically sharp, lancinating
pain that radiates unilaterally from posterior to anterior around the chest wall in
its dermatomal distribution correlating to the herniated thoracic disc.
Radiculopathy can develop without loss of neurological function because of the
overlap of innervation of adjacent intercostal nerves. Radiculopathy may be pre
sent without myelopathic manifestations. Isolated thoracic radiculopathy often
responds to conservative management such as nonsteroidal anti-inflammatory
medications, oral steroids, epidural injection, intercostal nerve injection, activity
Perez-Cruet/Kim/Sandhu/Fessler 268
dramroo
modification, and hyperextension brace [41]. Pain relief after thoracic discec
tomy is usually favorable in patients with radiculopathy, even for those with
long-standing radiculopathy. Although it is controversial whether patients with
localized back pain and axial pain without myelopathy require surgery or not,
most surgeons do not recommend thoracic discectomy for isolated back pain
unless it is associated with a neurological deficit [28, 29, 37, 38, 40]. This tech
nique is not recommended for a large calcified midline thoracic disc that may
require an anterior approach.
Technique
Posterolateral Microendoscopic Discectomy

We have developed a novel posterolateral, minimally invasive TMED
technique that potentially provides a less morbid treatment for thoracic disc
herniation compared to traditional techniques. This method has been used
successfully to treat lateralized and central thoracic disc herniations causing
radicular and myelopathic symptoms. This technique was developed with similar
instruments used to perform a microendoscopic lumbar discectomy using the
METRx system of tubular dilators and retractor initially developed by Smith and
Foley, see Perez-Cruet et al. [31]. The procedure is performed as follows.
Patient Positioning and Operative Setup
After general endotracheal induction and placement of a Foley catheter,

the patient is positioned prone on a radiolucent Wilson frame. All pressure
points are padded adequately and the arms are placed above the patient making
sure not to overextend at the shoulder beyond 90 (fig. 1a). Somatosensory
evoked potentials are measured throughout the procedure. Any change in
somatosensory evoked potentials during the procedure warrants investigation
as to the probable cause. The fluoroscopic C-arm is positioned to allow for a
lateral x-ray image of the operative area. A video monitor is placed opposite the
surgeon to allow for a clear field of view of the monitor. The back is shaved and
prepped in a routine surgical fashion.
Under fluoroscopic guidance a spinal needle was used to locate the level
of the herniated thoracic disc. The exact location of the thoracic disc herniation
was confirmed by counting from the sacrum below and/or from the occiput
above. Once the correct location was identified, a Steinmann pin was placed at
the medial aspect of the transverse process at the level of the thoracic disc
herniation. An incision of approximately 2 cm was made 34 cm lateral to the

dramroo
midline, depending on the size of the patient, through which a series of tubular
muscle dilators was placed (fig. 1b, c). A tubular retractor was placed over the
final muscle dilator and then affixed to a flexible arm secured firmly to the
operative table (fig. 1d). The endoscopic assembly was then focused, white
balanced, and passed down the center of the tubular retractor (fig. 1e). Once the
endoscope was in place the orientation of the endoscopic image seen on the
monitor was adjusted so that medial anatomy was at the top of the monitor
screen, lateral anatomy at the bottom, and rostral and caudal anatomy was in the
same orientation as the patient on the operative table. This greatly facilitated the
procedure and aided in orienting the surgeon throughout the procedure.
Surgical orientation throughout the case is of upmost importance to prevent
inadvertent entrance into the spinal canal and this is facilitated by lateral and
AP fluoroscopic images to help confirm the location of dissection. The proper
level was then reconfirmed using lateral fluoroscopy and the TMED procedure
was performed under endoscopic visualization. The muscle and soft tissue over
lying the proximal portion of the transverse process and lateral facet complex
were removed using an insulated Bovey cautery. Careful probing with a ball-
ended probe helps in defining the bone margins and liberal use of fluoroscope
imaging also helps to orient the surgeon thoroughout the procedure. The
approach is similar to that of a transforaminal approach and adequate laterality
of the approach reduces any manipulation of the spinal cord during disc
removal. Access to the disc space required removal of a small amount of bone.
This is facilitated by the use of a long tapered high-speed drill. Once the bone
architecture is defined a small portion of the proximal rib head and transverse
process and a portion of the lateral aspect of the facet complex are removed. This
procedure facilitated widening of the neural foramen. The thoracic pedicle of the
more caudal vertebral body was identified and followed to the appropriate disc
level. The disc annulus was identified and the epidural veins coagulated with
a bipolar cautery. The annulus was cut with a thin anulotomy knife and the
discectomy was performed. The lateral projectory of the tubular retractor
through a transforaminal type approach allowed for extensive disc removal under
the thoracic spinal cord and dura enabling even midline placed thoracic disc
herniations to be removed (fig. 2a, b). Following complete discectomy further
inspection underneath the dura and spinal cord using a Woodson elevator or
similar instrument assured complete decompression (fig. 1f). The tubular retrac
tor was removed and a suture was placed in the thoracodorsal fascia. Interrupted
subcuticular sutures were placed and the skin closed with Dermabond.
This technique has a number of advantages over other traditional tech
niques used to perform thoracic discectomy and includes the following: avoid
ance of traversing through the thoracic cavity, maintaining the integrity of the
disc, avoidance of the need for thoracic fusion, and avoidance of extensive
dramroo
a b
c d
e f
Fig. 1. a Positioning of the patient in the prone position on a radiolucent frame with
lateral fluoroscopy for performing TMED procedure. After placing a K wire on the trans
verse process (b, c) a series of muscle dilators are used over which a tubular retractor is
placed (d). e The operation is then performed under endoscopic visualization. f A Woodson
elevator instrument is useful in removing a central disc herniation.

dramroo
T6, 7
a b
10
c d
T9, 10
T9, 10
e f
Fig. 2. Pre (a, c, e)- and respective post (b, d, f )-operative images showing decom
pression of the spinal cord following a three-level (T6, 7; T8, 9, and T9, 10) TMED.
dramroo
Fig. 3. Postoperative scar after three-level TMED procedure. Of note, no fusion was
required nor entry into the thoracic cavity for performing a multilevel thoracic discectomy.
posterior muscle dissection. As a result, patient stays are brief with the major
ity of patients being discharge the following morning.
Illustrative Case
A 45-year-old female presented with polyradiculopathy and myelopathy
from three thoracic disc herniations one at T6, 7, another at T8, 9 and another
at T9, 10. The patient underwent a right-sided T6, 7 TMED and a left-sided T8, 9
and T9, 10 TMED. Pre- and postoperative sagittal and axial MRI images show
the extent of disc removal (fig. 2cf) with final decompression of the spinal
cord. The patient went on to make a full recovery with minimal postoperative
wound scar (fig. 3) and returned to work. In a series of 5 patients treated in this
manner, aged 2354 years, operative times averaged 1.8 h per level and blood
loss was approximately 113 ml per level. No cases required conversion to an
open procedure and all patients showed improvement in functional outcome as
measured by visual pain analog, Oswestry scores, and SF-36.
Discussion
The natural history of thoracic disc herniation is not well delineated.

Currently, the only absolute indication for surgery is myelopathy. Surgery for
thoracic disc herniation for controlling radicular thoracic pain is controversial
as is the role of fusion. A number of approaches currently exist for the treatment

dramroo
of this condition including the posterior (laminectomy, transpedicular, trans-
facet pedicle-sparing), anterolateral (transthroacic, thoracoscopic), anterior
(transsternal), and lateral (costotransversectomy, lateral rachiotomy, lateral
extracavitary). Though many of these approaches are effective, they require
significant dissection and retraction of normal anatomic structures, which can
increase patient morbidity. The novel approach described above which uses a
series of muscle dilators, tubular retractor, and endoscopic visualization can
reduce much of the morbidity associated with this procedure and avoids the
need for fusion and entrance into the thoracic cavity. By using a tubular retrac
tor and endoscope, less muscle, rib and transverse process resection is required.
The 30 angle endoscope allows for visualization under the dural sack to reduce
any spinal cord manipulation during the procedure and facilitate removal of the
herniated thoracic disc. Since the majority of the functioning disc is left in
place, no iatrogenic instability is created from extensive disc removal that
would require bone fusion. Although the series presented is small, the technique
is promising as a new minimally invasive approach to thoracic spine pathology
that can lead to reduced operative times, less blood loss, and quicker patient
recoveries.
References
1 Anand N, Regan JJ: Video-assisted thoracoscopic surgery for thoracic disc disease: Classification
and outcome study of 100 consecutive cases with a 2-year minimum follow-up period. Spine
2002;27:871879.
2 Arlet V: Anterior thoracoscopic spine release in deformity surgery: A meta-analysis and review.
Eur Spine J 2000;9(suppl 1):S17S23.
3 Awwad EE, Martin DS, et al: Asymptomatic versus symptomatic herniated thoracic disc: Their
frequency and characteristics as detected by computed tomography after myelography.
4 Birch BD, Desai RD, et al: Surgical approaches to the thoracolumbar spine. Neurosurg Clin N Am
1997;8:471485.
5 Burgos J, Rapariz JM, et al: Anterior endoscopic approach to the thoracolumbar spine. Spine
1998;23:24272431.
6 Cunningham BW, Kotani Y, et al: Video-assisted thoracoscopic surgery versus open thoracotomy
for anterior thoracic spinal fusion. A comparative radiographic, biomechanical, and histologic
analysis in a sheep model. Spine 1998;23:13331340.
7 Dickman CA, Mican CA: Multilevel anterior thoracic discectomies and anterior interbody fusion
using a microsurgical thoracoscopic approach. J Neurosurg 1996;84:104109.
8 Dickman CA, Rosenthal D, et al: Thoracic vertebrectomy and reconstruction using a microsurgi
cal thoracoscopic approach. Neurosurgery 1996;38:279293.
9 Dickman CA, Rosenthal D, et al: Reoperation for herniated thoracic discs. J Neurosurg
1999;91(suppl 2):157162.
10 Faciszewski T, Winter RB, et al: The surgical and medical perioperative complications of anterior
spinal fusion surgery in the thoracic and lumbar spine in adults. Spine 1995;20:15921599.
11 Ferson PF, Landreneau RJ, Dowling RD, Hazelrigg SR, Ritter P, Nunchuck S, Perrino MK,
Bowers CM, Mack MJ, Magee MJ: Comparison of open versus thoracoscopic lung biopsy for
diffuse infiltrative pulmonary disease. J Thorac Cardiovasc Surg 1993;106:194199.
dramroo
12 Fessler RG, Sturgill M: Review: Complications of surgery for thoracic disc disease. Surg Neurol
1998;49:609618.
13 Hazelrigg SR, Landreneau RJ, Boley TM, Priesmeyer M, Schmaltz RA, Nawarawong W,
Johnson JA, Walls JT, Curtis JJ: The effect of muscle-sparing versus standard posterolateral tho
racotomy on pulmonary function, muscle strength, and postoperative pain. J Thorac Cardiovasc
Surg 1991;101:394401.
14 Horowirz BH, Moossy JJ, et al: Thoracic discectomy using video assisted thoracoscopy. Spine
1994;19:10821086.
15 Huang TJ, Hsu RW, et al: Video-assisted thoracoscopic treatment of spinal lesions in the thora
columbar junction. Surg Endosc 1997;11:11891193.
16 Huang TJ, Hsu RW, et al: Complications in thoracoscopic spinal surgery: A study of 90 consecu
tive patients. Surg Endosc 1999;13:346350.
17 Huntington CF, Murrell WD, et al: Comparison of thoracoscopic and open thoracic discectomy in
a live ovine model for anterior spinal fusion. Spine 1998;23:16991702.
18 Karahalios DG, Apostolides PJ, et al: Thoracoscopic spinal surgery. Treatment of thoracic insta
bility. Neurosurg Clin N Am 1997;8:555573.
19 Kuklo TR, Lenke LG: Thoracoscopic spine surgery: Current indications and techniques. Orthop
Nurs 2000;19/6:1522.
20 Landreneau RJ, Hazelrigg SR, Mack MJ, Dowling RD, Burke D, Gavlick J, Perrino MK, Ritter PS,
Bowers CM, DeFino J: Postoperative pain-related morbidity: Video-assisted thoracic surgery
versus thoracotomy. Ann Thorac Surg 1993;56:12851289.
21 Larson SJ, Holst RA, Hemmy DC, Sances A: Lateral extracavitary approach to traumatic lesions
of the thoracic and lumbar spine. J Neurosurg 1976;45:628637.
22 McAfee PC, Regan JR, et al: The incidence of complication in endoscopic anterior thoracolum
bar spinal reconstructive surgery. Spine 1995;20:16241632.
23 Newton PO, Shea KG, et al: Defining the pediatric spinal thoracoscopy learning curve: Sixty-five
consecutive cases. Spine 2000;25:10281035.
24 Newton PO, Wenger DR, et al: Anterior release and fusion in pediatric spinal deformity. A com
parison of early outcome and cost of thoracoscopic and open thoracotomy approaches. Spine
1997;22:13981406.
25 Niemeyer T, Freeman BJ, et al: Anterior thoracoscopic surgery followed by posterior instrumen
tation and fusion in spinal deformity. Eur Spine J 2000;9:499504.
26 Nymberg SM, Crawford AH: Video-assisted thoracoscopic releases of scoliotic anterior spines.
AORN J 1996;63:561562, 565569, 571575; quiz 576, pass.
27 Oskouian RJ Jr, Johnson JP, et al: Thoracoscopic microdiscectomy. Neurosurgery 2002;50/1:
103109.
28 Otani K, Yoshida M, et al: Thoracic disc herniation. Surgical treatment in 23 patients. Spine
1988;13:12621267.
29 Peker S, Akkurt C, et al: Multiple thoracic disc herniations. Acta Neurochir (Wien) 1990;
107/34:167170.
30 Perez-Cruet MJ, Fessler RG, Noel PI: Review: Complications of minimally invasive spinal
surgery. Neurosurgery 2002;51(suppl 2):2636.
31 Perez-Cruet MJ, Foley KT, Isaacs RE, Rice-Wyllie L, Wellington R, Smith MM, Fessler RG:
Microendoscopic lumbar discectomy: Technical note. Neurosurgery 2002;51(suppl 2):129136.
32 Perez-Cruet MJ, Smith MM, Foley KT: Microendoscopic lumbar discectomy; in Perez-Cruet MJ,
Fessler RG (eds): Outpatient Spinal Surgery. St Louis, Quality Medical Publishing, 2002,
pp 171183.
33 Perot PL, Munro DD: Transthoracic removal of midline thoracic disc protrusions causing spinal
cord compression. J Neurosurg 1969;31:458.
34 Ransohoff J, Spencer F, et al: Transthoracic removal of thoracic discs: Case reports and technical
notes. J Neurosurg 1969;31:459461.
35 Regan JJ: Percutaneous endoscopic thoracic discectomy. Neurosurg Clin N Am 1996;7/1:8798.
36 Regan JJ, Guyer RD: Endoscopic techniques in spinal surgery. Clin Orthop 1997;335:122139.
37 Rosenthal D, Dickman CA: Thoracoscopic microsurgical excision of herniated thoracic discs.
J Neurosurg 1998;89/2:224235.

dramroo
38 Rosenthal D, Rosenthal R, et al: Removal of a protruded thoracic disc using microsurgical
endoscopy. A new technique. Spine 1994;19:10871091.
39 Roush TF, Crawford AH, et al: Tension pneumothorax as a complication of video-assisted thoras
copic surgery for anterior correction of idiopathic scoliosis in an adolescent female. Spine
2001;26:448450.
40 Stillerman CB, Chen TC, et al: Experience in the surgical management of 82 symptomatic herni
ated thoracic discs and review of the literature. J Neurosurg 1998;88:623633.
41 Stillerman CB, Chen TC, et al: The transfacet pedicle-sparing approach for thoracic disc removal:
Cadaveric morphometric analysis and preliminary clinical experience. J Neurosurg 1995;83:
971976.
Mick J. Perez-Cruet, MD, MS
Institute for Spine Care, Department of Neurosurgery
Chicago Institute of Neurosurgery and Neuroresearch,
Rush-Presbyterian-St. Lukes Medical Center, 1725 West Harrison Street
Suite 970, Chicago, IL 60612 (USA)
dramroo
Laparoscopic versus Mini-Open
Anterior Lumbar Interbody Fusion
Michael G. Kaiser, Regis W. Haid, Jr., Brian R. Subach,

Gerald E. Rodts, Jr.
Department of Neurosurgery, Emory Clinic, Atlanta, Ga., USA
The anterior approach for a lumbar interbody fusion was originally designed
to treat Potts disease [1]. Since that time, the anterior lumbar interbody fusion
(ALIF) has evolved into an effective and popular alternative in the treatment of
a variety of lumbar degenerative disorders, including degenerative disc disease,
low-grade spondylolisthesis, and posterior pseudoarthrosis. Compared to tradi
tional posterior fusion techniques, the ALIF operation is associated with shorter
operative times, decreased blood loss, less postoperative pain, reduced hospital
stay, and shorter recovery periods [25].
Supporters of the ALIF argue that reconstruction of the anterior column is
biomechanically superior, avoids paraspinal muscle trauma and denervation,
indirectly decompresses the intervertebral foramen, improves sagittal balance,
and allows a more efficient restoration of disc interspace height compared to
posterior fusion techniques [68]. The anterior position of the interbody graft
increases the likelihood of fusion by exposing the graft to fusion-promoting
forces in accordance with Wolffs law [9]. The interbody space also provides an
increased surface area for fusion formation and robust blood supply following
decortication of the vertebral endplates [10].
Early in the development of the ALIF technique, open approaches, such as
the transperitoneal or retroperitoneal approach, were utilized to expose the ante
rior lumbar spine. Although providing adequate visualization, these more exten
sive exposures were associated with increased postoperative morbidity. As the
ALIF technique evolved, emphasis was placed on exposing the spine through
less invasive approaches. These minimally invasive techniques are intended to
decrease postoperative morbidity, reduce hospitalization time, and shorten the
dramroo
recovery period with comparable or superior treatment outcomes compared to
more traditional techniques.
Laparoscopic ALIF
Todays spine surgeon has two such options for approaching an ALIF: the
laparoscopic approach and the mini-open laparotomy. Zucherman et al. [11]
were the first to report the use of the laparoscopic approach for an anterior
interbody fusion. This technique is considered by many as the least invasive
approach to the ventral lumbar spine and in many centers has become the stan
dard technique when performing an ALIF, particularly at the L5/S1 disc inter
space. The safety and efficacy of an anterior laparoscopic fusion have been
reported in numerous reports [2, 3, 1216].
The small incisions required for insertion of the laparoscopic working
channels reduce the extent of abdominal wall muscle dissection and blood loss,
both contributing to a decrease in postoperative pain. The laparoscopic
approach is also associated with less direct manipulation of the abdominal con
tents, resulting in a decreased incidence of postoperative ileus. These charac
teristics are thought to reduce postoperative morbidity and contribute to a
shorter length of hospitalization.
Mini-Open ALIF
More recently the traditional open transperitoneal approach has been

modified into a mini-open technique that incorporates similar principles as the
anterior cervical exposure. Originally reported by Mayer [17] in 1997, the mini-
open technique reduces the degree of postoperative morbidity associated with
the traditional open laparotomy by utilizing a smaller incision combined with a
muscle splitting exposure, dividing the abdominal muscles in the direction of
the fiber orientation. In this initial report, a total of 25 patients underwent an
ALIF across various lumbar levels utilizing the mini-open laparotomy. It was
the authors belief that this technique was associated with negligible surgical
trauma, decreased operative time and blood loss, decreased postoperative mor
bidity, and a shorter recovery time. The authors concluded that this approach
could be employed for a variety of lumbar spinal disorders and fusion tech
niques. The mini-open approach has been described with both a transperitoneal
and retroperitoneal route to the anterior lumbar spine [6, 17, 18]. Recently there
has been evidence to suggest that the laparoscopic approach does not offer a
Kaiser/Haid/Subach/Rodts 278
dramroo
significant advantage over the mini-open approach when performing an ALIF
at the L4/5 interspace [5].
The mini-open approach provides direct access to the disc interspace
and allows easier identification of the anatomical midline. The open technique
allows complete disc removal, resection of any herniated fragments, and
increased endplate exposure, improving the preparation of the disc interspace
for fusion formation compared to the trephine technique used with the laparo
scopic approach. The mini-open laparotomy is less technically demanding,
contributing to a shorter preparation and operative time. The reduced inci
dence of retrograde ejaculation has also been observed with the mini-open
laparotomy [4].
Surgical Technique
Patient Positioning
There are a number of previously published reports describing the details
of both the laparoscopic and mini-open techniques [3, 6, 14, 18]. For both
approaches the assistance of either a general or vascular surgeon, familiar with
laparoscopic techniques when indicated, is advised. This team approach will
optimize the speed and safety of the procedure, providing the patient with
maximal benefit.
Patient positioning is essentially identical for each approach, however
accommodations for the video monitor are required with the laparoscopic
exposure. The patient is positioned supine on a radiolucent operative table
with a pillow or bolster placed under the patients hips to accentuate the
natural lumbar lordosis and under the knees to avoid hyperextension. The
patient is securely strapped to the operative table. This is particularly neces
sary for the laparoscopic approach due to the steep Trendelenburg position
required to allow the abdominal viscera to move rostrally out of the pelvis.
Fluoroscopic imaging is used with both techniques during positioning of the
implants.
Laparoscopic Technique
The patient is prepared and draped under sterile conditions in the usual
fashion. The fluoroscopic equipment is brought into the operative field to con
firm the midline prior to any incisions. For the laparoscopic approach, intra-
abdominal access is obtained through four small incisions in the anterior
abdominal wall (fig. 1). The viewing camera is inserted through a curvilinear
umbilical incision and two lower paramedian incisions provide portals for the
Laparoscopic versus Mini-Open ALIF 279

dramroo
10-mm trocar
(laparoscope)
5- or 10-mm
5- or 10-mm trocar
trocar
L4/L5 or
18-mm trocar L5/S1
(working port)
suprapubic
Fig. 1. Diagram demonstrating the trocar placement for laparoscopic ALIF. The chan
nels for working instruments are provided through two paramedian incisions. The laparo
scope is inserted through an umbilical incision and the instrumentation portal through a
suprapubic incision.
working instruments. The channel for interbody implant insertion is passed

through a midline suprapubic incision and measures approximately 24 cm in
length (fig. 2).
Harvesting of the bone graft is performed at the beginning of the proce
dure. During this period the laparoscopic surgeon obtains access and insufflates
the abdominal cavity. At this point the operative table is placed in a steep
Trendelenburg position to assist in mobilizing the abdominal viscera out of the
pelvis inlet.
Adequate exposure of the disc is critical. It is essential to identify the
appropriate anatomical landmarks and midline (fig. 3). The sacral promontory
is identified by palpation and the midline determined with fluoroscopic imag
ing. The posterior peritoneum is opened sharply. In an attempt to avoid retro
grade ejaculation in male patients the presacral sympathetic plexus is mobilized
through blunt dissection and the use of monopolar cautery avoided at the disc
interspace. More liberal use of the monopolar cautery is allowed with female
patients. The middle sacral artery and vein are then identified, ligated, and
divided.
The insertion of interbody implants is performed only after confirmation
of the midline is repeated. The implants are inserted through a trephine
dramroo
Fig. 2. Final placement of abdominal
trocars following insufflation of the
abdomen. The instrumentation trocar is
positioned toward the top and the laparo
scope toward the bottom of the photograph.
3 4
Fig. 3. Anterior view of the lumbosacral spine prior to retroperitoneal exposure. The
dotted line indicates a possible peritoneal incision to access the retroperitoneum and anterior
disc space.
Fig. 4. Lateral view of the instrumentation port positioned within the disc interspace
for trephination of the disc material and insertion of implant.
technique according to the guidelines specific for the implant chosen (fig. 4).
Following radiographic confirmation of implant position, the posterior
peritoneum is closed using clip ligation. The abdominal incisions are closed
with interrupted absorbable sutures and Steri-Strips.

dramroo
Fig. 5. Typically a transverse incision is used for single level exposure for cosmesis, how
ever a longitudinal incision can be used for obese patients, history of prior abdominal surgery,
or if more than one level is addressed. The incision is usually 46 cm in length.
Mini-Open Technique
Excluding the equipment required for a laparoscopic approach, the opera
tive setup is identical. The level of incision is determined with fluoroscopic
imaging, with a bias toward the caudal border of the level of arthrodesis. This
bias will provide a more tangent approach into the disc space and avoid exces
sive reaming into the inferior vertebral body. A longitudinal or transverse
incision approximately 46 cm in length is made in the suprapubic region.
A longitudinal incision is reserved for two-level ALIF procedures or for obese
patients (fig. 5). After the skin incision is made, monopolar cautery is used to
dissect down to the rectus abdominus muscle that is split in a longitudinal fash
ion parallel to the muscle fiber plane. The posterior rectus sheath and transver
salis fascia are divided to expose the underlying peritoneum. The peritoneum is
incised and the abdominal contents packed superiorly held in position with a
table-mounted retractor (fig. 6). Although the anterior lumbar spine can be
approached through either a transperitoneal or retroperitoneal exposure, our
preference has been the transperitoneal route since a more direct anterior
trajectory is obtained.
The posterior peritoneum overlying the disc interspace is identified and
sharply divided. Exposure of the L5/S1 interspace is usually easier due to the
more rostral bifurcation of the great vessels. In order to expose L4/5 interspace
the left iliac artery and vein are mobilized. It is imperative that the iliolumbar vein
be identified and ligated. Transection of this vein without adequate control can
dramroo
Fig. 6. A table-mounted retractor system is placed to optimize exposure and retract the
abdominal contents rostrally.
Fig. 7. Once the retractor blades are positioned and the posterior peritoneum opened
the anterior surface of the disc is easily identified.
lead to catastrophic blood loss. Mobilizing the presacral sympathetic plexus in

male patients is performed with blunt dissection, avoiding the use of monopolar
cautery. In females monopolar cautery may be used more liberally to expose the
anterior surface of the vertebral bodies and disc space (fig. 7). Midline identifi
cation is made through direct visualization and fluoroscopic imaging.

dramroo
Fig. 8. Intraoperative photograph
demonstrated the generous discectomy, with
preservation of the lateral annulus, required
for adequate endplate exposure prior to
implant insertion.
Fig. 9. Intraoperative view of the double-barrel channel in position for implant insertion.
A complete removal of the disc is performed and the entire endplate prepared
for graft insertion (fig. 8). The interbody implant is then inserted according to
manufacturers guidelines along with additional autologous bone (fig. 9). Once the
appropriate implant position is verified with intraoperative fluoroscopic images,
the posterior peritoneum is primarily closed with a running absorbable suture.
The anterior peritoneum, transversalis fascia, and rectus sheath are closed with
interrupted absorbable sutures and the skin closed with staples.
dramroo
Postoperative Care
The postoperative course for both the laparoscopic and mini-open groups
is essentially identical. Mobilization occurs early, typically on the first postop
erative day, and their diet advanced with the initiation of bowel sounds. Patients
are generally discharged once they are able to tolerate an oral diet, able to
ambulate, and voiding without difficulty. Occasionally this may be as early as
the 2nd day following surgery. Radiographic images, to determine implant
position, are obtained prior to discharge and during scheduled follow-up visits,
at approximately 6 weeks, 3, 6, 9, 12 and 24 months postoperatively.
Laparoscopic versus Mini-Open ALIF
There have been few studies that directly compared the mini-open laparot
omy to the laparoscopic ALIF. Regan et al. [19] compared their experience of 65
laparoscopic fusions to a large number of both anterior and posterior open
interbody fusions. The mean operative time was significantly shorter for the mini-
open ALIF compared to the laparoscopic approach, 149 compared to 207 min.
The average blood loss during the mini-open procedure was 224 ml (range
202,000 ml) and for the laparoscopic approach 176 ml (range 102,200 ml). The
trend for decreased blood loss with the laparoscopic approach did not prove sta
tistically significant. The mean length of hospitalization was essentially the same
for both groups, approximately 3.9 days for the laparoscopic group and 4.0 days
following the mini-open approach. There were several technical complications
associated with the laparoscopic approach that occurred early in the surgeons
operative experience. Nine attempts at a laparoscopic fusion were aborted and
converted to an open approach, 3 secondary to laceration of the iliac vein.
In 2 cases a herniated disc was noted to impinge upon a nerve root requiring a
second operation for removal of the herniated fragment.
More recently Zdeblick and David [5] presented their experience with both
the mini-open and laparoscopic approaches. Over a 3-year period data was
prospectively collected on 50 patients who underwent an ALIF involving the
L4/5 disc interspace, with an equal number of laparoscopic and mini-open
laparotomy procedures. The authors found no statistical difference in operating
time, intraoperative blood loss, or length of hospital stay between the two
groups. When two-level procedures were considered separately, a significant
increase in the mean operative time was noted with the laparoscopic approach,
185 versus 160 min. The laparoscopic group demonstrated a significantly
higher complication rate, 20 versus 4%. Based on these observations, the
authors concluded that the laparoscopic approach offered no advantage over the
mini-open laparotomy when performing an ALIF involving the L4/5 level.

dramroo
Table 1. Comparison of operative variables for all
ALIF procedures
Laparoscopic Mini-open
Preparation time, min 50.8 26.2

Surgical time, min 185.0 171.9
Operative blood loss, ml 188.2 231.3
Hospitalization, days 2.9 3.7
Table 2. Comparison of operative variables for L5/S1

ALIF
Laparoscopic Mini-open
Preparation time, min 50.2 24.8

Surgical time, min 173.6 145.1
Operative blood loss, ml 178.3 160.1
Hospitalization, days 2.8 3.5
Complications associated with an ALIF approach can be categorized as

either major or minor and divided into visceral, vascular, neural, and fusion-
related complications. Rajaraman et al. [20] recently reported the complications
recorded in their series of 60 patients undergoing open anterior lumbar surgery.
A total of 24 complications were noted in 23 patients (38.3%), including
sympathetic dysfunction in 6, vascular injury in 4, somatic nerve injury in 3,
sexual dysfunction in 3, prolonged ileus in 3, wound incompetence in 2, and
bowel injury, deep venous thrombosis, and acute pancreatitis in 1 patient. The
majority of complications were only transient in nature.
In a study focusing on the complications associated with endoscopic spinal
surgery, McAfee et al. [16] included 22 patients that underwent a laparoscopic
ALIF at either L4/5 or L5/S1. One approach-related complication was noted, a
tear in the left common iliac vein that required conversion to an open proce
dure. The authors concluded that a laparoscopic fusion was a safe and effective
procedure compared to traditional open techniques.
Over the past 5 years our group has gained extensive experience with
both the laparoscopic and mini-open ALIF procedures [3, 6, 14, 18].
Between 1996 and 2001 a total of 98 ALIF procedures were performed at our
institution, equally distributed between a laparoscopic and mini-open
approach. Recently we performed a retrospective review of our ALIF expe
rience, comparing data regarding operative parameters, length of hospital
dramroo
Table 3. Intraoperative complications Table 4. Postoperative complications
Laparoscopic Laparoscopic
Bladder perforation UTI
Iliac vein laceration New onset radiculopathy
Conversion to open procedure (n 2) Mini-open
Mini-open Transient ileus (n 5)
Iliac vein laceration (n 2) Retroperitoneal hematoma
UTI
Wound infection
Worsened radiculopathy
UTI Urinary tract infection.
stay, and approach-associated complications (tables 14). Comparisons

were made for the entire series as well as for patients undergoing an L5/S1
ALIF [21].
One of the more striking differences observed between the two
approaches was the time required to organize the operating room. This inter
val was estimated from the time of anesthesia induction to the point of
surgical incision. The operative preparation time was significantly shorter
for the mini-open laparotomy. For the entire patient series as well as patients
undergoing an L5/S1 ALIF the operative organization time was reduced by
approximately 50% during a mini-open approach. Although this time inter
val is an arbitrary measurement, the increased preparation time is directly
related to the technically demanding equipment required for the laparoscopic
approach.
The procedure time was also reduced following a mini-open laparotomy,
although the decrease only proved statistically significant for patients undergo
ing an L5/S1 ALIF. The lack of significance observed for the entire series is
likely the result of a selection bias for two-level and L4/5 interspace procedures
performed with the mini-open approach. As with the operative preparation
interval, we concluded that the increased procedure time was related to the
equipment required for the laparoscopic approach.
No significant differences, for the entire series or the L5/S1 cohort, were
observed regarding the degree of intraoperative blood loss or intraoperative
complications. Vascular and visceral injuries occurred following both
approaches. A clear advantage of the mini-open approach is the immediate and
direct access provided in the event of such an injury.
An increased incidence of immediate postoperative complications was noted
with the mini-open laparotomy. The majority of these complications consisted
of a transient ileus that resolved by the 5th postoperative day. A retroperitoneal

dramroo
hematoma was the only additional complication directly related to the mini-
open approach and required observation only.
Retrograde ejaculation resulting from injury to the superior hypogastric
plexus has a reported incidence of up to 22% in men undergoing an open
anterior lumbar fusion [4, 5, 20, 22] and up to 20% following a laparoscopic
fusion [35, 12]. Up to one third to half of these patients will return to nor
mal function within 2 years [22]. In our entire series of male patients a
significantly higher incidence of retrograde ejaculation occurred following
the laparoscopic ALIF, 45.4 versus 6.0%. The percent of men suffering an
episode of retrograde ejaculation following a laparoscopic L5/S1 ALIF was
unchanged, however following a mini-open laparotomy the incidence
decreased to 0%. Although an exact explanation for the increased incidence
observed in the laparoscopic group is lacking, we speculate that sweeping
the neural plexus off the midline using the laparoscope instead of a cotton
sponge is more difficult and damaging. It is unclear why our observed inci
dence was so high; however, as with previous reports the majority of these
cases were transient in nature.
Conclusions
Based on the theoretical considerations described, previous published

series, and our own operative experience, we have developed a bias toward the
mini-open laparotomy when performing an ALIF at either the L4/5 or L5/S1
level. It is our contention that the mini-open ALIF is technically less demand
ing, allows for a shorter operative time, provides the access necessary for com
plete disc removal, allows for resection of a herniated fragment, and provides
direct visualization of the disc interspace for optimal interbody graft insertion.
In addition the incidence of retrograde ejaculation is significantly reduced with
the mini-open laparotomy. Both the mini-open and laparoscopic techniques,
however, are effective and safe and ultimately surgeon preference will dictate
the choice of approach utilized when performing an ALIF.
References
1 Ito H, Tsuchiya J, Asami G: A new radical operation for Potts disease. J Bone Joint Surg Br
1934;16:499515.
2 Dickman C, Sonntag V, Russell J: The laparoscopic approach for instrumentation and fusion of the
lumbar spine. BNI Q 1997;13:2636.
3 McLaughlin M, Zhang J, Subach B, Haid R, Rodts G: Laparoscopic anterior lumbar interbody
fusion: Technical note. Neurosurg Focus 1999;7:16.
dramroo
4 Reagan J, Yuan H, McAfee P: Laparoscopic fusion of the lumbar spine: Minimally invasive spine
surgery. A prospective multicenter study evaluating open and laparoscopic lumbar fusion. Spine
1999;24:402411.
5 Zdeblick T, David S: A prospective comparison of surgical approach for anterior L4L5 fusion:
Laparoscopic versus mini anterior lumbar interbody fusion. Spine 2000;25:26822687.
6 McLaughlin M, Haid R, Rodts G, Miller J: Current role of anterior lumbar interbody fusion in
lumbar spine disorders. Semin Neurosurg 2000;11:221229.
7 Voor MJ, Mehta S, Wang M, Zhang YM, Mahan J, Johnson JR: Biomechanical evaluation of
posterior and anterior lumbar interbody fusion techniques. J Spinal Disord 1998;11:328334.
8 Greenough CG, Peterson MD, Hadlow S, Fraser RD: Instrumented posterolateral lumbar fusion.
Results and comparison with anterior interbody fusion. Spine 1998;23:479486.

9 Egger E, Gottsauner-Wolf F, Palmer J, et al: Effects of axial dynamization on bone healing.
J Trauma 1993;34:185192.
10 Lin P: Posterior lumbar interbody fusion; in Cauthen J (ed): Lumbar Spine Surgery. Indications,
Techniques, Failures, and Alternatives, ed 2. Baltimore, Williams & Wilkins, 1988, pp 228247.
11 Zucherman JF, Zdeblick TA, Bailey SA, Mahvi D, Hsu KY, Kohrs D: Instrumented laparoscopic
spinal fusion. Preliminary results. Spine 1995;20:20292035.
12 Zdeblick TA: Laparoscopic spinal fusion. Orthop Clin North Am 1998;29:635645.
13 Mahvi DM, Zdeblick TA: A prospective study of laparoscopic spinal fusion. Technique and
operative complications. Ann Surg 1996;224:8590.
14 McLaughlin M, Comey C, Haid R: Laparoscopic anterior lumbar interbody fusion. Cont Neurosurg
1998;20:19.
15 Bhatnagar MK, Mathur SK, Mess CF: Laparoscopic spinal fusion. Md Med J 1999;48:161164.
16 McAfee PC, Regan JR, Zdeblick T, Zuckerman J, Picetti GD 3rd, Heim S, Geis WP, Fedder IL: The
incidence of complications in endoscopic anterior thoracolumbar spinal reconstructive surgery. A
prospective multicenter study comprising the first 100 consecutive cases. Spine 1995;20:16241632.
17 Mayer HM: A new microsurgical technique for minimally invasive anterior lumbar interbody
fusion. Spine 1997;22:691700.
18 McLaughlin M, Haid R, Rodts G, Miller J: Open anterior lumbar interbody fusion. Tech Neurosurg
2001;7:140151.
19 Regan J, Zdeblick T, Zucherman J: Comparison of open versus laparoscopic fusion of the lumbar
spine using the BAK threaded fusion cage. 10th Annual Meeting, North American Spine Society,
Washington, 1995, p 131.
20 Rajaraman V, Vingan R, Roth P, Heary RF, Conklin L, Jacobs GB: Visceral and vascular compli
cations resulting from anterior lumbar interbody fusion. J Neurosurg 1999;91:6064.
21 Kaiser M, Haid R, Subach B, Miller J, Smith C, Rodts G: Comparison of the mini-open versus
laparoscopic approach for anterior lumbar interbody fusion: A retrospective review. Neurosurgery,
in press.
22 Flynn JC, Price CT: Sexual complications of anterior fusion of the lumbar spine. Spine 1984;9:
489492.
Department of Neurosurgery, The Emory Clinic,
1365B Clifton Road, Atlanta, GA 30322 (USA)

dramroo
Socio-Economic Evolution
Epidemiology and Variations in
Care of Spine Disease
James P. Burke, Peter C. Gerszten

Department of Neurological Surgery, University of Pittsburgh Medical Center,
Pittsburgh, Pa., USA
Spine disease, with associated neck, back, and radicular pain, is common
and costly. In the current American health care environment, many types of
practitioners (primary care physicians, spine surgeons, physical therapists,
chiropractors, and others) are involved in managing patients with spine disease,
often with very different approaches. The varied nature and economic costs of
spine disease are driving a growing interest in research. The coming years may
reveal the fundamental aspects of this problem, as well as standardized treat
ment regimens, in more detail.
Epidemiology of Spine Disease
Degenerative spine disease is extremely common in our society. It is well

accepted that the vast majority of the population will develop degenerative spinal
abnormalities; some form of spondylosis is present radiographically in more than
80% of males and females older than 55 years [1]. In the more mobile cervical
and lumbar spinal segments, the development of bony and ligamentous hyper
trophy, along with facet arthropathy, is expected. While spinal degeneration may
develop at an earlier age in some patients, these asymptomatic degenerative
changes are almost universal in the elderly, active individual. The patient with a
congenitally narrow spinal canal may present at a young age with symptomatic
spinal stenosis [2].
Age-related spinal disease typically presents in the 6th through 8th
decades. Patients may complain of nonspecific neck pain, low back pain (LBP),
or both. Radicular complaints in the upper extremities or symptoms suggestive
dramroo
of neurogenic claudication are common. Typical age-related changes may be
expected on plain radiographs of the spine. These degenerative changes include
some minimal interspace collapse, moderate osteophyte formation with associ
ated foraminal encroachment, small ventral traction spurs, minimal vertebral
body compression or collapse in the osteoporotic patient, and moderate spinal
canal stenosis, without evidence of fracture, instability, or subluxation [2].
The true incidence of symptomatic cervical spondylosis is difficult to
assess because minor cervical discomfort or neck stiffness often goes unre
ported. Teresi et al. [3] reviewed cases in which cervical MRIs had been
obtained from patients with no subjective complaint or objective physical find
ing of degenerative spinal disease. They noted disc space narrowing in 67% of
patients older than 64 years, compared with a 24% occurrence rate for those
aged 4564 years. Patients with degenerative spinal disease in the cervical
region may present with a myriad of signs and symptoms. The mean age of
onset of symptoms ranges from 60 to 69 years, with men affected slightly more
than women [4]. The natural history of myelopathy has been well described,
with progression usually during a prolonged period. Spontaneous improvement
is uncommon [4, 5]. The natural history of radiculopathy and discogenic neck
pain is less well defined. It is believed that many patients with radiculopathy
improve with time; however, some patients will fail to improve, particularly
those with a soft disc herniation [6]. Up to 80% of patients with neck pain will
improve with time.
While the thoracic spine is much less mobile than either the cervical or the
lumbar spine and degenerative disease is less common, the thoracic spine is not
exempt from the effects of aging and degeneration. Kyphosis occurs commonly
among elderly women. Although symptomatic degeneration in the thoracic
spine is uncommon, relatively less compromise of the thoracic spine may result
in neurological deficits for two reasons. First, the spinal cord occupies a rela
tively larger percentage of the canal throughout the thoracic spine, and com
pression may occur from relatively smaller osteophytes. Second, the vascular
supply to the thoracic spinal cord comes primarily from segmental or radicular
arteries. This supply places the thoracic cord at an even greater risk of ischemia
from compression.
Symptomatic herniation of thoracic discs is uncommon. While early studies
reported that surgery for thoracic disc herniation accounted for only 35 of every
1,000 disc operations [7], the true incidence of symptomatic thoracic disc herni
ation probably is higher. MRI of the spine has allowed diagnosis of the problem
in many patients before neurological signs develop. Thoracic disc herniations
occur most commonly in adults aged 3060 years, although these lesions may
occur in children [8], and are equally common in men and women.
Approximately two thirds of thoracic disc herniations occur below the level of
Epidemiology and Variations in Care of Spine Disease 291

dramroo
T8, presumably because of the greater forces acting on the lower thoracic spine.
Midline herniations are several times more common than those that are laterally
placed [7, 9].
In general, degenerative disease of the lumbar spine is clinically stable,
with relatively few minor symptoms and signs. The progression of symptoms is
often slow, and the patient usually describes a history of steady worsening of
symptoms and signs over years or even decades. Given the varied descriptions
and definitions of lumbar spondylosis, it is not surprising that the prevalence
varies widely depending upon the reporting authority. In one series of 850
myelograms, researchers noted a 6% occurrence of stenosis of the thecal sac
[10]; however, there was no mention as to why the myelograms were performed
nor how many of these patients were symptomatic. On the other hand, Roberson
et al. [11] reported a 1.7% occurrence of lumbar stenosis in a series of 2,000
lumbar myelograms. Because a certain percentage of patients have lumbar
stenosis secondary to degenerative disease, the absolute prevalence of stenosis
varies according to the age and sex of the population studied. The prevalence in
the general population remains unknown.
The patients age at the onset of symptoms also varies with the criteria used
to define lumbar spondylosis. In a study by Paine [12] that included conditions
with or without disc herniation, the mean age of onset was 30 years in patients
with preexisting stenosis. The age of onset in patients with congenital or devel
opmental stenosis (such as achondroplasia) is similar. Most studies evaluating
acquired lumbar stenosis place the average age of onset of symptoms in the late
5th to early 7th decades [13]. Symptomatic degenerative spondylolisthesis is
infrequent in patients younger than 50 years of age [14].
Currently, there is some controversy in the literature regarding the sexual
preponderance of lumbar spondylosis. Both the developmental and acquired
forms of lumbar stenosis have shown a male preponderance in most early
studies [12, 15]. However, in more recent studies noting the increased recogni
tion of this disease process, the ratio appears to favor women over men by as
much as 5 to 1 [16, 17]. Symptomatic degenerative spondylolisthesis has
been reported to occur 46 times as frequently in women as in men, perhaps
because of hormonal factors leading to ligamentous laxity of the motion
segment [18].
In contrast to degenerative spine disease, there are approximately 11,000
new cases of acute spinal cord injuries per year or 4 per 100,000 persons. The
estimated prevalence of persons living with spinal cord injuries is between
300,000 and 500,000 cases. An increased prevalence during the past decade has
been mainly attributed to enhanced longevity of spine-injured patients.
Increased survivorship, as well as improvements in neurological outcome, has
been attributed to enhanced medical, surgical, and prehospital care. Spinal cord
Burke/Gerszten 292
dramroo
Table 1. Incidence of several neurosurgical disease states
Diagnosis Incidence per 100,000 persons
LBP 30,000
Carpal tunnel syndrome 2,800
Traumatic brain injury 150
Occlusive cerebral vascular disease 96
Intracerebral/intracranial hemorrhage 30
Malignant neoplasm of the brain 12
Subarachnoid hemorrhage (nontraumatic) 8
Traumatic spinal cord injury 4
injury is most common in young adult men. It is estimated that 63% of new
traumatic injuries occur in individuals between the ages of 16 and 30 years,
with a 4:1 male:female ratio [19].
Epidemiology of LBP
LBP is a very common health problem in western industrialized countries.

Up to 80% of Americans report having LBP at some point in their lives, most
often between the ages of 30 and 50 years. To put this in perspective, table 1 lists
the incidence of LBP compared to a variety of other neurosurgical disease states.
In addition, clinically significant sciatica is seen in 5% of the US population.
Recurrent episodes of LBP occur very frequently, and a considerable number of
people have permanent discomfort from LBP. Chronic LBP is present in 37%
of the population in western industrialized countries [2022]. In the USA, back
pain is the most common cause of activity limitation in people younger than 45
years [23], the second most frequent reason for visits to the physician, the fifth-
ranking cause of admission to the hospital, and the third most common cause of
surgical procedures [21]. About 2% of the US workforce are compensated for
back injuries each year [21]. In 1988, the annual combined cost of back pain-
related medical care and disability compensation in the United States was
estimated to be between USD 26 and USD 56 billion [24].
LBP affects men and women equally, with onset most often between the
ages of 30 and 50 years. It is the most expensive cause of work-related disabil
ity in terms of workers compensation and medical expenses. Risk factors for
chronic LBP include heavy lifting and twisting, bodily vibration, obesity, and
poor conditioning, although LBP is common even in people without these risk
factors [21, 22].

dramroo
Recovery from nonspecific LBP is generally rapid. One study reported
that 90% of patients seen within 3 days of LBP onset recovered within 20 weeks
[25]. However, in cross-sectional studies, which oversample patients with
multiple visits, the prognosis is less favorable. These studies may best reflect
the experience of primary care physicians, and suggest that one third of patients
are substantially improved at 1 week, with two thirds improved at 7 weeks [26,
27]. Recurrences are common, affecting 40% of patients within 6 months. Most
recurrences are not disabling, but the emerging picture is that of a chronic
problem with intermittent exacerbations, analogous to asthma, rather than an
acute disease that can be cured [22].
LBP is second only to upper respiratory problems as a symptom-related
reason for visits to a physician [20, 21]. Unfortunately, there are wide variations
in care, a fact that suggests there is uncertainty about the optimal approach to
management. In addition, there is evidence of excessive imaging and surgery
for LBP in the United States, and many experts believe the problem has been
overmedicalized [22].
The association between clinical symptoms and radiographic imaging
shows little correlation. Most diagnostic evaluations focus on exclusion of
herniated discs, stenosis, infection, neoplasms, and trauma. However, up to
85% of patients with isolated LBP cannot be given a precise pathoanatomical
diagnosis, with nonspecific terms, such as strain, sprain, or degenerative
process, commonly used [20]. Strain and sprain have never been anatomically
or histologically characterized, and patients given these diagnoses might accu
rately be said to have idiopathic LBP [22].
Patients with chronic LBP (i.e., pain duration more than 3 months) use
health services more often than most other patient groups [20]. Spontaneous
recovery in these patients is slow and unpredictable, and the return-to-work rate
after 2 years absence from work due to LBP has been shown to be exceedingly
low. At any given time, about 1% of the work force is chronically disabled
because of back problems [28].
In contrast, the natural history of herniated discs is more favorable.
Improvement is the norm, although it is often slower than improvement in LBP
alone. Only about 10% of patients have sufficient pain after 6 weeks that
surgical intervention is considered. Sequential MRI studies reveal that the
herniated portion of the disk tends to regress with time, with partial or complete
resolution in two thirds of cases after 6 months [29]. Patients with herniated
discs who undergo surgery do not return to work earlier than those who receive
nonsurgical therapy, although they have better symptomatic and functional
outcomes [30].
A paradox exists in that the American economy has become increasingly
postindustrial, with less heavy labor and more automation and more robotics,
Burke/Gerszten 294
dramroo
while medical advances in spine treatment with diagnostic imaging and new
forms of surgical and nonsurgical therapy have developed. At the same time,
however, work disability caused by back pain has steadily risen! The positive
aspect is that most back pain patients will substantially and rapidly recover,
even when their pain is severe. This prognosis holds true regardless of treatment
method or even without treatment. Only a minority of patients with back pain
will miss work because of it. Most patients who do leave work return within
6 weeks, and only a small percentage never return to their jobs.
Variations in Treatment for Spinal Disorders
Patients with neck and back pain seek care from general practitioners,
chiropractors, orthopedists, neurosurgeons, rheumatologists, and others. There
is a wide variation in how doctors care for patients with neck and back pain,
with evidence of excessive imaging and surgery for the problem. In most cases
of LBP, patients recover within a few weeks of the onset of symptoms. The fact
that LBP often resolves spontaneously may partially explain the proliferation of
unproven treatments that may seem to be effective. When patient descriptions
are standardized, physician recommendations for neck and back pain evalua
tions vary enormously. Rheumatologists are twice as likely to order blood work
to rule-out arthritic conditions. Neurosurgeons are twice as likely to order
imaging studies. Neurologists are 3 times more likely to order EMGs [28].
Surgical procedures for herniated discs and spinal stenosis accounted for
83% of the more than 188,000 spine surgeries done in Medicare patients in
19961997. There were approximately 39,000 discectomies, 90,000 lumbar
decompressions, and 27,000 cervical spine procedures. The remaining 32,000
procedures were for other spinal conditions. Overall, spine surgery rates
increased by 57% over the 10-year period from 1988 to 1996, from 2.1 to 3.4 per
1,000 Medicare enrollees (fig. 1) [31].
There remains significant variation within the United States in the use of
surgery for many spine-related problems. Rates of spine surgery vary more than
any other common inpatient procedure. In 19961997, rates of spine surgery
varied by a factor of 6, from 1.4 per 1,000 Medicare enrollees in the Bronx,
N.Y. hospital referral region to 8.6 per 1,000 Medicare residents in Bend, Oreg.
Cervical spine procedures accounted for 14% of the spine surgery performed in
the Medicare population in 19961997; rates of cervical spine surgery varied
by a factor of more than 10, ranging from 0.16 to 1.72 per 1,000 Medicare
enrollees in the different hospital referral regions [31].
The use of spinal fusion displays a wide variation among geographic
areas, as well as varying from 0.3 to 3.0 per 1,000 Medicare enrollees (fig. 2).

dramroo
Spine surgery per 1,000 Medicare enrollees
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997
Year
Fig. 1. Increase in rates of spine surgery among Medicare enrollees (19881997).

Overall surgery rates increased by 57% between 1988 and 1997 [from 31, p. 29].
3.2
2.7
Spinal fusion procedures per
1,000 Medicare enrollees
2.2
1.7
1.2
0.7
0.2
Fig. 2. Rates of spinal fusion varied by a factor of almost 10, from 0.3 to 3.0 per 1,000
Medicare enrollees, after adjustment for differences in population age, sex and race
(19961997). Each point represents 1 of 306 hospital referral regions in the United States
[from 31, p. 38].
The proportion of patients undergoing spine surgery who received a spinal fusion
increased from 23% in 1993 to 29% in 1997. During the same period, the
proportion of patients undergoing fusion who received hardware fixation devices
rose from 50 to 60%. Among Medicare patients undergoing laminectomy for
Burke/Gerszten 296
dramroo
Percent of decompression
procedures for lumbar
stenosis that included fusion
by hospital referral region (19961997)
40 or more (12)
30 to 40 (45)
20 to 30 (114)
10 to 20 (97)
Less than 10 (10)
Not populated
Suppressed for confidentiality
Fig. 3. Use of fusion with surgery for lumbar spinal stenosis (19961997). Fusion was
used in at least 40% of lumbar decompression procedures in 12 hospital referral regions.
In 10 regions, fusion was used in less than 10% of operations [from 31, p. 41].
lumbar spinal stenosis, the use of fusion varied from 4% of the operations to 56%.
Figure 3 shows the variation across the United States in the use of fusion with
surgery for lumbar spinal stenosis [31].
It is unlikely that the large degree of regional variation in the use of spine
surgery reflects regional differences in the incidence of disease. Instead,
regional variation in surgery reflects differences in physician practice style
physicians in some regions of the United States are simply more inclined to rec
ommend surgery in patients with surgically treatable conditions of the spine.
The likelihood that patients will undergo particular surgical procedures of the
spine is remarkably variable. This is apparent on a local level, as well as with
surgical signatures reflecting the practice patterns of individual physicians
and the local medical culture. Neighboring regions with similar demographics

dramroo
1.6
1.41
1.39
1.27
1.4
1.18
1.16
Ratio to US average (1996-1997)
1.09
1.08
1.2
1.04
0.98
0.98
0.94
0.93
0.92
0.88
1.0
0.86
0.84
0.83
0.72
0.72
0.8
0.62
0.61
0.53
0.51
0.6
0.46
0.4
0.2
0.0
Fresno Modesto Alameda Salinas San Francisco San Jose San Mateo Stockton
County County
Lumbar discectomy Decompression for lumbar stenosis Cervical spine surgery
Fig. 4. The surgical signature of spine surgery in 8 California hospital referral regions
(19961997). Patterns of spine surgery the surgical signatures of hospital referral regions
varied in idiosyncratic ways. This graph compares rates of three kinds of spine surgery in 8
California hospital referral regions to the national average [from 31, p. 44].
and about the same per capita numbers of spine surgeons can have very differ
ent signatures for spine surgery. Figure 4 shows the variation in spine surgery
in eight different hospital referral regions. The rate of lumbar discectomy was
41% higher than the national average in the Stockton, Calif. hospital referral
region, but the rate of decompression for lumbar stenosis was 7% lower than
the average. By contrast, the rate of lumbar discectomy in Fresno, Calif. was
39% below the average, but the rate of cervical spine surgery was 9% higher
than the national average. In San Jose, Calif. the rate of decompression was
54% below the average, but the rate of lumbar discectomy was only 2% below
the average [31].
Among the surgeons who performed spine surgery in Medicare enrollees in
1996, 3,011 were orthopedic surgeons and 2,934 were neurosurgeons. Overall,
neurosurgeons performed 64% of all spine surgery among Medicare enrollees,
compared to 36% by orthopedists. The proportion of spine surgery performed by
either neurosurgeons varied markedly among hospital referral regions, from 19%
in Akron, Ohio to 99% in Rapid City, S.D. The relative contributions of orthope
dists and neurosurgeons also varied widely according to the kind of procedure.
While neurosurgeons performed 85% of surgical procedures on the cervical spine,
they performed only 59% of decompressions for lumbar stenosis. Neurosurgeons
and orthopedic surgeons were quite different in their use of fusion for some types
of spine surgery. While both performed noninstrumented fusions in about one
third of cervical procedures, orthopedic surgeons were much more likely to
Burke/Gerszten 298
dramroo
Proportion of surgery using no fusion, uninstrumented
%
100 9% 3%
2% 8%
6% 21% 5%
27%
fusion, and fusion with hardware
80 43%
16%
60
32%
95%
85% 87%
40 32%
64%
20 42%
25%
0
Orthopedic Neuro Orthopedic Neuro Orthopedic Neuro
surgeons surgeons surgeons surgeons surgeons surgeons
Lumbar discectomy Lumbar decompression Cervical spine surgery
No fusion Uninstrumented fusion Fusion with hardware
Fig. 5. Use of fusion (uninstrumented and with hardware) by orthopedists and neuro
surgeons in spine surgery, by indication (1996) [from 31, p. 49].
perform a fusion during lumbar procedures and were much more likely to perform
an instrumented fusion during both cervical and lumbar procedures (fig. 5) [31].
The Cost of Spine Disease
It has been said that next to the common cold, no condition afflicts the
American population with greater incidence and prevalence than LBP. Back pain
affects about 31 million people annually, and over 10 million people are disabled
because of back pain, with upwards of 250 million workdays lost per year. Over
USD 50 billion is expended on the management of back pain each year, with
approximately USD 11 billion of that in the workers compensation system. Two
thirds of Americans suffer an incapacitating episode of back pain at least once
in their lives, one third are suffering at any one time, and over one tenth are
seeking medical care [19]. An estimated 25% of American workers will experi
ence some back pain each year, with 5060% experiencing some disabling pain
during their working career [32]. While the annual rate for compensable back
pain is approximately 2%, this varies with occupation: less than 1% for admin
istrative and clerical personnel, and 115% for industrial workers [19].
Of the total money spent annually for the management of back pain, the
distribution of total direct costs is 45% for permanent disability payments,

dramroo
22% for temporary disability payments, and 33% for medical expenses. The
medical expenses are distributed with 33% to physicians, 33% for hospital
bills, 7% for medication, 5% for appliances, 9% for physical therapy, and 12%
for diagnostic tests [32]. Surgical intervention represents only a small fraction
of the total socioeconomic impact of this disease process, with less than 0.5%
of patients with back pain undergoing surgery [33].
Spinal cord injuries also represent a significant cost to our society. There are
very few accurate data on the total cost of spinal cord injuries. The estimated
annual cost to support and treat all patients with a spinal cord injury is over USD
4 billion [34]. As expected, the initial hospitalization costs (including acute care
and rehabilitation) and the annual follow-up expenses vary tremendously depend
ing on the level of the injury. In one study, researchers found that costs of acute
and rehabilitation care vary from USD 67,950 in the low-level paraplegic patient
to USD 426,592 in the respirator-dependent, high-quadriplegic patient. Similarly,
these researchers noted that annual follow-up expenses for the low-level paraplegic
patient were USD 10,109 compared with USD 141,238 for the respirator-
dependent, high quadriplegic patient [32]. Regarding economic opportunities lost,
foregone lifetime earnings vary from USD 141,253 in the patient who is incom
pletely paraplegic to USD 308,054 in the completely quadriplegic patient [35].
Swift diagnosis and treatment that result in a rapid return to a premorbid level of
activity would greatly reduce costs to the patient and to society.
References
1 Lawrence JS, Bremmer JN, Bier F: Osteoarthrosis: Prevalence in the population and relationship
between symptoms and x-ray changes. Ann Rheum Dis 1966;25:124.
2 Hadley MN, Robinson RD: Clinical assessment of degenerative disorders of the spine; in Tindall GT,
Cooper PR, Barrow DL (eds): The Practice of Neurosurgery. Philadelphia, Williams & Wilkins,
1996, pp 23672375.
3 Teresi LM, Lufkin RB, Reicher MA, Moffit BJ, Vinuela FV, Wilson GM, Bentson JR, Hanafee WN:
Asymptomatic degenerative disk disease and spondylosis of the cervical spine: MR imaging.
Radiology 1987;164:8388.
4 Clarke E, Robinson P: Cervical myelopathy: A complication of cervical spondylosis. Brain 1956;
79:483510.
5 Wilberger JE Jr, Chedid MK: Acute cervical spondylotic myelopathy. Neurosurgery 1988;22:
145146.
6 Zeidman SM, Ducker TB: Cervical disc disease. Neurosurg Q 1992;2:116163.
7 Love JG, Schorn VG: Thoracic disc protrusions. J Am Med Assoc 1965;191:9195.
8 Brennan M, Perrin JCS, Canady A, Wesolowski D: Paraparesis in a child with a herniated thoracic
disc. Arch Phys Med Rehabil 1987;68:806808.
9 Arce CA, Dohrmann GJ: Herniated thoracic disks. Neurol Clin 1985;3:383392.
10 De Villiers PD, Booysen EL: Fibrous spinal stenosis: A report on 850 myelograms with a water-
soluble contrast medium. Clin Orthop 1976;115:140144.
11 Roberson GH, Llewellyn HJ, Tavenas JM: The narrow lumbar spinal canal syndrome. Radiology
1973;107:8997.
12 Paine KWE: Clinical features of lumbar spinal stenosis. Clin Orthop 1976;115:7782.
Burke/Gerszten 300
dramroo
13 Getty CJM: Lumbar spinal stenosis: The clinical spectrum and the results of operation. J Bone
Joint Surg Br 1980;62:481485.
14 Epstein JA, Epstein BS, Lavine LS, Carras R, Rosenthal AD: Degenerative lumbar spondylolis
thesis with an intact neural arch (pseudospondylolisthesis). J Neurosurg 1976;44:139147.
15 Verbeist H: Pathomorphologic aspects of developmental lumbar stenosis. Orthop Clin North Am
1975;6:177196.
16 Hall S, Bartleson JD, Onofrio BM, Baker HL Jr, Okazaki H, ODuffy JD: Lumbar spinal steno
sis. Clinical features, diagnostic procedures, and results of surgical treatment in 68 patients. Ann
Intern Med 1985;103:271275.
17 Spengler DM: Degenerative stenosis of the lumbar spine. J Bone Joint Surg Am 1987;69:305308.
18 Rosenberg NJ: Degenerative spondylolisthesis. Predisposing factors. J Bone Joint Surg Am 1975;
57:467474.
19 Gerszten PG: The socioeconomic implications of neurosurgical disease; in Linskey ME,
Rutigliano MJ (eds): Quality & Costs in Neurological Surgery. Philadelphia, Lippincott, Williams
& Wilkins, 2001, pp 295302.
20 Hart LG, Deyo RA, Cherkin DC: Physician office visits for low back pain: Frequency, clinical
evaluation, and treatment patterns from a U.S. national survey. Spine 1995;20:1119.
21 Andersson GBJ: Epidemiological features of chronic low-back pain. Lancet 1999;354:581585.
22 Deyo RA, Weistein JN: Low back pain. N Engl J Med 2001;344:363370.
23 Kelsey JL, White AA: Epidemiology of low back pain. Spine 1980;6:133142.
24 Mitchell LV, Lawler FH, Bowen D, Mote W, Asundi P, Purswell J: Effectiveness and cost-effectiveness
of employer-issued back belts in areas of high risk for back injury. J Occup Med 1994;36:9094.
25 Coste J, Delecoeuillerie G, Cohen de Lara A, Le Parc JM, Paolaggi JB: Clinical course and prog
nostic factors in acute low back pain: An inception cohort study in primary care practice. BMJ
1994;308:577580.
26 Cherkin DC, Deyo RA, Street JH, Barlow W: Predicting poor outcomes for back pain seen in
primary care using patients own criteria. Spine 1996;21:29002907.
27 Croft PR, Macfarlane GJ, Papageorgiou AC, Thomas E, Silman AJ: Outcome of low back pain in
general practice: A prospective study. Br Med J 1998;316:13561359.
28 Deyo RA: Low-back pain. Sci Am 1998;279:4853.
29 Bozzao A, Gallucci M, Masciocchi C, Aprile I, Barile A, Passariello R: Lumbar disk herniation:
MR imaging assessment of natural history in patients treated without surgery. Radiology
1992;185:135141.
30 Atlas SJ, Chang Y, Kammann E, Keller RB, Deyo RA, Singer DE: Long-term disability and return
to work among patients who have a herniated lumbar disc: The effect of disability compensation.
31 Weinstein JN, Birkmeyer JD: The Dartmouth Atlas of Musculoskeletal Health Care. The Trustees
of Dartmouth College, Dartmouth Medical School, 2000.
32 Watts C, Esser GB 3rd: Economic overview of spinal disorders; in Menezes AH, Sonntag VK
(eds): Principles of Spinal Surgery. New York, McGraw-Hill, 1996, pp 2536.
33 Conrad DA: Low back pain: Economic analysis of its impact on the U.S. non-elderly population; in
Bigos S (ed): Report of the Low Back Pain Guidelines Panel to the Office of the Forum of the Agency
for Health Care Policy and Research. Washington, Department of Health and Human Services, 1993.
34 Stripling TE: The cost of economic consequences of traumatic spinal cord injury. Paraplegia
News, August 1990, pp 5054.
35 DeVivo MJ: The cost of spinal cord injury: A growing national dilemma; in Apple DF Jr, Hydson
LM (eds): The Spinal Cord Injury Model: Proceedings of the National Consensus Conference on
Catastrophic Illness and Injury. Atlanta, 1989, pp 109113.
Peter C. Gerszten, MD, MPH
Department of Neurological Surgery, University of Pittsburgh Medical Center
200 Lothrop Street, Suite B-400, Pittsburgh, PA 15213 (USA)

dramroo
The Changing Economics of
Spine Surgery
Gregory J. Przybylski
Seton Hall University, Orange, N.J., USA
Advances in minimally invasive as well as complex spinal surgery during

the past decade have been revolutionary. The practical application of molecular,
biomechanical and computer engineering have enabled the spinal surgeon to
have more options available in treating patients with a variety of simple and
complex spinal disorders. However, the contraction of health care dollars avail
able to pay for this wonderful growth of technology has begun to limit the
choices available to the physician that hospitals are willing to purchase and
maintain. This chapter will examine the separate issues facing spinal surgeons
in their personal practice as well as the problems that hospitals face in manag
ing the ever-changing and concurrently more expensive technological innova
tions brought to us.
The complexity of managing the spinal physicians practice has faced sim
ilar dramatic change, requiring increasing direct physician involvement to
reduce the risk of fraud and abuse accusations and even prosecution. Accurate
use of current procedural terminology (CPT) coding and application of a
resource-based relative value system (RBRVS) methodology for physician
payment and practice expense accounting have been the hallmarks of this change.
In order to better adapt to these changes, it is important for the spinal practitioner
to understand the methodology behind CPT coding as it applies to spinal
surgery, as well as the use of RBRVS to the fee schedule and cost accounting
of a practice.
CPT remains the standard method by which physicians communicate to
the payer the services that were provided to their patients. This method for com
piling and communicating physician procedural services was standardized and
developed by the American Medical Association (AMA) in 1966 as current
procedural terminology (CPT) [1]. Although initially reflecting predominantly
dramroo
surgical procedures, the second version 4 years later expanded the description
of medical services using a five-digit coding system. The current fourth edition
was completed in 1977 and contained substantial revisions to include improve
ments in medical technology. Each year, the book is updated to reflect the elim
ination of older procedures that are no longer performed as well as the addition
of new procedures that reflect improvements in technology. The Health Care
Finance Administration (HCFA, now renamed CMS for the Centers for
Medicare and Medicaid Services) did not adopt CPT as part of their Common
Procedure Coding System (HCPCS) until 1983, after which it mandated use of
this system to report services for payment under Part B of the Medicare
program. Three years later, HCFA also required Medicaid agencies to use the
method.
The CPT system undergoes annual revision under the direction of the CPT
Editorial Panel. This 16-member panel meets quarterly and is comprised of 11
AMA-appointed physicians that serve 4-year terms. Four of these seats rotate
among specialists to allow a multidisciplinary influence. The other members of
the panel include the co-chairman of the Health Care Professionals Advisory
Committee (HCPAC), a representative from CMS, and appointees from the
Blue Cross and Blue Shield Association, the Health Insurance Association of
America, and the American Hospital Association. The panel is assisted by
AMA staff as well as the CPT Advisory Committee, which is comprised
predominantly of physicians selected by national medical specialty societies.
The largest change in CPT affecting the spine practitioner has been in the
reporting of anterior thoracolumbar spinal surgery performed by more than one
physician. With improvements in anesthetic technique and available instrumen
tation, greater attention had been drawn to anterior surgical approaches to spinal
diseases. Although some spinal surgeons perform their own thoracolumbar expo
sures, many utilize the expertise of other surgeons to perform the initial approach
to the anterolateral spinal column. Over several years, the American College of
Surgeons and Society of Thoracic Surgeons in cooperation with orthopedists and
neurosurgeons from 5 additional specialty societies worked to develop a consen
sus proposal to describe the approach component of anterior thoracolumbar spine
surgery. Using the model of skull base surgery that separated the approach from
the definitive procedure, emphasis was placed upon creating a scheme that would
allow separate coding of the surgical exposure and closure from the spinal
decompression or reconstruction. After several years of committee and work
group discussions and multiple presentations before the CPT Editorial Panel, the
surgical representatives concluded that the substantial variability in physician
practice made such a proposed system excessively complicated. Consequently, a
consensus proposal was presented to the CPT Editorial Panel in February of 2001
requesting an expanded application of the 62 co-surgery modifier.
Economics of Spine Surgery 303

dramroo
For several years previously, the 62 co-surgery modifier could only be
used once per operative session. Since the approach has been considered
included in the work value of the decompression and arthrodesis codes by the
Relative Value Update Committee (RUC) and by CMS, it was proposed that the
additional level codes describing adjacent segment decompression and/or
arthrodesis must also contain a component representing the additional work of
exposing the additional level(s). Concerned about the potential financial impact
of expanding the use of the 62 co-surgery modifier, representatives from CMS
presented data regarding the current usage of 62 as well as a summary of
actual claims data in which anterior thoracolumbar surgery was performed by
more than one surgeon. Several interesting observations made included less
frequent usage of 62 to describe this work than was expected, but more impor
tantly, serious concerns were raised about actual payments made on claims. For
example, the correct method for reporting co-surgery requires both the
approach surgeon and the spinal surgeon to submit the same code (though not
necessarily the same charge) appended with the 62 modifier. However, CMS
identified claims in which one surgeon used the modifier, whereas the other did
not. Rather than paying both surgeons 62.5% of the Medicare allowable, only
the surgeon coding correctly with the modifier was paid 62.5% of the allow
able. In contrast, the other surgeon was paid 100% of the allowable, raising the
vigilance of CMS in scrutinizing past and future claims regarding anterior
thoracolumbar spine surgery. The CPT Editorial Panel accepted the consensus
proposal to expand the use of the 62 modifier to allow reporting of the addi
tional physician work involved in approaching adjacent segments for decom
pression or arthrodesis. The modifier will not be applicable to instrumentation
or bone graft harvest codes, as specific language will be provided listing codes
excluded from use of the modifier. These changes were implemented for
January 2002.
With changes in the structure and management of CMS, an effort has been
made to improve the transparency of certain activities, one of which is the
correct coding initiative (CCI). Despite the large number of codes available to
describe spinal procedures, the work included in many of these codes overlaps.
The process of unbundling, which involves description of a larger procedure
with several codes that contain overlapping work, resulted in significant
increases in health care expenditures. On January 1, 1996, Medicare initiated
the CCI to reduce unbundling and inappropriate reporting of CPT codes. At that
time, HCFA contracted with Administar Federal, an Indiana Medicare carrier,
to create and maintain a nationally used computer software program to preclude
payment of otherwise bundled services. Despite the USD 700,000 development
cost, Medicare reported savings in excess of USD 700 million since the
programs inception. Most of the edits represented payment policies in which
Przybylski 304
dramroo
a comprehensive code would be paid while the component code would be
disallowed, whereas a small percentage of edits identified mutually exclusive
codes which would not be performed concurrently.
An example of the limitations of CCI placed on spinal surgery can be found
in posterior lumbar interbody fusion (PLIF, CPT code 22630). The service
description of posterior interbody fusion includes laminectomy, facetectomy,
and discectomy to approach and prepare the endplates for the interbody
arthrodesis. Consequently, CCI includes edits that preclude coding PLIF with
many decompression procedures. However, this does not account for circum
stances in which decompression is performed beyond the typically bony removal
required to perform a PLIF. An editorial revision to 22630 was instituted in 2001
that included only the laminar, facet and disk removal required for performance
of the PLIF, thereby allowing for coding of additional decompressive work if
performed and medically indicated. The CCI process prevents payment for these
additional physician services for the most part, unless a 59 modifier is also
appended to identify the code as a distinct procedural service from the PLIF.
This example distinguishes the differences between coding rules (developed and
maintained by the AMA) and reimbursement rules (developed individually by
CMS and third party payers).
However, the development of RBRVS has had the greatest impact upon the
billing practices of physicians this past decade after its full implementation on
January 1, 1996. The impetus to revise the Medicare payment system arose
from the rapidly increasing expenditures for payment of physician and hospital
services by Medicare. Since hospital services accounted for more than two
thirds of Medicare expenditures, cost containment efforts were naturally
directed at hospitals first. In 1983, a prospective pricing system (PPS) for hos
pital services using a diagnosis-related group (DRG) payment was developed
for approximately 500 diseases. Using the national average cost of hospital care
for that particular illness, it was assumed that the average cost for providing
care for patients with a range of illness severity would equal the calculated
DRG payment. Additional payments were also authorized to account for unusu
ally severe illnesses requiring prolonged hospital stays. Since the payment was
identical regardless of the hospital cost, the PPS provided a strong incentive for
hospitals to improve cost-efficiency. Not surprisingly, the annual growth of
Medicare expenditures was reduced by more than half between 1975 and 1990.
In 1994, physician payments on behalf of Medicare beneficiaries repre
sented more than three quarters of the expenditures from Medicare Part B [2].
The original method for determining the physician payment schedule was based
on customary, prevailing and reasonable (CPR) charges. This resembled the
usual, customary, and reasonable (UCR) charge system utilized by private insur
ers to pay for physicians services based upon their actual fees, but included

dramroo
some adjustments to keep government costs predictable. However, the wide
variation in the amount Medicare paid for physician services both among physi
cian specialties as well as among geographical regions caused dissatisfaction
within the medical profession [3]. Although costs were initially controlled by
reducing payments from the 90th to the 75th percentile prevailing charge,
Medicare later introduced a temporary price freeze on physicians services. In
1976, this was replaced by linking increases in prevailing charges to the
Medicare Economic Index (MEI). As a result, Medicare payments were based
upon prevailing charges in 1971 and remained nearly unchanged until 1992.
After a second freeze on payment levels in the 1980s and reduced pay
ments for overpriced surgical procedures, several payment reform proposals
were suggested by the government. Although additional modifications to CPR
were considered, it seemed that this method evolved into a complex system that
no longer reflected physician fee schedules. Consideration was given toward
developing a DRG system similar to that developed for hospital payments
under Medicare Part A [4]. Another option was to create a managed care or
capitation model of payment. Finally, a proposal was offered for replacing the
CPR method with a payment schedule based on a relative value scale.
The concept of a relative value system (RVS) was not new. In fact, the
California Medical Association developed an RVS in 1956, with regular updates
for nearly two decades. Although this method still represented a CPR charge sys
tem, physicians used the RVS to determine fee schedules, whereas government
and private insurers used the system to establish payment rates. However, the
Federal Trade Commission raised concerns regarding the possible antitrust
violations from price fixing, prompting the California Medical Association to
discontinue updating the charge data collected.
Rather than using a charge-based RVS, a resource-based RVS was adopted
in which physician services were valued based upon the relative costs incurred
in providing them. Since antitrust concerns precluded direct physician involve
ment in the development of RBRVS, the AMA accepted a proposal submitted
by the Harvard University School of Public Health to perform a national study
of resource-based relative value scales (RBRVS) for physician services.
On July 1, 1986, the Consolidated Omnibus Budget Reconciliation Act
mandated that the department of Health and Human Services develop an RBRVS
to be submitted to Congress. In addition, the law created the Physician Payment
Review Commission (PPRC) to study a variety of additional options for
Medicare payment reform including changing CPR, capitation, and physician
DRG [5]. The PPRC recommended development of a payment schedule linked
to an RBRVS [6].
Six months earlier, the Harvard study, under the direction of William
Hsiao, PhD and Peter Braun, MD, began with funding from HCFA. The first
Przybylski 306
dramroo
phase of the national study supported development of an RBRVS for 12 physi
cian specialties. In addition, independent funding was obtained for the study of
6 additional specialties. Not only were specialty-specific scales developed, but
also a method for creating cross-specialty links allowed integration of a single
cross-specialty RBRVS. The Omnibus Budget Reconciliation Act of 1986
provided a 2-year extension for submission of RBRVS to Congress as well as
mandated inclusion of 15 additional specialties during the second phase of the
study [710]. Although the AMA adopted in principle the results of the Harvard
study, they also recommended that the new Medicare payment system include
geographical differences in practice costs and professional liability as well as a
transition period to prevent disruptive changes between the CPR and RBRVS
systems. The PPRC likewise endorsed the study and supported these AMA
recommendations.
In December of 1989, Congress enacted the Omnibus Budget
Reconciliation Act (OBRA 89) that mandated a Medicare payment schedule
based on RBRVS from the Harvard study with inclusion of physician work,
practice expense, and professional liability costs. Geographical adjustments to
all three components were included. Calculation of a relative value unit (RVU)
of a physicians service under RBRVS is as follows:
MFS [(RVUw GPCIw) (RVUpe GPCIpe) (RVUm GPCIm)] CF
where MFS Medicare fee schedule, GPCI geographical practice cost

index, w work, pe practice expense, m malpractice and CF conversion
factor.
The conversion factor for the calculated RVU was based on keeping the
overall Medicare expenditure the same as the cost using the CPR system.
A process was created to annually adjust the conversion factor, maintaining a
budget-neutral value, which limited increases in expenditures to USD 20
million annually. The final comprehensive RBRVS for physician services was
published in the November 25, 1992 Federal Register.
After implementation of the RBRVS fee schedule by Medicare, additional
health insurers have gradually implemented an RVS as well. Although RBRVS
is the method most commonly used, an alternative RVS called St. Anthonys is
also used in certain regions. This privately maintained system, formerly known
as McGraw-Hill, utilizes RVUs based upon the time, risk, and complexity of the
physician service. In contrast, the RBRVS uses physician work, practice
expense, and malpractice expense as the components for determining the rela
tive value of a particular physician service. The physician work component com
prises approximately 55% of the total relative value of the service, whereas
practice expense comprises 42%. In addition, a geographic practice cost index is

dramroo
45 Conversion factor
40
35
30
25
USD
20
15
10
0
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year
Fig. 1. Change in Medicares surgical conversion factor over the past decade.
incorporated to adjust for geographical differences. Consequently, conversion

factors between RBRVS and St. Anthonys RVS are different. Based upon the
congressional mandate to maintain budget neutrality, CMS adjusts the conver
sion factor annually to assure budgetary control of expenditures for physician
services (fig. 1). The recent 5.4% reduction in the conversion factor has
prompted multitudes of physician organizations to petition Congress for legisla
tion to reform the method for tying Medicare expenses to the gross domestic
product. Growing evidence of reduced access for Medicare patients, including
an American Academy of Family Physicians survey finding 17% of the respon
dents are no longer accepting new Medicare patients and a similar survey by the
North American Spine Society showing a greater proportion of spine practition
ers limiting access to Medicare patients, may prompt change in the continued
contraction of reimbursement.
In order to coordinate changes in CPT with assignment of RVU by CMS,
the AMA/Specialty Relative Value Update Committee (RUC) was created in
1991. Twenty-three of the twenty-eight members are appointed by major
national medical specialty societies. The other five panelists include the RUC
chair, the co-chair of HCPAC, and members of the AMA, American Osteopathic
Association, and CPT Editorial Panel. A RUC Advisory Committee composed
of members appointed by 94 specialty societies develop and suggest RVU for
new codes to the RUC. Specialty society representatives are responsible for
compiling physician survey data to determine the time spent in performing the
medical service and ranking the service relative to existing services.
Subsequently, consensus recommendations are forwarded to HCFA for annual
consideration. During the 5-year period ending in 1998, nearly 2,300 relative
Przybylski 308
dramroo
value recommendations were made to HCFA with a recent acceptance rate of
more than 90%.
The spine practitioner can take advantage of the RBRVS system in deter
mining the cost required to provide physician services and thereby in developing
a fee schedule that reflects these costs. Although practice costs are not linearly
related to physician work, the RVU can serve as a surrogate. Consequently, a fee
schedule can be constructed based on a conversion factor determined by the
practice and applied to the RVU assigned by Medicare to procedural codes.
However, the appropriate conversion factor is influenced by many factors includ
ing personnel, equipment, and insurance (disability, health, malpractice) costs
among others. The practice manager should determine the average annual RVU
performed for the entire practice as well as stratified by individual physician.
As a result, simply dividing the total practice (or individual physician) costs by
the RVU performed during the same period provides a cost/RVU figure that
reflects the expense to provide physician services.
Determination of cost/RVU is essential in negotiating contracts with third-
party payers. Since many third-party payers have adopted fee schedules based
on RBRVS, it is to your advantage to determine your costs on a similar scale.
First of all, your analysis of the payment schedule is much more meaningful
once you have determined your practice cost to provide the service. Secondly,
the cost analysis allows the physician to focus on elements of the practice in
which costs can be reduced. Finally, one can assess the time required to perform
particular services. Certain services may be more economical than others,
thereby allowing the physicians to focus efforts on their most efficient services.
Although work values are resource-based, practice expense was based
upon the AMAs Socioeconomic Monitoring System 1989 Core Survey (SMS).
While overall practice costs including office rent, wages of nonphysician per
sonnel, equipment and supplies were measured, costs specific to a given single
procedural service were unknown. Moreover, practice expenses varied among
specialties. For example, these expenses represented 52.2% of family physi
cians practice costs but only 38.9% of neurosurgeons costs. In contrast, the
average neurosurgical professional liability component of practice cost was
7.6%, compared with a 3.9% proportion for the family physician. The method
enacted by OBRA 89 involved multiplying the specialty-specific practice
expense factor by the average Medicare payment of the service in 1991.
Similarly, professional liability was calculated based upon the proportion of
cost multiplied by the Medicare payment.
The Omnibus Budget Reconciliation Act of 1993 mandated reductions in
the practice expense of overvalued services. Over a 3-year period, the practice
expense was reduced annually by 25% of the amount that the value exceeded the
physician work RVU until it was no greater than 128% of the work component.

dramroo
However, concerns over the non-resource-based OBRA 89 method of calculat
ing practice expense as well as a 1994 Social Security Act amendment mandat
ing resource-based practice expense calculations prompted HCFA to contract
with Abt Associates, to perform a national study of physicians practice expense.
Fifteen Clinical Practice Expert Panels (CPEPs) were formed from nominations
by medical associations to develop a list of direct cost components of a selected
group of reference codes. In addition, a national mail survey of 5,000 practices
was performed to obtain a sample of practice costs and service mix to validate
the CPEP estimates. However, poor response rates from physicians led HCFA to
abandon efforts at obtaining actual survey data. In order to meet the deadline of
January 1998 set by Congress, HCFA planned to implement new practice
expense values based on CPEP data. Lack of validation of the CPEP data with
actual practice expense information as well as failures to account for actual
differences in practice cost among all specialties prompted the AMA to urge
Congress to extend the deadline for implementation of new practice expense
values. As a result, Congress not only delayed implementation until 1999, but
also directed the General Accounting Office (GAO) to review HCFAs method
ology and make recommendations for a valid resource-based model. The GAO
report supported the concerns raised by the AMA.
Beginning in January 1999, HCFA initiated a transition to a resource-based
practice expense valuation that differs based on the site of the service. If a med
ical service can be performed either in an office or a hospital, then both a non-
facility and facility practice expense value will be assigned. Whereas locations
assigned nonfacility practice expense include physician offices and independent
imaging or laboratory centers, facility practice expense is attributed to hospitals,
surgical centers, and nursing homes. The method for estimating practice costs is
based upon the AMAs SMS data. Since the SMS data came from small sample
sizes with sufficient variability to introduce sampling bias, there continues to be
concern regarding this methodology. The transition to a resource-based method
for calculating practice expense has been largely responsible for gradual decline
in RVU assigned to spinal and other surgical codes.
More recently, rapidly escalating professional liability costs have had a
substantial impact on the spinal surgeons practice. A combination of a decade
of underestimating the underwriting costs of these policies by insurers and
increasingly frequent large judgments has resulted in a national professional
liability crisis. For example, it has been estimated that USD 1 in premiums have
been collected for every USD 1.36 in claims paid [11]. At the same time, the
average medical liability award increased from USD 1.95 million to USD 3.5
million between 1993 and 1999 [12]. The Pennsylvania Medical Society has
found that liability insurance charges were hiked 2191% in 2001 alone.
Similarly, a recent survey presented at the Council of State Neurosurgical
Przybylski 310
dramroo
Societies meeting in 2001 revealed substantial increases in insurance costs in
the past year, which have led some physicians to curb practice or move to more
favorable geographical regions. For example, Indiana and California have
adopted tort reform legislation that includes caps on noneconomic damages that
have helped in reducing the unreasonable judgments awarded by some juries.
Unfortunately, CMS does not update the professional liability component of the
RVU on an annual basis, leading to a significant underestimation of these costs
for certain specialties including orthopedic and neurological surgery.
The AMA estimates that 12 states are currently experiencing a medical
liability crisis, whereas 30 additional states are on the verge of similar problems.
For example, the closure of a level 1 University Trauma Center in Las Vegas,
Nev. prompted the state legislature to temporarily give the physicians state
employee status, thereby limiting their liability to USD 50,000. The growing
liability crisis has prompted a bipartisan bill termed the Health Act that seeks
to institute national tort reform and includes limits on noneconomic damages
modeled after California and Indiana.
Finally, the increasing influence of governmental regulation on the practice
of medicine will continue to have a significant influence upon the manner in
which spinal surgeons provide health care. For example, the Kennedy-
Kassebaum Health Insurance Portability and Accountability Act 1996 (HIPAA)
changed the US Governments Fraud and Abuse regulations by increasing civil
monetary damages from USD 2,000 to USD 10,000 and by applying fraud and
abuse laws to the private as well as the public sector, by permitting confiscation
of personal property for health care fraud convictions, and by changing health
care frauds from misdemeanors to federal felonies with mandatory prison sen
tences. Surgeons are liable for fraud and abuse violations in the documentation,
coding and billing tasks of their practice. Compliance plans and programs are
aimed at satisfying the Office of the Inspector Generals requirements for the
constant surveillance of these responsibilities. The requirement to comply with
the complex regulations that have evolved from the law begins next year and
poses a significant economic burden for smaller practices in developing and
maintaining compliance programs.
Similarly, the regulations that have arisen from the Emergency Medical
Treatment and Labor Act (EMTALA) of 1986 apply to physicians providing on-
call services to emergency departments. The requirements that have evolved
from the expanded interpretation of EMTALA include the need for the on-call
physician to arrive to the emergency department within 30 min of notification
of a true emergency. Possible violations of the act include inability to see a
patient at a second hospital if call is simultaneously taken at two hospitals and
both require concurrent patient evaluation and inability to leave an elective
procedure if a patient arrives in the emergency room requiring attention by

dramroo
the operating surgeon. Both hospitals and physicians can be fined up to
USD 50,000 for each violation, in addition to the damages allowed by state law
for injury sustained by the patient from an EMTALA violation.
Although the changing economic environment for the physician has been
quite dramatic over the past decade, hospitals have faced similar issues that
have resulted in more frequent closures secondary to bankruptcy. With increas
ing granularity of actual hospital costs to provide patient services, administra
tors have identified areas that are being targeted for cost containment. For
example, the cost of spinal implants has had a significant impact upon the
hospitals expense in treating spine patients. In the absence of carve-outs that
exclude implant cost from the DRG, much if not all of the hospital reimburse
ment for a patients particular diagnosis is needed for purchase of the implants,
thereby creating a shortfall in accounting for the nursing, medication, imaging,
and facility costs of the procedure.
As a result, hospital administrators are becoming increasingly influential
in determining the types and source of implants available for the spinal surgeon
to use. With many manufacturers vying for an opportunity to provide implants
and services, increasing price competitiveness may result in single-source
purchase of equipment and implants that may be different from the physicians
preference. Efforts at proactive cost containment by the physicians may help the
hospital while simultaneously allowing the physician to participate in the
decision-making process. For example, the Lahey Clinic implemented a system
by which the equipment vendors supplied all of the implants as well as the
disposable and nondisposable instruments to perform a particular procedure
(termed single price/single case purchasing). As a result, costs were reduced by
2345% [13]. Although some savings required a change of vendors, some of
the equipment savings were achieved with maintenance of the same provider.
Another area of increasing scrutiny involves outpatient surgery. Although
there have been efforts in place for some time now to reduce length of hospital
stay (and therefore cost) for patients recovering after spinal procedures, some
procedures like lumbar discectomy are now performed on an outpatient basis.
However, CMS has suggested that the costs to provide an outpatient procedure are
typically less than the cost to provide the service to an inpatient. Consequently,
CMS has identified a group of outpatient procedures with a different payment
schedule. As a result, the hospital may not receive sufficient payment for proce
dures performed on an outpatient basis that are identified as inpatient procedures.
This policy may curtail efforts at cost containment, as procedures are not
approved by the hospital unless the patient is kept overnight.
In conclusion, the dramatic improvements in technology that have allowed
the spinal surgeon a vastly greater armamentarium for the management of
spinal disorders have been mirrored by an increasingly complex socioeconomic
Przybylski 312
dramroo
environment that has both directly and indirectly affected the daily practice of
the spinal surgeon. An awareness of and adaptation to this changing environ
ment should help the spinal surgeon in meeting both the clinical demands of
treating patients with the economic constraints that may limit options going
forward. Increasing individual as well as professional societal involvement in
the legislative process is imperative to shaping a more favorable landscape for
practicing spinal surgery.
References
1 Gordon BL: Current Procedural Terminology. Chicago, American Medical Association, 1973.
2 Committee on Ways and Means: Overview of entitlement programs: 1994 Green Book. Washington,
Committee on Ways and Means, 1994, p 1075.
3 Burney IL, Schreiber GJ, Blaxall MO, Gabel JR: Geographic variation in physicians fees. JAMA
1978;240:13681371.
4 Mitchell JB: Physician DRGs. N Engl J Med 1985;313:670675.
5 Physician Payment Review Commission: Medicare physician payment: An agenda for reform;
in Annual Report to Congress No 68-227. Washington, US Government Printing Office, 1987.
6 Physician Payment Review Commission: Medicare physician payment; in Annual Report to
Congress. Washington, US Government Printing Office, 1988.
7 Hsiao WC, Braun P, Becker ER, Thomas SR: The resource based relative value scale. JAMA
1987;258:799802.
8 Hsiao WC, Braun P, Dunn D, Becker ER: Resource based relative values: An overview. JAMA
1988;260:23472353.
9 Hsiao WC, Braun P, Dunn D, Becker ER: Results and policy implications of the resource based
relative value study. N Engl J Med 1988;319:881888.
10 Hsiao WC, Braun P, Yntema D, Becker ER: Estimating physicians work for a resource-based rel
ative value scale. N Engl J Med 1988;319:835841.
11 Zimmerman R, Oster C: Insurers price wars contributed to doctors facing soaring costs. Wall
Street Journal, June 24, 2002.
12 Marcus MB: Healthcares perfect storm; soaring malpractice costs are causing critical shortages
of doctors. US News and World Report, July 1, 2002.
13 Healy WL, Iorio R, Lemos MJ, Patch DA, Pfeifer BA, Smiley PM, Wilk RM: Single price/case
price purchasing in orthopaedic surgery: Experience at the Lahey Clinic. J Bone Joint Surg Am
2000;82:607612.
Gregory J. Przybylski, MD
The New Jersey Neuroscience Institute at JFK,
Seton Hall University, 65 James Street, Edison, NJ 08818 (USA)

dramroo
Author Index
Alexander, J.T. 39 Kim, B.-S. 266 Sonntag, V.K.H. IX

Anderson, D.G. 14 Kuklo, T.R. 96 Subach, B.R. 1, 277
Swartz, K.R. 29
Bartolomei, J. 71 Lenke, L.G. 188
Bertagnoli, R. 176 Lieberman, I. 240 Traynelis, V.C. 165
Birch, B.D. 128, 154 Lunsford, L.D. VIII Trost, G.R. 29
Burke, J.P. 290
Burkus, J.K. 55 McLaughlin, M.R. 71 Welch, W.C. 39
Mummaneni, P.V. 154, 165,
Casey, A.T.H. 142 213, 225
Chun, J.Y. 154
Crockard, H.A. 142 Ondra, S.L. 213, 225
Fessler, R.G. 251, 266 Perez-Cruet, M.J. 251, 266

Fiore, A.J. 128 Petraglia, C.A. 1
Foley, K.T. 55, 84 Polly, D.W., Jr. 96
Przybylski, G.J. 302
Gerszten, P.C. 290
Rampersaud, Y.R. 84
Haid, R.W., Jr. 1, 55, 128, Resnick, D.K. 39
277 Rodts, G.E., Jr. 1, 204, 277
Helm, G.A. 14
Sandhu, F. 266
Kaiser, M.G. 277 Sasso, R.C. 154, 165, 213,
Kalfas, I.H. 107 225
314
dramroo
Subject Index
Adeno-associated virus, gene therapy prospects 23, 24
vector 17, 18
vectors 1719
Adenovirus 17, 35
osteoinductive activities 2, 3
Anterior lumbar interbody fusion,

processing 16
see Lumbar interbody fusion purification 3, 4
Atlantoaxial instability, see C1 lateral receptor signaling 16
mass fixation structural homology between
osteoinductive proteins 1, 15
Bone formation, see Osteogenesis types and functions 1, 34
Bone grafts
Bracing
cadaveric sources 30
adolescent idiopathic scoliosis 220
carriers, see also specific carriers

structural kyphosis 223
classification 31
Brain Lab system, virtual fluoroscopy
commercial products 31
7981
pros and cons of substances 32
comparisons of autografts, allografts,

C-arm fluoroscopy, see Virtual
and substitutes 36, 40
fluoroscopy
cost benefits of synthetic materials
C1 lateral mass fixation
40, 41
advantages 128, 136, 137, 140
growth factors and delivery system

atlantoaxial fixation 136, 137
3335
biomechanical characteristics
ideal properties 29
138
success factors 36, 37

image guidance 139
toxicity of allograft products 40

large C2 nerve roots 139
Bone morphogenetic proteins occipitocervical fixation 137
BMP-2 recombinant protein and outcomes 130, 132135
clinical use, see rhBMP-2

patient selection 129, 136138
BMP-7 studies in autografts 36

technique 129, 130, 140
gene therapy
training 140
direct 19, 20
C1C2 screw fixation
ex vivo 20, 22, 23

contraindications to transarticular
overview 4, 16
screw placement 136
315
dramroo
C1C2 screw fixation (continued)
closure 261, 262
image guidance
complications and avoidance 262264
C1C2 segmental screw fixation 119,

development 251
121
indications 252255
transarticular screw fixation 115,

instrumentation 253
117119
operative setup 255257
Calcium sulfate, bone repair 41

postoperative care 262
Carbon fiber reinforced plastic implants

preoperative evaluation 253
advantages 176, 185, 186

technique 257, 259261
animal studies 176, 177

training 264
development 176
Cervical pedicle screws
interbody fusion study

biomechanical stability 156, 157
cervical outcomes 181184

breach rates by technique 156
devices 178
C2 pars screw placement 158
lumbar outcomes 184, 185

C2 pedicle screw placement 159, 160
neurologic status assessment 181

C3-C6 pedicle screw placement
pain assessment 181, 183

160162
patient satisfaction 181185

C7 pedicle screw placement 162
patient selection 179, 180

clinical studies 157, 158
radiographic assessment of fusion

historical perspective 154
180, 181, 183

image guidance 162, 163
manufacture 176
indications 163
mechanical tests 177, 178, 181

pedicle dimensions 154156
radiolucency 176
Cervifix, posterior cervical lateral mass
Cervical laminoplasty
fixation 166
advantages of expansive cervical

Chemotherapy, vertical compression
laminoplasty 142, 143, 149, 151

fracture considerations 241
alternatives 148, 149

Civil damages 311
complications
Clinical Practice Expert Panel 310
pain 150, 151

Cobb angle, measurement 219
radiculopathy 149, 150

Computed tomography
contraindications 143
lumbar interbody fusion assessment 58,
French door vs open door 143

59, 6267
historical perspective 142

myelography 253
indications 143
spinal stereotaxis 72, 73
kyphosis 142
thoracic pedicle screw placement 104,
outcomes 147149
189, 198
queen square technique 145147

virtual fluoroscopy registration 78, 79,
saws and drills 145

108, 111
Z plasty 143
Computer-assisted fluoroscopy, see Virtual
Cervical lateral mass fixation,

fluoroscopy
see C1 lateral mass fixation, Posterior

Coralline grafts, bone repair 41, 42
cervical lateral mass fixation

Correct coding initiative, Medicare 304,
Cervical microendoscopic discectomy

305
and laminoforaminotomy
Current procedural terminology
advantages 251253
correct coding initiative limitations on
anesthesia 255
spinal surgery 305
Subject Index 316

dramroo
development 302, 303
ex vivo therapy 20, 22, 23
Medicare acceptance 303

heterodimer production 24
revision 303, 304

overview 4, 16
spine surgery 303

prospects 23, 24
Customary, prevailing, and reasonable

comparison of ex vivo and in vivo
charges 305, 306
techniques 34
direct plasmid injection 18
Diagnosis-related group, Medicare

electroporation 18, 19
payment 305, 312

gene gun 19
Dynamic reference array, virtual

liposome delivery 19
fluoroscopy 86, 88, 89, 93

LMP-1 23, 35
polymer-DNA complexes for delivery
Economic constraints, spinal surgery 19
302312 principles 16, 33
Electromyography viral vectors
cervical microendoscopic discectomy

adeno-associated virus 17, 18
and laminoforaminotomy planning

adenovirus 17, 35
254, 255
herpes virus 18
intraoperative 72
thoracic pedicle screw placement

Health Insurance Portability and
198
Accountability Act 311
Emergency Medical Treatment and

Herniated disc, see Cervical
Labor Act 311, 312
microendoscopic discectomy and
laminoforaminotomy, Thoracic
Fixation devices microendoscopic discectomy
bioresorbable devices, see also specific
Herpes virus 18
devices and materials

Hyaff-11, tissue engineering 31
advantages 45
Hydroxyapatite, coralline grafts 41, 42
applications 47, 50, 51
foreign-body reactions 47
Image-guided spine surgery
prospects 51
anatomic landmarks 110
resorption 46
C1 lateral mass fixation 139
wear 45
cadaver testing 113
Fluoronav, see Virtual fluoroscopy

cervical pedicle screw placement 162,
Fluoroscopy, see Virtual fluoroscopy

163
Food and Drug Administration approval

cervical spine applications
resorbable biomaterials 39, 40

advantages over fluoroscopic
rhBMP-2 7
navigation 124, 125
Foraminal stenosis, see Cervical C1C2 segmental screw fixation
microendoscopic discectomy and 119, 121
laminoforaminotomy C1C2 transarticular screw fixation
Free-hand pedicle screw technique, 115, 117119
see Thoracic pedicle screws limitations 123, 124
navigational technique 114, 115
Gene therapy
overview 113, 121, 123
bone morphogenetic proteins

prospects 125, 126
direct gene therapy 19, 20

transoral decompression 121
Subject Index 317

dramroo
Image-guided spine surgery (continued)
Lenke classification, adolescent idiopathic
equipment 74, 75
scoliosis 226
frameless navigation and advantages

Liability insurance
7476, 81
crisis 311
instrumentation 108, 109

trends in rates 310, 311
patient movement 112

LIM-mineralizing protein-1, gene therapy
probe orientation 112, 113
23, 35
reference arc generation 74

Lordosis, see Thoracolumbar deformity
registration
Low back pain
computed tomography with virtual

diagnosis 294
fluoroscopy data registration 78,

economic impact 293, 299
79, 108, 111

epidemiology 293295
paired point registration 110, 111

recovery 294, 295
purpose 111
LT cage
surface matching 111

bone morphogenetic protein applications
spinal stereotaxis 7276

7, 8
thoracic pedicle screw placement 100,

radiography 62, 63
101
Lumbar interbody fusion
tracking devices 86, 111, 112

anterior lumbar interbody fusion
unreliability of intraoperative
advantages 277
radiography 71, 72, 84, 85, 107, 108

comparison of laparoscopic and
virtual fluoroscopy, see Virtual

mini-open approaches
fluoroscopy
complications 286288
Implants, cost control measures 312
operating room organization time
InFUSE, see rhBMP-2
286
Interbody carbon fiber, see Carbon fiber

operative time 285
reinforced plastic implants recommendations 288
retrograde ejaculation 288
King classification, adolescent idiopathic

evolution 277, 278
scoliosis 226
laparoscopic approach
Kyphoplasty
overview 278
outcomes 246248
technique 279281
overview 241, 242, 244, 245

mini-open technique
percutaneous vertebroplasty comparison

overview 278, 279
247
technique 282284
technique 245, 246

patient positioning 279
Kyphosis
postoperative care 285
epidemiology 291
assessment
treatment, see Thoracolumbar deformity

bone formation and remodeling 63,
64, 68
Lactosorb
computed tomography 58, 59, 6267
applications 47
device-bone interface 6163
implantation technique 48
dynamic plain radiographs 57, 58, 68
materials 45, 46
histologic biopsy 55
resorption 46, 48, 49

plain radiographs 56, 57
strength 48
segmental spinal stabilization 61
Laminoplasty, see Cervical laminoplasty spinal alignment 59, 60, 68
Subject Index 318

dramroo
fusion zones 6567
vertical compression fracture,
LT cage
see Vertical compression fracture
bone morphogenetic protein

Outpatient surgery, cost control measures
applications 7, 8
312
radiography 62, 63
rhBMP-2 application 6, 7, 35, 64, 65

Pedicle screw
variability in fusion rates 55
adolescent idiopathic scoliosis surgery
Lumbar pedicle screws, see Percutaneous

anterior derotation and fusion 230, 231
lumbar pedicle screws

posterior fusion 228230
Lumbar spondylosis, epidemiology 292

cervical screws, see Cervical pedicle
screws
Macropore
lumbar screws, see Percutaneous lumbar
applications 50, 51
pedicle screws
imaging 50
thoracic screws, see Thoracic pedicle
materials 46
screws
resorption 46, 49
unreliability of intraoperative
strength 49
radiography in placement 71, 72, 84,
Magnetic resonance imaging 85, 107, 108
cervical microendoscopic discectomy

virtual fluoroscopy
and laminoforaminotomy planning

misplacement rates 91, 92, 94
253, 254
percutaneous pedicle screw placement
spinal stereotaxis 72, 73

88, 89
Medicare Economic Index 306

posterior spinal procedures 88
Medicare fee schedule 307

Pedicle subtraction osteotomy, kyphosis
Mesenchymal stem cell, see Stem cell

management 236
Microendoscopic discectomy, see Cervical

Percutaneous lumbar pedicle screws
microendoscopic discectomy and

closure 210
laminoforaminotomy, Thoracic
indications 204, 205, 211
microendoscopic discectomy
K wire drilling 208
medial transverse process as entry point
Osteoconduction
206, 207
bone grafts 2, 31
operating time 210, 211
definition 30
outcomes 210, 211
Osteogenesis
overview 204
growth factors 30
screw extender 208, 209
lumbar interbody fusion assessment 63,

Sextant device 208, 209
64, 68
stab incisions 208
overview 29, 30
virtual fluoroscopy 88, 89, 205, 206, 212
Osteoinduction
Percutaneous vertebroplasty
bone grafts 29
complications 243, 244
definition 30
contraindications 243
Osteolysis
development 242
prevalence 240
kyphoplasty comparison 247
vertical compression fracture,

outcomes 243, 248
see Vertical compression fracture

overview 241
Osteoporosis
pain relief mechanism 242, 243
prevalence 240
technique 242, 243
Subject Index 319

dramroo
Physical therapy
rhBMP-2
adult patients with progressive
Food and Drug Administration
approval 7
postural kyphosis 222

LT cage application 7, 8
PLA-DX-PEG, tissue engineering 32

preclinical studies 35
Polyactive 70/30, animal vs human bone

production 2, 4, 9
repair 44, 45
prospects for study 8, 9
Polyetheretherketone, tissue engineering

spinal fusion studies
32, 33
human studies 6, 7, 35, 64, 65
Polyglycolic acid
interbody spinal fusion animal
biocompatibility 47
models 6
degradation 45
posterolateral lumbar fusion animal
fixation devices 45
models 4, 5
Poly(L-lactic acid)
stem cell differentiation induction 3
biocompatibility 47
degradation 45, 48
Scheuermann disease, structural kyphosis
fixation devices 45, 46
222
tissue engineering 33
Scoliosis, see Thoracolumbar deformity
Polymethyl methacrylate
Segmental spinal stabilization, lumbar
complications in use 244

interbody fusion assessment 61
percutaneous vertebroplasty 241244

Sextant device, percutaneous lumbar
tissue engineering 32
pedicle screw placement 208, 209
Posterior cervical lateral mass fixation

Smith-Peterson osteotomy, kyphosis
advantages 165
management 236, 237
commercial system overview

Spinal alignment, lumbar interbody
Cervifix 166
fusion assessment 59, 60, 68
ideal systems 171

Spinal cord injury
Starlock 166
economic impact 300
SUMMIT 167, 171173

epidemiology 292, 293
VERTEX 166, 167, 171173

Spinal stereotaxis, see Image-guided
complications 171
spine surgery
disadvantages 165
Spine surgery rates
exposure 167, 168

fusion surgery 295297
screw placement with VERTEX system

incidence by type of surgery 295
168170 regional variation in United States
297, 298
Radiation therapy, vertical compression

specialists and procedures 298, 299
fracture considerations 241

trends among Medicare enrollees 295,
Radiculopathy
296
epidemiology 290, 291

Spondylosis, epidemiology 290292
natural history 291

Starlock, posterior cervical lateral mass
Registration, see Image-guided spine fixation 166
surgery
Stealthstation, spinal stereotaxis 73
Relative value system 306

Stem cell
Relative value unit 307309

advantages in osteogenesis promotion
Resource-based relative value system 302,

14
305309 gene therapy 22, 23
Subject Index 320

dramroo
mesenchymal stem cell bone formation
surgery
in dog models 15
adolescent idiopathic scoliosis
SUMMIT, posterior cervical lateral mass

fixation 167, 171173

adolescent idiopathic scoliosis
232, 233
Thoracic laminectomy, disadvantages

anterior derotation and fusion 230,
266
231
Thoracic microendoscopic discectomy

crankshaft phenomenon 226, 227
advantages 268, 270, 274

curve classification 226
herniation incidence 266, 291, 292

degenerative adult-onset scoliosis
indications 267269
233235
outcomes 273
double major curves 231, 232
pain relief 269

double thoracic curves 231
patient positioning and operative setup

indications 225
269, 270
pedicle screws 227, 228
posterolateral approach 269

posterior fusion 228230
thoracoscopy development 267

preoperative planning 225, 226
Thoracic pedicle screws

structural right thoracic curve
anatomy 98, 99
228231
assessment of placement 98
history 225
cancellous feel in placement 99

kyphosis 235237
evaluation of pedicle tract 101, 104

morbidity 237, 238
free-hand technique
treatment, nonoperative
exposure 190
outcomes 199, 201

overview of steps 190192

adolescent idiopathic scoliosis 220,
pedicle gearshift-lateral 196

221
pedicle palpation 196

degenerative adult-onset scoliosis
pedicle tapping 197

221, 222
preoperative assessment 188, 189

history 213
probe 193
kyphosis 222, 223
screw placement and confirmation

Transoral decompression, image guidance
197, 198
for cervical spine 121
starting point and cortical burr 190,

Tricalcium phosphate
193, 196
bone repair 42, 43
training 202
Vitoss
navigation techniques 96, 98101

bone repair 43
preoperative planning 96
porosity 43
successful screw placement definition
98, 104
VERTEX
T1 pedicle screw placement 162

adolescent idiopathic scoliosis surgery
Thoracolumbar deformity
229
evaluation
posterior cervical lateral mass fixation
curve measurement 219

166, 167, 171173
history 214
screw placement 168170
physical examination 214216

Vertical compression fracture
radiography 216218
epidemiology 240
Subject Index 321

dramroo
Vertical compression fracture (continued) Brain Lab system 7981
morbidity 240 C-arm 76, 77, 85, 86, 94
nonoperative management 240, 241 computed tomography registration 78,
prevention 248, 249 79, 108, 111
surgical management contraindications 78
decompression 241 disadvantages 92, 93
kyphoplasty dynamic reference array 86, 88, 89, 93
outcomes 246248 equipment 76, 77
overview 241, 242, 244, 245 Fluoronav system 76, 77, 91, 92, 206
percutaneous vertebroplasty image quality 87, 88
comparison 247 instrument projection on screen 78
technique 245, 246 limitations 92, 93
percutaneous vertebroplasty pedicle screw misplacement rates 91, 92,
complications 243, 244 94
contraindications 243 prospects 93, 94
development 242 signal transmission 77
kyphoplasty comparison 247 software 79, 80
outcomes 243, 248 steps in process 86, 87
overview 241 tracking devices 86
pain relief mechanism 242, 243 Vitoss
technique 242, 243 bone repair 43
prospects 248, 249 porosity 43
Virtual fluoroscopy
accuracy validation 87 X-ray
advantages 76, 81, 85 lumbar interbody fusion assessment
applications dynamic plain radiographs 57, 58,
anterior-cervical spine 89 68
overview 85, 88 plain radiographs 56, 57
percutaneous pedicle screw placement thoracic pedicle screw placement 189
88, 89, 205, 206, 212 thoracolumbar deformity 216218
posterior spinal procedures 88 unreliability of intraoperative
thoracic pedicle screw placement 100 radiography 71, 72
Subject Index 322

dramroo

Advances in Spinal Stabilization PDF

Uploaded by

Copyright:

Available Formats

Advances in Spinal Stabilization PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Advances in Spinal Stabilization PDF

Uploaded by

Copyright:

Available Formats

dramroo

Advances in Spinal Stabilization

L. Dade Lunsford Pittsburgh, Pa.

Regis W. Haid, Jr. Atlanta, Ga.

145 figures, 2 in color, and 17 tables, 2003

Basel Freiburg Paris London New York

Library of Congress Cataloging-in-Publication Data

Includes bibliographical references and index.

ISBN 3805574312 (hard cover : alk. paper)

Gerald E. IV. Series.

[DNLM: 1. Spinesurgery. 2. Orthopedic Proceduresmethods. 3. Spinal

Fusionmethods. WE 725 A2447 2003]

RD533 A285 2003

Copyright 2003 by S. Karger AG, P.O. Box, CH4009 Basel (Switzerland)

VIII Series Editors Note

1 Bone Morphogenetic Protein (rhBMP-2): Experimental Review

71 Overview of Spinal Navigation

Instrumentation and Technique Update

128 C1 Lateral Mass Fixation

Minimally Invasive Update

251 Endoscopic Posterior Cervical Foraminotomy

290 Epidemiology and Variations in Care of Spine Disease

314 Author Index

315 Subject Index

Haid RW Jr, Subach BR, Rodts GE Jr (eds): Advances in Spinal Stabilization.

Prog Neurol Surg. Basel, Karger, 2003, vol 16, pp 113

Bone Morphogenetic Protein

(rhBMP-2): Experimental Review

and Clinical Update

Brian R. Subach, Regis W. Haid, Jr., Gerald E. Rodts, Jr.,

Emory University, Atlanta, Ga., USA

The Basics of BMP Biology

Each of the six known osteoinductive BMPs shares significant similarities

An Osteoinductive Role for BMP

In order to function as a suitable graft for bridging bone defects or fusing

From the Laboratory

Early work in the isolation of proteins with osteoinductive activity suggested

autogenous and allogeneic bone grafting, investigators began applying BMP

Posterolateral fusion models attempted to replace autograft with a BMP/

criteria in 11 patients undergoing anterior lumbar interbody fusion procedures

Fig. 4. Computed tomographic reconstructions of a human patient receiving a

The widespread use of spinal fusion procedures in the management of

Brian R. Subach, MD, Neurosurgery

550 Peachtree Street/ Suite 806, Atlanta, GA 30308-2247 (USA)

Tel. 1 404 686 8101, Fax 1 404 686 4805

Prog Neurol Surg. Basel, Karger, 2003, vol 16, pp 1428

Cellular and Genetic Approaches for

A variety of evolving technologies are currently being evaluated in pre

Stem Cell Technologies

The direct application of osteoinductive stem cells on a bioresorbable

Bone Morphogenetic Protein Gene Therapy

Bone Morphogenetic Protein Overview

Cellular and Genetic Approaches 15

Gene Therapy Overview

Cellular and Genetic Approaches 17

Direct BMP Gene Therapy

Cellular and Genetic Approaches 19

Ex vivo BMP Gene Therapy

Fig. 2. Three-dimensional CT recon