Reconstruction of Upper Cervical Spine and Craniovertebral Junction

Reconstruction of Upper Cervical Spine and Craniovertebral Junction

Petr Suchomel  •  Ondrˇej Choutka

Reconstruction of Upper Cervical Spine and Craniovertebral Junction With contributions by Jan Hradil Lubomir Jurák Radim Brabec Pavel Buchvald Pavel Barsa Vladimír Beneš Radek Frič

Authors Petr Suchomel Regional Hospital Liberec Department of Neurosurgery, Neurocenter Husova St. 10 460 63 Liberec Czech Republic [email protected]

ISBN  978-3-642-13157-8

Ondrˇej Choutka University of Cincinnati Medical Center Dept. Neurosurgery Albert Sabin Way 231 45267-0515 Cincinnati Ohio USA [email protected]

e-ISBN  978-3-642-13158-5

DOI  10.1007/978-3-642-13158-5 Springer Heidelberg Dordrecht London New York Library of Congress Control Number: 2010937980 © Springer-Verlag Berlin Heidelberg 2011 This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Product liability: The publishers cannot guarantee the accuracy of any information about dosage and application contained in this book. In every individual case the user must check such information by consulting the relevant literature. Cover design: eStudioCalamar, Figueres/Berlin Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

To my wife Jana. She has unconditionally supported my efforts for 25 years and without her patience and love, this book would never have been finished. Petr Suchomel To my wife Petra and my boys Honzik and Lukas. Your eternal patience, support, tolerance, and love made this endeavor possible. Ondrˇej Choutka



v

Preface

The spine is a “scaffold” for the erect human body and spinal cord that allows for information to travel between the central nervous system and the peripheral movement executers. Without good signal transition or intact scaffold, a human being cannot walk efficiently. Last century has seen a growing interest and need by many surgeons to strengthen collapsing scaffold and to improve relay of neural signals along the spinal cord. Craniovertebral junction represents the ultimate link between the head and spine with its absolute need for structural support as well as mobility. Historically, orthopedic surgeons and neurosurgeons became intimately involved in the care of the spinal patient, rarely working together. One was more interested in the strength and shape of the scaffold; the other was more concerned about the quality of information passing through the spinal cord and assuring it remained free from compression. The two differing approaches resulted in two schools of spinal practice: one perfecting reconstructive and fusion techniques, the other mastering microsurgical decompressive aspects of spinal care. Both sides failed to realize that for a patient to enjoy a functional, ambulatory life, they are both necessary. The multilevel decompressive procedure that potentially results in spinal instability may require good structural support with anatomical alignment. The era of admiration of beautiful constructs without respect for neural structures or microsurgical decompression without the thought for good structural support is over. Spine surgery has undergone tremendous development in last 30 years allowing surgeons to operate safely and effectively in previously forbidden or dangerous areas. Development of imaging modalities, surgical instruments, implants, intraoperative monitoring, and anesthetic techniques allowed for spinal techniques to flourish with improved safety and ambulating patients! The new generation spine surgeon is here to stay and rid us off the artificial separation between structure and nervous system. Our daily work clearly demonstrates that there is a whole array of common spinal problems treated frequently. On the other hand, there are certain, more complex diagnoses even in spinal care that require special expertise, skills, and equipment. There are still some super specialized topics which, in our opinion will remain under the wings of original specialties. It is the orthopedic correction of thoracolumbar deformities namely those congenital and neurosurgical microsurgery of spinal cord pathologies. All the other surgically treatable diseases would encompass the “general spine surgery.” Spinal trauma, degenerative disorders, tumors, and inflammatory diseases all need fully devoted people able to be at service in a 24 h regime. This book, based on our own experience with nearly 300 upper cervical spine reconstruction surgeries, should serve to all those who would not only like to begin with surgery in this region but also to those who are already involved, offering them

vii

viii

Preface

a summarized information about the current possibilities of upper cervical spine reconstruction and a step by step guide of modern potential treatment options for disorders in the CVJ. This book would not be complete without the beautiful illustrations of Petr Polda and radiographic contributions by Dr. Ladislav Endrych, Chairman of the Radiology Department in Regional Hospital Liberec. Last but not least, our thanks goes to Drs. Jan Hradil, Vladimir Benes, Pavel Buchvald, Radek Fricˇ (currently Rikshospitalet Oslo), Pavel Barsa, Robert Frohlich, Lubomir Jurak, Miroslav Kaiser, and Radim Brabec for their significant contributions to this book. Their relentlessness reflects the team spirit of the Neurosurgery Department in Liberec, Czech Republic. Liberec, Czech Republic Cincinatti, Ohio, USA

Petr Suchomel Ondrˇej Choutka

Contents

Section I  Anatomy, Biomechanics and Radiology   1 Surgical Anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P. Suchomel, O. Choutka, and P. Barsa

3

  2 Biomechanical Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P. Suchomel and P. Buchvald

17

  3 Special Radiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . O. Choutka and P. Suchomel

23

Section II  Principles of Reconstruction Techniques 4 Surgical Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P. Suchomel, J. Hradil, and R. Fricˇ

39

5

55

Basic Principles of Reconstruction Techniques . . . . . . . . . . . . . . . . . . . O. Choutka and P. Suchomel

6 Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction . . . . . . . . . . . . . . . . . . . . . . . . . . . . P. Suchomel and O. Choutka

65

  7 Virtual and Real Time Navigational Techniques . . . . . . . . . . . . . . . . . . 125 P. Suchomel and O. Choutka

Section III  Indications for Surgery and Examples of Reconstruction   8 Traumatic Atlantooccipital Dislocation . . . . . . . . . . . . . . . . . . . . . . . . . 139 P. Suchomel and V. Beneš   9 Occipital Condyle Fractures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 P. Suchomel and L. Jurák



ix

x

10 Atlas Fractures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 P. Suchomel and R. Brabec 11 Odontoid Process Fractures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 P. Suchomel and L. Jurák 12 Fractures of the Ring of Axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 P. Suchomel and J. Hradil 13 Miscellaneous C2 Fractures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 P. Suchomel and J. Hradil 14 Multiple Fractures of Axis and Atlas-Axis Fracture Combinations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 P. Suchomel and J. Hradil 15 Acute Traumatic Atlantoaxial Dislocation in Adults . . . . . . . . . . . . . . 215 P. Suchomel and R. Fricˇ 16 Posttraumatic Deformity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 P. Suchomel and R. Fricˇ 17 Non Specific Inflammation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 P. Suchomel and O. Choutka 18 Rheumatoid Arthritis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 P. Suchomel, P. Buchvald, and O. Choutka 19 Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 P. Suchomel, V. Benes, and M. Kaiser 20 Congenital and Developmental Abnormalities . . . . . . . . . . . . . . . . . . . 285 P. Suchomel and O. Choutka 21 Degenerative Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 P. Suchomel and P. Barsa 22 Surgical Failures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 P. Suchomel and O. Choutka Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

Contents

Abbreviations

2D 3D AA AAD AADI AAI AAOA AARF ABC ACDF ALL AO AOI AOD AP AREZ ASA AT ATB BAI BDI BMP CCI CAD CCJ CEP CMA CN CTA CTA 3D CVJ DRA DREZ EEA EOP EP ES

Two dimensional Three dimensional Atlantoaxial Atlantoaxial dislocation Anterior atlantodental interval Atlantoaxial instability Atlantoaxial osteoarthritis Atlantoaxial rotatory fixation Aneurysmal bone cyst Anterior cervical discectomy and fusion Anterior longitudinal ligament Atlantooccipital Atlantooccipital interval Atlantooccipital dislocation Anteroposterior Anterior root exit zone Anterior spinal artery Anterior translation Antibiotic Basion-posterior axial line interval Basion-dental interval Bone morphogenetic protein C1-condyle interval Computer aided design Craniocervical junction Condylar entry point Cervicomedullary angle Cranial nerve CT angiography Spatial, 3 dimensional CT angiography Craniovertebral junction Dynamic reference array Dorsal root entry zone Expanded endonasal approach External occipital protuberance Evoked potentials Ewing sarcoma xi

xii

ETO FOM FT GCT IAAD IAR ICA INL IOM LCH LTA IOP MDCT MRA MRI MVA NPV NSAID OAA OC OCF OF OS PADI PLL PICA PMA PSA RA SAC SAS SCM SNL SOMI TAL TBC TBI TO UCS VA VAAII

Abbreviations

Endoscopic transcervical odontoidectomy Foramen occipitale magnum Foramen transversarium, transverse foramen Giant cell tumor Irreducible atlantoaxial dislocation Instantaneous axis of rotation Internal carotid artery Inferior nuchal line Intraoperative electrophysiological monitoring Langerhans cells histiocytosis Ligamental tubercle avulsion Internal occipital protuberance Multi-detector row CT Magnetic resonance angiography Magnetic resonance imaging Motor vehicle accidents Negative predictive value Nonsteroidal antiinflammatory drugs Occipito-atlanto-axial Occipitocervical Occipital condyle fracture Odontoid fracture Osteogenic sarcoma Posterior atlantodental interval Posterior longitudinal ligament Posterior inferior cerebellar artery Posterior meningeal artery Posterior spinal artery Rheumatoid arthritis Space available for the spinal cord Space available for screw Sternocleidomastoid muscle Superior nuchal line Sternal occipital mandibular immobilizer Transverse atlantal ligament Tuberculosis Traumatic brain injury Transoral Upper cervical spine Vertebral artery Vertical atlantoaxial instability index

Section Anatomy, Biomechanics and Radiology

I

1

Surgical Anatomy P. Suchomel, O. Choutka, and P. Barsa

The goal of surgical anatomy is to avoid the descriptive aspect of “pure form.” On the other hand, it emphasizes important structures with respect to the pathological condition and surgical approach. In descriptive anatomy of bony structures, one has to realize what the origin of its data is. Obviously, there are differences owing to gender variations (e.g., lower values in females); however, other factors also can influence anatomical variations, such as race or age. For example, data from Asian population show lower values in general relative to their population height. Old anatomical data can show slightly lower values due to a change in the average population height over a longer time period. The other differences can arise from the study design. CT measurements are frequently performed in young individuals due to traumatic injuries in another spinal region, whereas cadaveric data are often obtained from old or diseased people with possibly smaller vertebral sizes. In general, the exact descriptive anatomical data can only be used to give the proportional anatomical relationships. The absolute values have to be used cautiously and cannot be blindly applied to the individual patient. Pure anatomical knowledge has to be supported by the exact imaging and precise measurements of each individual patient. Nowadays, bony structures can be

clearly visualized by CT with 3D reconstructions, the status of soft tissue (spinal cord, disks, and ligaments) by MRI, and vascular structures by CTA and/or MRA. Plain films, although a good initial screening tool, are, in general, less helpful in surgical planning. Modern surgical anatomy should be comprehensive but practical so that readers can follow the guidelines and confirm the data in their daily experience. In this chapter, we are provide such a guide through the surgical anatomy the craniovertebral junction.

1.1 Bony Structures 1.1.1 Occipital Bone (C0) Understanding occipital bone anatomy is important as the posterior squamous part is often used as cranial anchor in occipito-cervical constructs. Foramen magnum is the exit foramen of the skull and is frequently involved in surgical procedures. Occipital condyles are unique joint projections connecting the spine to the skull. Anteriorly, the clivus is a structure also frequently involved in decompressive or reconstructive procedures.

1.1.1.1 Occipital Squama P. Suchomel (), P. Barsa Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic O. Choutka Department of Neurosurgery, University of Cincinnati College of Medicine, 231 Albert Sabin Way, Cincinnati, OH 45267-0515, USA

This part of occipital bone creates an externally convex surface directly visible during most dorsal approaches to CVJ. From surgical viewpoint, it is not just the bone thickness that is relevant, it is the relationship of exterior landmarks to the underlying intracranial venous sinuses and neural tissue that is of utmost importance when it comes to occipital bone.

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_1, © Springer-Verlag Berlin Heidelberg 2011

3

4

There are only a few external landmarks visible during surgery – superior nuchal lines (SNL) and inferior nuchal lines (INL), external occipital protuberance (EOP, Inion), external midline occipital crest, and the edge of foramen magnum. There is great variability in the position of the superior nuchal line and therefore, it does not reflect the internal position of transverse sinus accurately. The relation of confluence of sinuses (torcular Herophili) to EOP is more consistent [47]. According to work of Nadim et al. [39], the safe zones where the injury of transverse sinuses and torcular can be avoided are located more than 2 cm caudal from EOP and SNL. The bone thickness is greatest at the EOP and decreases radially [11, 61]. Most authors describe bone thickness in the region of EOP round 15 mm in males and 12 mm in female Caucasian population but approximately 6 mm or less is available over the cerebellar hemispheres [11, 14, 42]. The safe zone for an 8 mm occipital squama screw insertion covers an area 2 cm laterally from EOP and narrows down inferiorly. So, at 1 cm below EOP, the safe zone is only within 1cm of the midline, and at 2 cm below the EOP, it falls to 0.5 cm of the midline (Fig. 1.1). The thinnest bone (sometimes less than 1 mm) was measured laterally from midline between INL and foramen magnum [47]. The outer cortex contributes 45% to bone thickness whereas the inner table only 10% [61]. Practical conclusion: The thickest bone is available around the SNL and EOP and then along the midline occipital crest, but one has to be aware of injury of intracranial venous sinuses in this region. The exact preoperative bone thickness as well as localization of principal venous sinuses should be determined on

1  Surgical Anatomy

preoperative CT in each individual patient planned for occipital bone fixation.

1.1.1.2 Occipital Condyles There is a great variability in the shape and size of occipital condyles. Most often, the occipital condyles are kidney shaped, biconvex, and medially oriented bone structures localized in the anterior half of foramen magnum. The mean distance between both condyles is 41.6 mm posteriorly and 21 mm anteriorly [38]. The mean condyle length is 23.6 mm (15–30.6), the width 10.5  mm (6.5–15.8), and the height 9.2  mm (5.8–18.2) [28, 29, 36, 38]. Both condyles form an angle of 50°–60° in transverse plane and 124°–127° (male-female) in frontal plane (atlanto-occipital joint angle) [29]. The single condyle axis angle to midline is in average 30° but can vary 10°–54° in adults [36, 38]. Hypoglossal nerve canal is passing transversally through the bone just above the base of the condyle antero-laterally in an axial angle of 45°. It is directed slightly superior with the mean distance of 11.5  mm between the hypoglossal foramen and the inferior border of the condyle [36]. The canal itself is 6.2  mm long, ovoid in shape with 4 mm internal diameter [28]. Jugular foramen with its important contents (jugular vein, n IX,X,XI.) is located 12–25 mm antero-laterally from the condyle. Condylar emissary vein can be identified in the dorsal superolateral condyle border during dissection.

Protuberantia occipitalis externa Linea nuchalis superior Linea nuchalis inferior

Foramen magnum Condylus occipitalis

Fig. 1.1  Schematic drawing of the occipital bone external surface with depicted areas for safe occipital squama screw purchase

1.1  Bony Structures

Carotid artery is often located more than 5 mm anterolateral to the anterior condylar cortex [28]. The condyles are separated by intra-occipital synchondrosis in two (anterior, posterior) parts until the age of six but it can sometimes persist bifacet into adulthood. Third occipital condyle (condylus tertius) is an ossified remnant of the hypochondral bow of the fourth sclerotome (proatlas) at the distal end of clivus that can occasionally be seen as an individual or multiple ossicles directly above and anterior to arch of C1. Practical conclusion: The occipital condyle is normally twice as long (approx. 20  mm) as it is wide (approx. 10  mm), medially oriented structure (20°– 30°). Because of its variability, only CT scan can depict its exact shape, orientation, mass, and relationship to neighboring structures in each individual case.

1.1.1.3 Clivus The upper part of clivus belongs to sphenoid bone whereas the lower part to basilar portion of occipital bone. These two parts are separated by spheno-occipital synchondrosis till the age of 16.5 (13–18) in males and 14.4 (12–15) in females [50]. The suture allows for growth and correct formation of the skull. In normal adults, the length of the entire clivus is 4.5  cm (3.7–5.2), with the basilar portion of occipital bone representing 3.1 cm (SD 0.3). In occipital hypoplasia, the basilar portion could be as short as 1.7  cm [29]. The thicker part of clivus is located anterosuperiorly and contains cancellous bone. The thinner part, formed by cortical bone only, is in the region of foramen magnum (FOM). Usually, the outer cortex is more solid and thicker than the inner one [29]. Practical conclusion: The clivus can vary in shape and size especially in developmental anomalies. As a part of bone firmly connected to skull, it can be navigated using the skull data.

1.1.2 Atlas (C1) The atlas is a ring-shaped unique vertebra having no vertebral body and no intervertebral disk attached. It consists of two lateral masses connected with short anterior and longer posterior arches (Fig. 1.2). Its anatomic integrity is crucial for stability of the CVJ and movement of the head.

5

In older European anatomical studies of atlas, the average outer distance between anterior and posterior tubercle (length) was 46.3 mm in males and 43.2 mm in females. The external transverse diameter (width) was 83  mm and 72  mm, respectively [6]. Exact measurements performed later by Doherty [7] on 80 European C1 specimens, generally confirmed the old anatomical data. An average atlas outer length was 45.8 mm (SD = 2.9) and outer width 78.6 mm (SD = 8.1). The internal width was 32.2 mm (SD = 2.3) and internal length was 31.7 mm (SD = 2.2). Similar values were obtained by Kandziora from 50 dry specimens [25] and also when later measured electronically by Rocha in 20 cadaveric bones [48]. Atlas is the vertebra with the widest internal diameter. The internal length and width were 32.6  mm (range 29.6–36.4  mm) and 29.7 (25.7– 32.2  mm), respectively [48]. The transverse ligament tubercles serve as an attachment place of the ligament and are located internally on the medial wall of the lateral masses. Rocha measured the internal “intertubercle” distance to be 22.9 mm (18.7–27.9 mm) [48]. The largest published anatomical series was done by Christensen et al. [4] who measured 120 dried atlases from a defined American population (average age 52.9 years, average height 169.7 cm). Electronic caliper was used with the average outer width being 75.61  mm (SD = 5.94) and outer length 45.67 mm (SD = 3.61). Anterior arch is a strong structure that harbors anterior tubercle in the midline. The anterior tubercle is regularly visible on plain lateral radiographs and often serves as anatomical and radiographic landmark during surgical procedures especially during instrumentation. The internal wall of the anterior arch is in contact with odontoid process forming a facet (fovea dentis). The height of the anterior ring is 15.4 mm (SD = 3.2) [7], the length is 30 mm [25, 29], and the thickness is 6 mm in the midline [4]. This thickest cortical bone of the whole C1 is in agreement with its biomechanical load demands. Posterior arch is longer (usually 2/5 of C1 circumference) and weaker because of the bony groove for horizontal segment of the vertebral artery. The posterior arch has the thinnest cortex of the entire vertebra [7]. The posterior midline arch height is 9.58  mm (SD = 2.26) and its thickness is 7.82 mm (SD = 2.64) [4], in the area of vertebral artery(VA) groove, the height is a mere 4.5 mm (4.3–6.1) [48]. The distance between the posterior midline tubercle and the most medial aspect of the VA groove is approximately 15 mm [37, 48]. Lateral masses are in fact the most voluminous bony parts of atlas forming four facet joints. The superior

6

1  Surgical Anatomy

Fig. 1.2  Artistic drawing of C1 with marked parameters simplified for practical purpose

joints differ from the inferior ones in size and shape. Its average midportion (central) length is 16.82 mm (SD 1.0), width 16.06 mm (SD 0.91), and height 15.68 mm (SD 0.98) [51]. The lateral mass cones in medially in the coronal plane and wedges posteriorly in the sagittal plane. Medial wall height was measured 11 mm (SD = 1.21) and lateral 22  mm (SD = 1.89) [25]. Its mean height anteriorly is 18.5 mm (SD = 2.39 mm) and posteriorly 10.2 mm (SD = 2.0) [1]. The superior articular surface (fovea articularis superior) is a medially tilted, concave ellipsoid (kidney shaped), with an approximate length of 20 mm and width of 10 mm [4, 29, 56]. This naturally corresponds to the shape and size of the articular surface of the c­ ondyle. In coronal plane, it overlaps the smaller lower facet. Sometimes it can be developmentally divided into two contact surfaces. Superior facet angle to ­midline was measured in horizontal plane 22.4° (SD = 1.52) [25].

The lower articular surface (fovea articularis inferior) is less concave, round shaped, and smaller than the upper one with a common length of 17 mm (14– 23 mm) and width 17 mm (14–23 mm) [1, 13, 29, 32]. The pillar between lower facet and posterior arch is often used for screw anchorage. The height of this “working window” is 3.6  mm (2.13–4.09  mm) and width 9.5 mm (6.98–13.34 mm) [15]. Other studies on non Indian population are suggestive of a larger area with the mean height of 4.5 mm, ranging from 4.1 to 6.1 mm [48, 51]. Normally, the transverse process forms a transverse foramen (FT) for vertebral artery (VA). It has a variable diameter and position and can be opened in an anterolateral direction. Relatively often (15.6%), there is a partial or total covering (arcuate foramen) of artery (Fig. 1.3) in the C1 groove creating a so called “ponticulus posticus” [20, 30, 60]. This can be very

7

1.1  Bony Structures

1.1.3 Axis (Epistropheus, C2)

Fig. 1.3  Arcuate foramen of C1 and high riding VA in C2 pars. Sagittal CT reconstruction. Note the broken transarticular screw (case referred for revision from another department)

important during sub-periosteal exposure of C1 lamina and during C1 lateral mass screw placement. One can suppose that the lamina is too broad and the VA could thus be injured through a poor planning of the entry point. The distance of medial FT border to midline is approximately 25 mm [19, 25, 54]. For practical surgical purposes we can summarize: In average the atlas outer width is approximately 75 mm and outer length 45 mm. Internal AP diameter is usually 30 mm. Anterior arch is the strongest bony structure approximately 30 mm long, 15 mm high, and 6 mm thick. Posterior arch is the weakest part of atlantal ring having a medial height of approximately 10 mm but in the region of VA groove only 5 mm or less. The lateral mass is wedge shaped in sagittal and frontal planes. Superior facet is approximately 20 mm long and 10 mm wide with the axis angled in at about 20°–30° medially. Lower facet is round shaped with a 17 mm diameter. The lateral mass pillar below the posterior arch has a rectangular shape with 4.5 mm height and 10  mm width forming the working window for eventual screw introduction. The VA containing transverse foramina are approximately 25 mm from midline but the VA containing groove comes as close as 15 mm to midline.

Second cervical vertebra is also a unique spinal structure. It is composed of vertebral body with upward directed odontoid process, which articulates with the posterior aspect of anterior C1 arch (Fig.  1.4). The body is connected to the lateral mass by short and strong pedicles. Lateral mass pillar between upper and lower articular processes is pars interarticularis and its narrowest part is the isthmus. Transverse foramen can vary in shape and size. Posterior arch is similar to other subaxial cervical arches but is larger, in general. The spinous process is often bifid. The cortical bone covering the anterior aspect of odontoid process, its tip, and the anterior vertebral body surface is extremely thick, especially in the area of an anterior midline ridge named “promontory“ by Doherty [17]. The thickness here is approximately 1.7 mm whereas the lateral and posterior parts of odontoid and body are covered by only a 1.0 mm or less thick cortical bone. The inside dens trabecular bony architecture is organized to resist the antero-posterior and lateral forces. Strong trabeculae that span fanlike from the upper facets to the inferior bony endplates assist to bear and transmit the axial load. The weakest bone density is in the body below the base of odontoid [17]. The external axis width is 56  mm (48–69  mm) [25] and length approximately 55 mm. Vertebral body is connected to C3 by intervertebral disk. The C2 endplate shape is sagittally concave with a prominent anterior edge. The distance between this edge and the base of dens is on an average 22 mm (17– 31 mm) [25]. The caudal body width is 18–19 mm and AP diameter 15–17 mm [25, 27, 59]. Both parameters decrease superiorly. Consequently, the internal spinal canal AP diameter is smaller at the C2 body base (14.8 mm) than at the level of odontoid process attachment 17.35 mm [59]. The average canal width, measured as 21.6 mm [59], does not change throughout the height of the vertebra. The odontoid process diameter is usually smaller at the base (waist) than in the middle of its shaft. The process is composed of thick cortical bone surrounding internal cancellous component with an internal diameter reaching 4.3–6.2  mm [18]. It is usually 20 mm (15–25.4 mm) long and posteriorly tilted in an angle of 64° in respect the endplate [25, 59]. Anterior surface forms an articulation with the atlas. Basal (waist) odontoid diameter is approximately 9  mm

8

1  Surgical Anatomy

Fig. 1.4  Artistic drawing of C2 with marked parameters simplified for practical purpose

(7.8–14.1  mm) and the maximal one 11  mm (8.4– 14.1 mm) [25, 59]. The pedicles of C2 vertebra, despite being very short, are the strongest and widest pedicles in the cervical spine connecting the body to lateral masses nearly in right angle to sagittal plane. From a surgical viewpoint, the most critical value is the mean transverse pedicle diameter at the level of the VA. It was measured 6.4 mm (2.09–13.2 mm) [35]. Lateral masses form an oblique column between upper and lower facets. This part of bone is very important for possible screw purchase is usually called pars interarticularis or simply “pars” in the anatomical literature. This pillar is more or less thinned by the groove of VA in its narrowest point called the isthmus. Despite the above nomenclature corresponding exactly to that of the other spine regions, one has to be careful when interpreting publications even from wellaccepted authors [11, 26, 31, 33, 34, 46, 53, 59], who

commonly refer to the pars interarticularis as the C2 pedicle. The pars is inclined 35.2° (29°–41°) medially and 38.8° (22°–52°) rostro-caudally [21]. The width and height of the isthmus at the level of the transverse foramen is 7.9–8.6  mm (female vs. male) and 6.9– 7.7  mm (female vs. male), respectively [59]. It can, however, vary to values less than 3.5 mm at least on one side in approximately 18–23% of patients [23, 45, 46]. The upper facet is slightly convex and faces upward and outward with the shape and size corresponding to the inferior articular process of C1. The outward angle in coronal plane is approximately 24° [25]. Its length and width are similar, approximately 17 mm [25], depending on gender and body size. The articular surface atypically arises directly from the C2 pedicle laterally. The lower facet forms a typical, forward-facing subaxial spine joint surface to articulate with C3. C2 arch is the strongest arch in cervical spine usually containing enough cancellous bone to accommodate a

9

1.2  Ligaments and Joints

3.5 mm laminar screw. The mean laminar thickness was measured 5.77  mm (1.35–9.77  mm) [3]. On 37 adult specimens, Xu et  al. [58] measured the C2 laminar height to be 11.2  mm (SD = 1.1  mm), its half length 15.6 mm (SD = 1.2 mm), and average thickness 4.3 mm (SD = 0.9 mm). Both the laminas formed an angle of 99.1° (SD = 8.0°) called laminar width, which is the narrowest in the spine. The downslope laminar angle was determined as 111.7° (SD = 9.3°). The spinous process is likewise a strong structure serving as an attachment point of important suboccipital triangle muscles and the nuchal ligament. Transverse foramen incorporating VA is of utmost surgical importance. Usually, the VA enters the C2 transverse foramen vertically approximately 15  mm from the midline, passes cranially, and then courses 45° laterally to form an upward loop around the transverse process to reach the vertically oriented C1 transverse foramen. In about 80% of population, the VA bends sharply outward inside the C2 VA groove, leaving enough bone of the isthmus for transisthmic or transpedicular screw purchase. The VA curve located directly below the superior articular process of C2 can occasionally be more superior, dorsal, or medial than expected, thus directly influencing the pars and pedicle size. Such “high riding VA” occurs at least unilaterally in up to 23% of patients undergoing craniocervical procedures [34, 40, 45]. Despite this, it is clear that the diameter of the bony VA canal and foramen does not represent the external diameter of the actual artery [2, 35]. The artery is surrounded by venous plexus and connective and periostial tissue, and this fact often allows for certain amount of foraminal breach during screw placement.

For practical surgical purposes we can summarize: On an average, the axis outer width is around 56 mm and outer length 55 mm. Internal AP diameter increases from 15 mm at the level of C2/3 disk to 17 mm at the base of odontoid process. The internal width remains relatively constant measuring approximately 22  mm. The odontoid process has an average diameter of 10 mm and is tilted backwards relatively to the C2 endplate in an approximately 60° angle and about 10° relatively to horizontal plane. The distance from the anterior inferior edge of C2 body to the odontoid tip is approximately 40 mm (shorter in females). The AP C2 body diameter decreases with upward direction. The diameter of the isthmus (the narrowest part of the pars) is approximately 7–8 mm but in 18–23% of patients it can be significantly thinner due to a high riding VA.

1.2 Ligaments and Joints The UCS and CVJ ligamentous connections are very complex (Fig. 1.5) providing one of the most complicated movement patterns in the human body. Atlantooccipital together with atlantoaxial joints are always working together in a synchronized fashion. Upper cervical spine is the most mobile part of the entire vertebral column with a unique anatomical structure. There are no intervertebral disks and yellow ligament. Movement is restricted not only by the bony shape of the vertebra but mostly by the strong ligaments. The axis is firmly connected to the occiput and atlas is quite freely floating in between.

Membrana tectoria (cut off) Canalis hypoglossi Ligamentum apicis dentis Ligamenta alaria Articulatio atlantooccipitalis Ligamentum transversum atlantis (TAL) Ligamentum atlantoaxiale accesorium Articulatio atlantoaxialis

Fig. 1.5  Schematic picture of UCS ligamentous structures (internal posterior view)

Fasciculus longitudinalis ligamenti cruciformis

10

Posterior atlanto-occipital and atlantoaxial membranes are relatively weak structures compared to the interarcual subaxial ligamentum flavum. The anterior longitudinal ligament (ALL) loosely attaches to the vertebral bodies of the subaxial spine. However, it is firmly connected to the disk annulus at each level and finally inserts to the anterior tubercle of atlas. Anterior atlanto-occipital membrane replaces the ALL between atlas and the clivus. The posterior surface of C2 body and odontoid process is covered by tectorial membrane which is, in fact, a strongly developed cranial part of the posterior longitudinal ligament (PLL). Cranially, it is inserted into the clivus with lateral extend to hypoglossal canals. Caudally, the membrane is attached to C2 body continuing into the PLL. The most important structure for atlantoaxial translational stability is the transverse ligament. This ligament is the strongest of the entire complex and attaches to the bony tubercles located on the medial surface of lateral masses of atlas. It is 10 mm high and 2 mm thick with an average length of 23 mm [29, 48]. Together with the longitudinal bundles attached to the posterior aspect of C2 body and anterior edge of foramen magnum, the transverse ligament forms the cruciate (cruciform) ligament. The axis is connected to the occipital bone with three other ligamentous structures. The apical ligament, a possible remnant of chorda dorsalis, connects the tip of the dens to the anterior edge of foramen magnum. This relatively weak band runs forward in a 20° angle and is around 8  mm long and 2–5  mm wide [29, 43, 44]. Symmetrical allar ligaments are extended between the lateral odontoid apex and the medial surface of each occipital condyle. Regularly, these 10  mm-long ligaments also have small insertions into the lateral masses of atlas [9, 10]. Atlantoaxial accessory ligaments found irregularly on both sides are not only connecting the atlas to the axis but also continued cephalically to the occipital bone. The approximate length of this structure is 30 mm and thickness 5 mm [57]. Occasionally, one can also find atlantodental ligament connecting the base of the odontoid process with the anterior arch of atlas [9, 10].

1.2.1 Atlanto-Occipital Joints The two atlanto-occipital joints are true synovial joints similar to the others in UCS. The articulation between

1  Surgical Anatomy

the condyle and the upper C1 articular process allows mainly flexion and extension. The shape, angle, and congruence of joint surfaces are natural restraints of other movement directions. The joints contain synovial membrane and are covered by capsular ligaments.

1.2.2 Atlantoaxial Lateral Joints These two most mobile joints in the entire spine provide predominantly rotational movement; however, movement in other directions and planes is also possible. This is due to the naturally incongruent articular surfaces that do not limit any direction of movement and due to the laxity of restricting ligamentous structures. They consist of encapsulated synovial joint between inferior articular process of C1 and superior process of C2. Their capsular ligaments are reinforced by medial and posterior accessory ligaments.

1.2.3 Atlantodental Joint This synovial joint forms anterior and posterior ­articulation between the odontoid process and anterior arch of C1 and the odontoid process and transverse atlantal ligament, respectively. The transverse ligament is obviously so rigid to keep the odontoid process in contact with anterior arch of C1 under all circumstances. There is only a very limited freedom for lateral movement of the odontoid process. Further, a greater degree of elasticity in childhood allows for greater movement in this joint.

1.3 Muscles of CVJ and UCS Several complex muscular attachments of the upper cervical spine act together to provide three main functions: muscular tension stabilizes the position of head in space; multiple small muscles attached to the skull, C1, and C2 provide movement of the head in all directions; and the massive posterior muscular layer aids in protection of the CVJ from external violence. Good working knowledge of the muscular attachments allows for anatomical dissection during

11

1.3  Muscles of CVJ and UCS

exposures of the CVJ and prevents unnecessary damage to soft tissues. Similarly to the other spine regions, the musculature can be divided in musculi brevii (proprii) connecting one motion segment only and musculi longi bridging two or more segments. In UCS, the short muscles are more important and more specifically developed than in subaxial cervical spine (Fig. 1.6). The nuchal ligament has two portions and knowledge of the presence of fatty areolar tissue between the two leaves of the deeper lamellar portion can prevent blood loss during posterior exposure of cervical spine [24]. The large, posterior superficial muscles of the neck consist of trapezius, semispinalis, sternocleidomastoid, and splenius capitus. They merely cross/attach at the CVJ but are encountered during posterior, posterolateral, and lateral approaches to the region. The deep short muscles are more specific in their structure and function as head extenders, rotators, and lateral

benders. The atlas is connected to the skull through a series of short capitis muscles (posterior rectus capitis minor and superior obliquus capitis). The axis is connected to the atlas by inferior obliquus capitis and to the skull by rectus capitis posterior major. The insertion of this muscle to the spinous process of C2 merges with the insertion of inferior obliquus capitis. These muscles allow mostly for rotation and extension. The anterior muscles of the CVJ include the paired, short rectus capitis anterior that connect the atlas to the clivus. Rectus capitis lateralis runs vertically between the transverse process of C1 and the jugular process of the occipital bone. These two muscles are separated by the ventral ramus of the first cervical nerve. The longus capitis muscle originates on transverse processes of lower cervical vertebrae crosses the CVJ anteriorly to attach to the base of the skull. The function of the anterior muscle group is mostly stabilization of the skull on the vertebral column.

M. obliquus capitis superior M.splenium capitis A. occipitalis N. occipitalis major M. semispinalis capitis M. rectus capitis post. minor M. rectus capitis post. major Membrana atlantooccipitalis post. A. vertebralis N. suboccipitalis Tuberculum post. atlantis Procesus spinosus axis M. obliquus capitis infeerior M. spinalis cervicis M. splenium capitis Proc. articularis inf. axis M. semispinalis capitis

Fig. 1.6  Schematic drawing of anatomical structures of suboccipital triangle

Proc. transversus atlantis M. longissimus cappitis

12

1  Surgical Anatomy

1.4 Vascular Anatomy of CVJ and UCS 1.4.1 Vertebral Artery (VA) Vertebral artery course in the cervical spine can be divided in four segments (V1–4). The first segment represents the course of the artery between its origin on the subclavian artery and its entrance into the transverse foramen of the C6 vertebra (most frequently). The second segment involves the cervical transverse foraminal portion of the VA course (C6 to C1). The horizontal portion of the VA (V3) is from the transverse foramen of the atlas to entrance to the dura. The VA runs in the groove of the C1 lamina, is surrounded by venous plexus, and ultimately passing the posterior wall of the condyle to pierce the atlanto-occipital membrane in its lateral aspect. The intradural course of the VA represents the fourth segment (V4) to terminate in the formation of basilar artery after joining the contralateral VA. The left VA is dominant in 35.8% of patients, hypoplastic in 5.7%, and absent in 1.8%. The right VA is dominant in 23.4% of subjects, hypoplastic in 8.8%, and absent in3.1%. Equivalent right and left VA can be detected in 40.8% of subjects; however, a great diversity exists in the percentual representation of these varieties [55]. The VA course in the region of upper cervical spine is curved and with some redundancy, particularly between C1 and C2 to allow not only for flexion and extension but for rotation so prominent at this spinal segment (Fig. 1.7). The VA redundancy decreases with age [8]. In subjects with a healthy upper cervical spine, the typical five-curve course of VA at the CVJ was seen in 81.8% of CTA evaluations. The rest of the subjects carried various anomalies of the VA course at the CVJ [8]. Surprisingly, in up to 15.6% of patients, one can discover a partial or total bony covering (arcuate foramen) of the horizontal segment of VA, so called “ponticulus posticus” [4, 20]. It is important to be aware of a rather dangerous variable that is, the persistent primitive first cervical intersegmental artery. This aberrant vessel may partially or completely substitute the VA and course below the posterior arch of atlas. Such course would complicate a subarcuate approach to the posterior lateral mass of C1 for screw insertion. In a very large series of 1,013 patients with CT angiography, Hong et al found persistent first

Fig. 1.7  CT angiogram showing the AV redundancy below the C1 entry allowing free C1–2 rotation

intersegmental artery on one side in 3.8% and bilaterally in 0.8 % [20]. Reports of tortuous VA coursing below the posterior arch of atlas without passing through the transverse foramen were also described [22].

1.4.1.1 Branches of VA Certain branches of the VA may have anomalous origins and thus become susceptible to injury during procedures of the CVJ. The posterior inferior cerebellar artery (PICA) usually originates from the fourth segment of the VA intradurally. However, an extradural origin of PICA may be present in 5–20% of people [12]. This makes it a relatively common variation. An extradural origin may be highly variable and PICA can arise close to the entrance of the VA into the dura or as far as atlantoaxial portion of the artery and course below the C1 arch. An extradural origin PICA, usually, does not supply anterior medulla. PICA may originate from other vessels in the region (ascending pharyngeal, ICA etc.) also. Posterior meningeal artery (PMA) should not be confused with an extradural PICA. It usually arises from the extracranial segment of VA and supplies posterior fossa dura and falx cerebelli and cerebri. It originates from the left VA in 17–30% of people and right VA 8–40% [16, 41]; however, just like PICA, it can

13

1.5  Neural Anatomy

originate from other vessels in the area (ascending pharyngeal, ICA, and occipital artery). Posterior spinal artery (PSA) usually originates from the VA, 50% intradurally and 46% extradurally from V3 [52]. However, PSA has also been described to originate from PICA, usually with an extradural origin. Anterior spinal artery (ASA) arises invariably intradurally from the vertebral arteries; however, the relationship of its origin to PICA and vertebrobasilar junction varies. ASA was a direct branch of left VA in 30% of cadaveric specimens, right in 8%, and directly from basilar artery in 2%. The “typical” pattern of dual anterior ventral spinal arteries merging into a single ASA was observed only in 18% of examined brainstems [49].

1.4.2 Internal Carotid Artery (ICA) Although, the ICA is not directly involved in UCS and CVJ, its adjacent position could be of importance in some UCS reconstructive techniques. The lumen of internal carotid artery (ICA) is medial to the transverse foramen of C1 in more than 80% of cases [5] (Fig. 1.8). In such cases, it lies directly in front of C1 lateral masses. With tortuous ICA, the vessel may even be located in front of the C2 vertebral body. Knowledge of ICA variation becomes relevant during direct anterior or anterolateral exposure of CVJ or during posterior reconstruction with instrumentation potentially perforating anterior cortex of vertebrae of UCS and putting the ICA at risk.

Fig. 1.8  Axial CT with contrast media application depicting the normal position of carotid artery in front of the atlas

1.5 Neural Anatomy 1.5.1 Spinal Cord Neural structures are occupying funnel-like cavity of craniocervical junction. The medulla oblongata merges into the spinal cord at the CVJ. The upper limit of spinal cord is defined by anatomists as an exit point of the uppermost root fibers of C1 or the lower end of pyramidal tract decussation. The morphology of spinal cord changes at different levels. There is significant individual variation in size. Nonetheless, it is flattened in anteroposterior direction and usually has a larger transverse diameter. Its surface is divided by the longitudinal fissure and several sulci. The anteromedial fissure and posteromedial sulcus divide spinal cord sagittally into symmetrical halves. The central canal originating from the fourth ventricle passes in the midline and is surrounded by an inner butterfly-shaped gray matter. The gray matter consists of cell columns that extend in posterolateral directions almost to the surface (the posterior horns) and anterolaterally, not reaching the anterior surface of the cord (the anterior horns). Posterior horns contain somatosensory neurons while anterior horns somatomotor neurons. A gray commissure connects the gray substances encircling the central canal. The white matter comprises ascending and descending fibers organized into distinct tracts. Anatomically, it is divided into three columns symmetrically in both halves of the cord: posterior, lateral, and anterior. The posterior column is ascending one localized between the posterior horns of the gray matter. Medially, it is symmetrically divided by the posteromedial sulcus that cranially extends to the caudal cusp of the fourth ventricle in the brain stem. Lateral column is located between anterior and lateral root entry zones and consists of the lateral corticospinal tract intermediating voluntary discrete and skillful motor function and the lateral spinothalamic tract transmitting painful and thermal sense from contralateral side. The anterior columns lie between the anterior entry zones and are symmetrically divided by the anterior spinal fissure. Its most important structure is the descending corticospinal tract concerned with fine motor skills. Of descending corticospinal axons localized in the anterior columns, 75–90% decussates, forming the crossed lateral corticospinal tract and anterior corticospinal tract involving uncrossed fibers.

14

1.5.2 Cervical Spine Nerves Spinal nerves arise from anterior and posterior root filaments. Ventral root filaments exit the anterolateral aspect of the cord in the anterolateral sulcus, in the region termed the anterior root exit zone (AREZ) and are purely motor. Posterior rootlets enter the spinal cord in dorsal root entry zone (DREZ), the region along the posterolateral sulcus and are sensitive ones. The rootlets pass obliquely and laterocaudally within the canal of craniocervical junction entering the root sleeve where the sensory and motor filaments are separated by the interradicular septum, a lateral extension of dura. The dorsal rootless present an oval bulge, the ganglion as it approaches or enters the intervertebral foramen. Distally to the ganglion, the dorsal and ventral roots combine to form a spinal nerve. The cervical nerve root occupies approximately one third of the foraminal section area, usually its inferior aspect. The residual foraminal space is filled with fat and associated veins. The first spinal cervical nerve leaves the canal through the orifice between the occiput and C1. Further cervical nerves exit above correspondingly numbered vertebrae.

References   1. Blagg, S.E., Don, A.S., Robertson, P.A.: Anatomic determination of optimal entry point and direction for C1 lateral mass screw placement. J Spinal Disord Tech 22, 233–239 (2009)   2. Cacciola, F., Phalke, U., Goel, A.: Vertebral artery in relationship to C1-C2 vertebrae: an anatomical study. Neurol India 52, 178–184 (2004)   3. Cassinelli, E.H., Lee, M., Skalak, A., et al.: Anatomic considerations for the placement of C2 laminar screws. Spine (Phila Pa 1976) 31, 2767–2771 (2006)   4. Christensen, D.M., Eastlack, R.K., Lynch, J.J., et  al.: C1 anatomy and dimensions relative to lateral mass screw placement. Spine (Phila Pa 1976) 32, 844–848 (2007)   5. Currier, B.L., Maus, T.P., Eck, J.C., et  al.: Relationship of the internal carotid artery to the anterior aspect of the C1 vertebra: implications for C1-C2 transarticular and C1 ­lateral mass fixation. Spine (Phila Pa 1976) 33, 635–639 (2008)   6. Debreuil-Chambardel, L.: Variations sexuelles de l’Atlas. Bull Soc Anthropologie de Paris 5, 399 (1907)   7. Doherty, B.J., Heggeness, M.H.: The quantitative anatomy of the atlas. Spine (Phila Pa 1976) 19, 2497–2500 (1994)   8. Duan, S., Lv, S., Ye, F., et al.: Imaging anatomy and variation of vertebral artery and bone structure at craniocervical junction. Eur Spine J 18, 1102–1108 (2009)

1  Surgical Anatomy   9. Dvorak, J.: Rotation of the cervical spine by using computerized-tomography (CT). Spine (Phila Pa 1976) 13, 595–597 (1988) 10. Dvorak, J., Panjabi, M.M.: Functional anatomy of the alar ligaments. Spine (Phila Pa 1976) 12, 183–189 (1987) 11. Ebraheim, N.A., Lu, J., Biyani, A., et al.: An anatomic study of the thickness of the occipital bone. Implications for occipitocervical instrumentation. Spine (Phila Pa 1976) 21, 1725– 1729 (1996). discussion 1729–1730 12. Fine, A.D., Cardoso, A., Rhoton Jr., A.L.: Microsurgical anatomy of the extracranial-extradural origin of the posterior inferior cerebellar artery. J Neurosurg 91, 645–652 (1999) 13. Francis, C.C.: Variations in the articular facets of the cervical vertebrae. Anat Rec 122, 589–602 (1955) 14. Grob, D., Dvorak, J., Panjabi, M.M., et al.: The role of plate and screw fixation in occipitocervical fusion in rheumatoid arthritis. Spine (Phila Pa 1976) 19, 2545–2551 (1994) 15. Gupta, T.: Cadaveric morphometric anatomy of C-1 vertebra in relation to lateral mass screw placement. Surg Radiol Anat 30, 589–593 (2008) 16. Hawkins, T.D., Melcher, D.H.: A meningeal artery in the falx cerebelli. Clin Radiol 17, 377–383 (1966) 17. Heggeness, M.H., Doherty, B.J.: The trabecular anatomy of the axis. Spine (Phila Pa 1976) 18, 1945–1949 (1993) 18. Heller, J.G., Alson, M.D., Schaffler, M.B., et al.: Quantitative internal dens morphology. Spine (Phila Pa 1976) 17, 861– 866 (1992) 19. Hong, X., Dong, Y., Yunbing, C., et  al.: Posterior screw placement on the lateral mass of atlas: an anatomic study. Spine (Phila Pa 1976) 29, 500–503 (2004) 20. Hong, J.T., Lee, S.W., Son, B.C., et al.: Analysis of anatomical variations of bone and vascular structures around the posterior atlantal arch using three-dimensional computed tomography angiography. J Neurosurg Spine 8, 230–236 (2008) 21. Howington, J.U., Kruse, J.J., Awasthi, D.: Surgical anatomy of the C-2 pedicle. J Neurosurg 95, 88–92 (2001) 22. Jian, F.Z., Santoro, A., Wang, X.W., et  al.: A vertebral artery tortuous course below the posterior arch of the atlas (with­out passing through the transverse foramen). Anato­ mical report and clinical significance. J Neurosurg Sci 47, 183–187 (2003) 23. Jun, B.Y.: Anatomic study for ideal and safe posterior C1-C2 transarticular screw fixation. Spine (Phila Pa 1976) 23, 1703–1707 (1998) 24. Kadri, P.A., Al-Mefty, O.: Anatomy of the nuchal ligament and its surgical applications. Neurosurgery 61, 301–304 (2007). discussion 304 25. Kandziora, F., Schulze-Stahl, N., Khodadadyan-Klostermann, C., et al.: Screw placement in transoral atlantoaxial plate systems: an anatomical study. J Neurosurg 95, 80–87 (2001) 26. Kazan, S., Yildirim, F., Sindel, M., et al.: Anatomical evaluation of the groove for the vertebral artery in the axis vertebrae for atlanto-axial transarticular screw fixation technique. Clin Anat 13, 237–243 (2000) 27. Koller, H., Kammermeier, V., Ulbricht, D., et al.: Anterior retropharyngeal fixation C1-2 for stabilization of atlantoaxial instabilities: study of feasibility, technical description and preliminary results. Eur Spine J 15, 1326–1338 (2006) 28. La Marca, F., Zubay, G., Morrison, T., et al.: Cadaveric study for placement of occipital condyle screws: technique and

References effects on surrounding anatomic structures. J Neurosurg Spine 9, 347–353 (2008) 29. Lang, J.: The cranio-cervical junction – Anatomy. In: Voth, D., Glees, P. (eds.) Diseases in the cranio-cervical junction. Anatomical and pathological aspects and detailed clinical accounts, pp. 27–61. Gruyter, Berlin, New York (1987) 30. Lee, M.J., Cassinelli, E., Riew, K.D.: The feasibility of inserting atlas lateral mass screws via the posterior arch. Spine (Phila Pa 1976) 31, 2798–2801 (2006) 31. Lee, J.H., Jahng, T.A., Chung, C.K.: C1-2 transarticular screw fixation in high-riding vertebral artery: suggestion of new trajectory. J Spinal Disord Tech 20, 499–504 (2007) 32. Lu, J., Ebraheim, N.A., Yang, H., et al.: Anatomic considerations of anterior transarticular screw fixation for atlantoaxial instability. Spine (Phila Pa 1976) 23, 1229–1235 (1998). discussion 1236 33. Madawi, A.A., Casey, A.T., Solanki, G.A., et al.: Radiological and anatomical evaluation of the atlantoaxial transarticular screw fixation technique. J Neurosurg 86, 961–968 (1997) 34. Madawi, A.A., Solanki, G., Casey, A.T., et al.: Variation of the groove in the axis vertebra for the vertebral artery. Implications for instrumentation. J Bone Joint Surg Br 79, 820–823 (1997) 35. Moftakhar, P., Gonzalez, N.R., Khoo, L.T., et al.: Osseous and vascular anatomical variations within the C1-C2 complex: a radiographical study using computed tomography angiography. Int J Med Robot 4, 158–164 (2008) 36. Muthukumar, N., Swaminathan, R., Venkatesh, G., et al.: A morphometric analysis of the foramen magnum region as it relates to the transcondylar approach. Acta Neurochir (Wien) 147, 889–895 (2005) 37. Naderi, S., Cakmakci, H., Acar, F., et  al.: Anatomical and computed tomographic analysis of C1 vertebra. Clin Neurol Neurosurg 105, 245–248 (2003) 38. Naderi, S., Korman, E., Citak, G., et al.: Morphometric analysis of human occipital condyle. Clin Neurol Neurosurg 107, 191–199 (2005) 39. Nadim, Y., Lu, J., Sabry, F.F., et  al.: Occipital screws in occipitocervical fusion and their relation to the venous sinuses: an anatomic and radiographic study. Orthopedics 23, 717–719 (2000) 40. Neo, M., Matsushita, M., Iwashita, Y., et  al.: Atlantoaxial transarticular screw fixation for a high-riding vertebral artery. Spine (Phila Pa 1976) 28, 666–670 (2003) 41. Newton, T.H.: The anterior and posterior meningeal branches of the vertebral artery. Radiology 91, 271–279 (1968) 42. Olivier, G.: Biometry of the human occipital bone. J Anat 120, 507–518 (1975) 43. Panjabi, M., Dvorak, J., Crisco 3rd, J.J., et al.: Effects of alar ligament transection on upper cervical spine rotation. J Orthop Res 9, 584–593 (1991) 44. Panjabi, M., Dvorak, J., Crisco 3rd, J., et al.: Flexion, extension, and lateral bending of the upper cervical spine in response to alar ligament transections. J Spinal Disord 4, 157–167 (1991)

15 45. Paramore, C.G., Dickman, C.A., Sonntag, V.K.: The anatomical suitability of the C1-2 complex for transarticular screw fixation. J Neurosurg 85, 221–224 (1996) 46. Resnick, D.K., Lapsiwala, S., Trost, G.R.: Anatomic suitability of the C1-C2 complex for pedicle screw fixation. Spine (Phila Pa 1976) 27, 1494–1498 (2002) 47. Roberts, D.A., Doherty, B.J., Heggeness, M.H.: Quantitative anatomy of the occiput and the biomechanics of occipital screw fixation. Spine (Phila Pa 1976) 23, 1100–1107 (1998). discussion 1107–1108 48. Rocha, R., Safavi-Abbasi, S., Reis, C., et al.: Working area, safety zones, and angles of approach for posterior C-1 lateral mass screw placement: a quantitative anatomical and morphometric evaluation. J Neurosurg Spine 6, 247–254 (2007) 49. Santos-Franco, J.A., de Oliveira, E., Mercado, R., et  al.: Microsurgical considerations of the anterior spinal and the anterior-ventral spinal arteries. Acta Neurochir (Wien) 148, 329–338 (2006). discussion 338 50. Schmidt, H., Fisher, E.: Die okzipitale Dysplasie. Thieme, Stuttgart (1960) 51. Seal, C., Zarro, C., Gelb, D., et al.: C1 lateral mass anatomy: proper placement of lateral mass screws. J Spinal Disord Tech 22, 516–523 (2009) 52. Seckin, H., Ates, O., Bauer, A.M., et al.: Microsurgical anatomy of the posterior spinal artery via a far-lateral transcondylar approach. J Neurosurg Spine 10, 228–233 (2009) 53. Solanki, G.A., Crockard, H.A.: Preoperative determination of safe superior transarticular screw trajectory through the lateral mass. Spine (Phila Pa 1976) 24, 1477–1482 (1999) 54. Tan, M., Wang, H., Wang, Y., et al.: Morphometric evaluation of screw fixation in atlas via posterior arch and lateral mass. Spine (Phila Pa 1976) 28, 888–895 (2003) 55. Tokuda, K., Miyasaka, K., Abe, H., et al.: Anomalous atlantoaxial portions of vertebral and posterior inferior cerebellar arteries. Neuroradiology 27, 410–413 (1985) 56. Tsusaki, T.: Über den Atlas und Epistropheus bei den eingeborenen Formosanern. Folia Anatomica Japonica 2, 221–246 (1924) 57. Tubbs, R.S., Salter, E.G., Oakes, W.J.: The accessory atlantoaxial ligament. Neurosurgery 55, 400–402 (2004). discussion 402–404 58. Xu, R., Burgar, A., Ebraheim, N.A., et al.: The quantitative anatomy of the laminas of the spine. Spine (Phila Pa 1976) 24, 107–113 (1999) 59. Xu, R., Nadaud, M.C., Ebraheim, N.A., et al.: Morphology of the second cervical vertebra and the posterior projection of the C2 pedicle axis. Spine (Phila Pa 1976) 20, 259–263 (1995) 60. Young, J.P., Young, P.H., Ackermann, M.J., et al.: The ponticulus posticus: implications for screw insertion into the first cervical lateral mass. J Bone Joint Surg Am 87, 2495–2498 (2005) 61. Zipnick, R.I., Merola, A.A., Gorup, J., et al.: Occipital morphology. An anatomic guide to internal fixation. Spine (Phila Pa 1976) 21, 1719–1724 (1996). discussion 1729–1730

2

Biomechanical Remarks P. Suchomel and P. Buchvald

Knowledge of normal biomechanics of the cervical spine is very important as it can be modified by various pathological situations. The changes that occur during injury and/or in consequence with other pathological conditions or surgical procedures can substantially influence the stability of this most important spinal joint complex. It is difficult to determine what the normal motion of the cervical spine is as it depends on the size, weight, anatomy, degree of degeneration, bone quality, and age of each person or specimen. Both in vivo and in vitro investigations have been undertaken to accumulate the clinically important biomechanical data. Performing the in vitro studies, various fresh cadaver spine specimens were tested. Most often, the six motion components were evaluated: flexion/extension, axial rotation, lateral bending, and translation about each axis. A number of techniques have been developed to apply loads and to measure these motion components. Pioneering work in this field is credited to Panjabi et al. [21]. They monitored the three- dimensional motion by an optoelectronic system based on the principles of stereophotogrammetry. In vivo motion analyses are usually based on the CT investigations [27], the electrogoniometer gauging technique, [1] or the stereophotogrammetry [25]. The occipitoatlantoaxial complex (C0-C1-C2) is a very complicated structure with motion determined by the bony morphology and orientation of the articular processes and limited by ligaments and joint capsules. It is composed of the occipitoatlantal (C0-C1) and atlantoaxial (C1-C2) joint complexes. We should emphasize that these two motion segments are intimately linked and the motion is always coupled. P. Suchomel () and P. Buchvald Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic

The atlantooccipital joints (C0-C1) are anteromedially oriented, concave spheroid articulations connected by very tight capsules. Their mechanical properties are determined mainly by the shape of bony elements. Flexion and extension reported between 13° and 25° (in total range), according to different investigators, is their dominant movement [10, 23, 26, 30, 33]. Flexion is limited by the tip of the dens impinging on the anterior margin of the foramen magnum (bursa apicis dentis) [33] and extension is restricted mainly by the tectorial membrane inserted to the body of axis and the anterior rim of the foramen magnum; nevertheless, the exact function of tectorial membrane is still a matter of debate [18, 30, 31, 33]. Translation at this junction is minimal under normal conditions and during sagittal movement should not change more than 1 mm [24, 37]. Allowed lateral bending is between 3° and 5° to each side [23, 26, 30]. Although the idea of possible axial rotation had been refused in the past, more recently some authors have documented existence of minimal axial rotation in this joint. The one-side rotational movement range was measured between 1° and 7.2° [4, 10, 23, 27]. The rotation and lateral bending of C0-C1 is controlled mainly by the joint capsules but also the allar ligaments. The instantaneous axis of axial rotation (IAR) for the C0-C1 articulation is ventral to foramen magnum. The atlantoaxial complex (C1-C2) is composed of four joints: two AA lateral joints, the atlantoaxial median joint (between the anterior arch of the atlas and the dens axis), and the joint between the posterior surface of the dens and the transverse ligament. Stability at this highly mobile junction is dependent predominantly on ligamentous structures. Sagittal plane motion (flexion-extension) in C1-C2 has been reported to be on an average 20° (10°– 30°) by several authors [7, 16, 33]. Lateral bending limited by allar ligaments is inconsequential under normal conditions by some authors [33] but reaching 7°–10° to

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_2, © Springer-Verlag Berlin Heidelberg 2011

17

18

one side by others [23, 27]. In the occipitoatlantoaxial complex, 85–90% of all axial rotation comes from the atlantoaxial segment [10, 23]. Penning and Wilmink [27] found that atlantoaxial complex accounted for 56% of the whole cervical rotation with respect to the first thoracic vertebra. Normal range of rotation between C1 and C2 is determined on an average about 40° to each side [3, 17, 34]. Range of axial rotation to one side in C1-C2 has been reported in the various studies to be between 23° [10] and 47° [33]. The great differences in the results are mostly due to the differences in the methods used and dissimilarities of in vitro and in vivo studies. For example, Dvorak et al. reported in their in vivo (CT) tests an average range of axial rotation 32.2° and consecutively, 43.1° [4]. The high rotational motion range is facilitated by very loose AA joint capsules and limited quite freely by allar ligaments. The allar ligaments (connecting the dens axis with occipital condyles and the anterior arch of the atlas) consist of high amount of collagen fibers and their major function is to prevent redundant axial rotation to the contralateral side [6, 22, 33]. These ligaments together with tectorial membrane also limit the flexion of the occiput and during lateral bending are responsible for the forced rotation of the axis [3]. The cruciate ligament is formed by horizontally oriented TAL and vertically oriented longitudinal fibers (Fig. 1.5, Chap. 1). Between the odontoid process and transverse ligament is a thin layer of cartilage, which allows for TAL to move freely during rotation and preserves it from friction damage.

2  Biomechanical Remarks

The transverse ligament consists of collagen fibers and is very resistant to breakage. Spence et al. in their original cadaver tests reported stress necessary to TAL rupture in average 580 N (38–104 kg) [29]. Dvorak et al. described experimental failure of TAL only under the force 170– 700 N (corresponding about 17–70 kg) [8]. Restriction of anterior translation of the atlas during flexion of the head is the main function of TAL while still permitting its axial rotation around the dens. The secondary restriction of this motion is secured by atlantodental component of the allar ligaments. The tertiary stabilizers are the accessory atlantoaxial ligaments and capsular ligaments [5, 6]. TAL partially protects the C1-C2 joint also from a rotary dislocation. Posterior translation is prevented by mechanical bracing of the anterior portion of C1 on the dens. The apical ligament due to its laxity probably has no important function is movement restriction [15]. The IAR for sagittal plane motion is located in the region of the middle third of the dens and for axial rotation in the central axis of the dens, respectively [33].

2.1 CVJ and UCS Axial Load Distribution The motion characteristics of UCS are unique; nevertheless, the axial load distribution in CVJ represents another exceptional situation irreproducible in other parts of spine (Fig.  2.1). The weight of the head and the axial

Fig. 2.1  Diagram showing the axial load distribution changing from two to three force vectors at the C2 level. (a) CT in 3D reconstruction in frontal plane. (b) CT in 3D reconstruction in sagittal plane. Note the consequent hangman type fracture

2.2  Clinical and Morphological Instability of CVJ and UCS

19

Fig. 2.2  Traumatic consequences of atypical axial load distribution in CVJ and UCS depicted on coronal CT reconstructions. (a) fracture of occipital condyle. (b) lateral C1 mass fracture

simultaneous with condyle fracture. (c) fracture of lateral facet C2 pillar. Note the TAL avulsion fragment and type III odontoid fracture

loads applied to the head are transmitted through the two occipital condyles to two AO joints. The wedge-shaped lateral masses of C1 transfer the weight downwards with logical tendency to distract laterally. If the C1 ring and TAL act properly, the force is further transmitted to two C2 facets; however, further load distribution in C2 vertebra is divided from two force vectors into three points at the C2-C3 interface [14]. Most of the load is thus transmitted to the C2-C3 disk and less to the posterior C2/C3 facetal joints. The force transmission divergence has some critical places where consequently fractures can arise in the case of overload (Fig. 2.2). First critical place is the atlas. Wedge-shaped lateral masses and the C1 free floating “washer” ring have to buffer the axial force and if overloaded the atlas bursting leading to Jefferson-like fractures can happen (Chap. 10). The second critical location is the C2 pars interarticularis and mainly its isthmus as locus of minor resistance. The axial force overload can lead to overstressing of the bone resistance and create the hangman type fractures. Certainly, previously described model situations can be further modified by concomitant rotation, lateral bending of sagittal movement. This physiological CVJ and UCS load distribution has to be respected during reconstruction procedures also.

choice in various types of pathological lesions in this region. Clinical stability at the C0-C1 and C1-C2 joints is intimately linked to their functional anatomy. Clinical instability can occur as a result of trauma, degenerative conditions, tumors, inflammation, or surgery; however, its definition is still controversial. Significant disagreement exists even among experts. White and Panjabi [35] defined clinical instability as the loss of the ability of the spine under physiologic loads to maintain relationships between vertebrae in such a way that there is neither initial nor subsequent damage to the spinal cord or nerve roots, and in addition, there is neither development of incapacitating deformity nor severe pain. They further defined physiological loads as loads that are incurred during normal activity. Incapacitating deformity was defined as gross deformity that the patient finds intolerable. Severe pain was defined as pain that cannot be controlled by non-opioid analgesic medications. In short, it means that the spine is unable to resist the physiological loads without pain, deformity and/or neurological deficit. Such a broad definition, in fact, encompasses nearly all the pathological conditions detectable in UCS. The definition of mechanical instability is more exactly specified than clinical one. In fact, whatever static or dynamic position of UCS beyond the physiological limits is detected must be considered as instability. Static and dynamic radiographic investigations with consecutive exact measurements are providing us information necessary to decide if the spine is stable or not (see Chap. 3).

2.2 Clinical and Morphological Instability of CVJ and UCS It is of utmost importance to decide whether the UCS is stable or not in order to determine correct treatment

20

2  Biomechanical Remarks

Table 2.1  Morphological criteria for upper cervical spine instability [35] >8°

Axial rotation C0-C1 to one side (measurable only on CT)

>1 mm

C0-C1 translation (measurable only on C)

>7 mm

Overhang C1-C2 (total right and left, on anteroposterior radiograph)

>45°

Axial rotation C1-C2 to one side

>3 mm

C1-C2 translation at anterior atlantodental interval (AADI)

<13 mm

posterior atlantodental interval (PADI)

Avulsed transverse ligament

2.3 Occipitoatlantal Joint Stability and Instability Stability in this joint is secured mainly by their tight capsules, the anterior and posterior atlanto-occipital membrane, and through the ligaments between the occiput and the axis: the tectorial membrane, allar ligaments, and apical ligament [13]. Instability in the C0-C1 joints is less common than at the C1-C2 level. It can be result of trauma, rheumatoid arthritis, infection, tumor, or destabilizing surgery. Vishteh et al. experimentally demonstrated the AO hypermobility caused by resection of occipital condyle. They found flexion-extension, lateral bending, and axial rotation increased 15.3%, 40.8%, and 28.1%, respectively after 50% condylectomy [32]. The AO joint is relatively unstable in children because of its structural characteristic and ligamentous laxity. Its stability increases in adulthood due to a decrease in elasticity of the ligaments [24]. OA dislocations (AOD) can be in the anterior, posterior, or longitudinal directions. Normally, the AO joint sagittal translation should not exceed 1mm, and the distraction distance (CCI) can reach 2 mm maximally on parasagittal CT images [19, 20, 36]. From other parameters used to evaluate the AOD, the Powers ratio is most frequently used [28]; however, the basion-dental interval (BDI) and basion-posterior axial line interval (BAI) both of which should not exceed 12  mm (“rule of twelve”) are considered as the most exact and AODspecific measurements [2, 11, 12]. Greater than 8° of unilateral axial rotation as other instability sign can be seriously measured only on superimposed CT scans

[4, 6]. Basilar invagination and/or basilar impression represent the vertical instability. It appears most often in developmental anomalies and rheumatoid arthritis but also can occur in tumors or trauma. Methods of determining vertical CVJ instability are described in detail in appropriate chapters (see Chaps. 3 and 20).

2.4 Atlantoaxial Joint Stability and Instability The transverse atlantal ligament has a key role in the maintaining AA stability. Especially, allar ligaments but also atlantoaxial accessory ligaments, apical ligament, and joint capsules provide secondary security [13, 22]. Instability at the AA joint is often presented as abnormal translation and/or axial rotation; nevertheless, other dislocations are also possible. Mainly TAL limits anterior translation of C1 on C2 as it will be described in further chapters. Fielding et al. [9] noted that anterior AADI is normally not more than 3  mm. AADI of 3–5 mm implies damage of the transverse ligament, and AADI measured 5 mm or more indicates that the accessory stabilizing system (especially, the allar ligaments) has been also damaged. A PADI of less than 13  mm may also denote anterior translational instability. The allar ligaments alone are not capable of preventing excessive anterior horizontal displacement if the transverse ligament is ruptured. If the odontoid process is hypoplastic, fractured, or resected logically the ligaments also cannot provide their stabilizing function. Posterior AA translation, despite being rare, can be detected in trauma, tumor, or other pathologies destroying odontoid – TAL catch system as well. Nonetheless, more frequently, the rotational dislocations of different types are diagnosed [3, 4, 6]. The rotation-limiting ability of the alar ligament was investigated by Dvorak et  al. [6] in cadaver studies. They observed a mean increase of 9° or 30% of the original mean rotation divided equally between the C0-C1 (+5°) and C1-C2 (+4°) complexes in axial rotation in response to an alar ligament lesion on the opposite side. The laboratory findings of Dvorak and Panjabi were verified with a clinical CT study of 9 healthy adults and 43 patients with cervical spine instability with conclusion that axial rotation of the C0-C1-C2 complex can be increased after trauma-lesions of the alar ligaments [5]. In general, these studies showed that the main function

References

of the alar ligament is to limit axial rotation to the contralateral side. The transverse ligament also protects the atlantoaxial joint from a rotatory dislocation. Fielding et al. [9] described that with the intact transverse ligament, a complete bilateral rotational AA dislocation can occur if 65° or more is reached. With transverse ligament disruption dislocation can occur at 45° of rotation already. Total lateral displacement of more than 6.9 mm of the lateral masses of C1 over the C2 facets, as measured on the cadaver tests, determines disruption or avulsion of the transverse ligament [29].

2.5 For Practical Purposes We Can Summarize The motion of AO-AA joint complex is always coupled. UCS is responsible for 60% of rotation and 40% flexion and extension of the whole cervical spine. Atlas has the widest range of motion of any vertebra in the spine. It is almost freely floating in between the occiput and C2 buffering the axial loads coming from head to spine. The AO joint mainly provides flexion and extension in the range of 20°, lateral bending is possible up to 10°. Negligible rotation is possible; however, translation of more than 1 mm is considered as pathological. The joint distraction measured on CT should not exceed 2 mm. The AA joint is responsible for 90% of axial rotation of the UCS complex with the average range of rotational motion to one side of 40°. Flexion-extension is possible at an average of 20° and the lateral bending can reach up to 10°. Whatever parameter measured beyond the physiological limits has to be considered as mechanical instability. Performing complex reconstruction of the UCS the load distribution typical for this spine region has to be respected and the non-affected segments spared to preserve as much movement as possible.

References   1. Alund, M., Larsson, S.E.: Three-dimensional analysis of neck motion. A clinical method. Spine (Phila Pa 1976) 15, 87–91 (1990)   2. Bono, C.M., Vaccaro, A.R., Fehlings, M., et al.: Measurement techniques for upper cervical spine injuries: consensus statement of the Spine Trauma Study Group. Spine (Phila Pa 1976) 32, 593–600 (2007)

21   3. Dvorak, J., Froehlich, D., Penning, L., et  al.: Functional radiographic diagnosis of the cervical spine: flexion/extension. Spine (Phila Pa 1976) 13, 748–755 (1988)   4. Dvorak, J., Hayek, J., Zehnder, R.: CT-functional diagnostics of the rotatory instability of the upper cervical spine. Part 2. An evaluation on healthy adults and patients with suspected instability. Spine (Phila Pa 1976) 12, 726–731 (1987)   5. Dvorak, J., Panjabi, M.M.: Functional anatomy of the alar ligaments. Spine (Phila Pa 1976) 12, 183–189 (1987)   6. Dvorak, J., Panjabi, M., Gerber, M., et  al.: CT-functional diagnostics of the rotatory instability of upper cervical spine. 1. An experimental study on cadavers. Spine (Phila Pa 1976) 12, 197–205 (1987)   7. Dvorak, J., Panjabi, M.M., Novotny, J.E., et al.: In vivo flexion/extension of the cervical spine. J Orthop Res 9, 824–834 (1991)   8. Dvorak, J., Schneider, E., Saldinger, P., et al.: Biomechanics of the craniocervical region: the alar and transverse ligaments. J Orthop Res 6, 452–461 (1988)   9. Fielding, J.W., Cochran, G.B., Lawsing 3rd, J.F., et al.: Tears of the transverse ligament of the atlas. A clinical and ­biomechanical study. J Bone Joint Surg Am 56, 1683–1691 (1974) 10. Goel, V.K., Clark, C.R., Gallaes, K., et al.: Moment-rotation relationships of the ligamentous occipito-atlanto-axial complex. J Biomech 21, 673–680 (1988) 11. Harris Jr., J.H., Carson, G.C., Wagner, L.K.: Radiologic diagnosis of traumatic occipitovertebral dissociation: 1.  Normal occipitovertebral relationships on lateral radiographs of supine subjects. AJR Am J Roentgenol 162, 881– 886 (1994) 12. Harris Jr., J.H., Carson, G.C., Wagner, L.K., et al.: Radiologic diagnosis of traumatic occipitovertebral dissociation: 2. Comparison of three methods of detecting occipitovertebral relationships on lateral radiographs of supine subjects. AJR Am J Roentgenol 162, 887–892 (1994) 13. Hecker, P.: Appareil ligamenteux occipito-atloidoaxoidien: étude d’anatomie comparée. Arch Anat Hist Embryol 2, 57–95 (1923) 14. Jeszenszky, D., Fekete, T.F., Melcher, R., et al.: C2 prosthesis: anterior upper cervical fixation device to reconstruct the second cervical vertebra. Eur Spine J 16, 1695–1700 (2007) 15. Lang, J.: The cranio-cervical junction – Anatomy. In: Voth, D., Glees, P. (eds.) Diseases in the cranio-cervical junction. Anatomical and pathological aspects and detailed clinical accounts, pp. 27–61. Gruyter, Berlin, New York (1987) 16. Lin, R.M., Tsai, K.H., Chu, L.P., et  al.: Characteristics of sagittal vertebral alignment in flexion determined by dynamic radiographs of the cervical spine. Spine (Phila Pa 1976) 26, 256–261 (2001) 17. Monckeberg, J.E., Tome, C.V., Matias, A., et  al.: CT scan study of atlantoaxial rotatory mobility in asymptomatic adult subjects: a basis for better understanding C1-C2 rotatory fixation and subluxation. Spine (Phila Pa 1976) 34, ­1292–1295 (2009) 18. Oda, T., Panjabi, M.M., Crisco 3rd, J.J., et al.: Role of tectorial membrane in the stability of the upper cervical spine. Clin Biomech 7, 201–207 (1992) 19. Pang, D., Nemzek, W.R., Zovickian, J.: Atlanto-occipital dislocation: part 1 – normal occipital condyle-C1 interval in 89 children. Neurosurgery 61, 514–521 (2007). discussion 521

22 20. Pang, D., Nemzek, W.R., Zovickian, J.: Atlanto-occipital dislocation – part 2: The clinical use of (occipital) condyleC1 interval, comparison with other diagnostic methods, and the manifestation, management, and outcome of atlantooccipital dislocation in children. Neurosurgery 61, 995–1015 (2007). discussion 1015 21. Panjabi, M.M., Duranceau, J., Goel, V., et  al.: Cervical human vertebrae. Quantitative three-dimensional anatomy of the middle and lower regions. Spine (Phila Pa 1976) 16, 861–869 (1991) 22. Panjabi, M., Dvorak, J., Crisco 3rd, J.J., et al.: Effects of alar ligament transection on upper cervical spine rotation. J Orthop Res 9, 584–593 (1991) 23. Panjabi, M., Dvorak, J., Duranceau, J., et  al.: Threedimensional movements of the upper cervical spine. Spine (Phila Pa 1976) 13, 726–730 (1988) 24. Panjabi, M.M., Yue, J.J., Dvorak, J., et  al.: Cervical spine kinematics and clinical instability. In: Clark, C.R., Benzel, E.C., Currier, B.L., et al. (eds.) The cervical spine, vol. 4, pp. 55–78. Lippincott, Philadelphia (2005) 25. Pearcy, M.J., Whittle, M.W.: Movements of the lumbar spine measured by three-dimensional X-ray analysis. J Biomed Eng 4, 107–112 (1982) 26. Penning, L.: Normal movements of the cervical spine. AJR Am J Roentgenol 130, 317–326 (1978) 27. Penning, L., Wilmink, J.T.: Rotation of the cervical spine. A CT study in normal subjects. Spine (Phila Pa 1976) 12, 732–738 (1987)

2  Biomechanical Remarks 28. Powers, B., Miller, M.D., Kramer, R.S., et  al.: Traumatic anterior atlanto-occipital dislocation. Neurosurgery 4, 12–17 (1979) 29. Spence Jr., K.F., Decker, S., Sell, K.W.: Bursting atlantal fracture associated with rupture of the transverse ligament. J Bone Joint Surg Am 52, 543–549 (1970) 30. Steinmetz, M.P., Mroz, T.E., Benzel, E.C.: Craniovertebral junction: biomechanical considerations. Neurosurgery 66, A7–A12 (2010) 31. Tubbs, R.S., Kelly, D.R., Humphrey, E.R., et al.: The tectorial membrane: anatomical, biomechanical, and histological analysis. Clin Anat 20, 382–386 (2007) 32. Vishteh, A.G., Crawford, N.R., Melton, M.S., et al.: Stability of the craniovertebral junction after unilateral occipital condyle resection: a biomechanical study. J Neurosurg 90, 91–98 (1999) 33. Werne, S.: Studies in spontaneous atlas dislocation. Acta Orthop Scand Suppl 23, 1–150 (1957) 34. White, A.P.M.: Kinematics of the spine. Lippincott, Philadelphia (1990) 35. White, A.A., Panjabi, M.M.: Clinical biomechanics of the spine. Lippincott, Philadelphia (1990) 36. Wiesel, S., Kraus, D., Rothman, R.H.: Atlanto-occipital hypermobility. Orthop Clin North Am 9, 969–972 (1978) 37. Wiesel, S.W., Rothman, R.H.: Occipitoatlantal hypermobility. Spine (Phila Pa 1976) 4, 187–191 (1979)

3

Special Radiology O. Choutka and P. Suchomel

The anatomy and pathology of the craniovertebral junction (CVJ) may be complex but can be readily visualized by a variety of radiological means. The primary modalities include simple plain radiographs, computer tomography (CT), and magnetic resonance imaging (MRI). Each clinical scenario warrants a ­different imaging modality or, more commonly, a ­combination of multiple modalities. This chapter describes these imaging modalities as they pertain to the CVJ. Specific pathologies are discussed in separate chapters. Historically, plain films have been the radiographic gold standard for assessment of the spine in general and form an essential part of spinal evaluation today, more than 100 years since Wilhelm Roentgen shared the firstever radiograph of his wife’s hand in 1895 [23]. Bony and some soft tissue abnormalities are well visualized on plain films. Lateral cervical radiographs are the most commonly used images in acute evaluation of the cervical spine. For example, they are used while patient is still on the stretcher and further determine the way patient can be handled through their traumatic work up. This is even more important for the unconscious patient. The most commonly missed traumatic injuries are at the lower end of the cervical spine [44] and different projections such as swimmer’s (“flying angel”) view have been designed to enhance the visibility of the cervicothoracic junction. Similarly, multiple views exist to carefully delineate CVJ.

O. Choutka Department of Neurosurgery, University of Cincinnati College of Medicine, 231 Albert Sabin Way, PO Box 670515, Cincinnati, OH 45267-0515, USA P. Suchomel Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic

Lateral projection allows for a good assessment of the alignment of the bony components of the spine and also sagittal balance when performed upright. Any abnormality detected on bony spinal canal (fracture, subluxation) necessitates further examination to determine its cause. Prevertebral soft tissue swelling can point one to the area of injury as it often indicates a presence of a hematoma secondary to a fracture. Anteroposterior view is commonly obliterated by the jaw; so, openmouth view films are particularly useful in assessing odontoid pathology as well as the integrity of the atlantoaxial and atlanto-occipital relationships. Allesandro Vallebona proposed to represent a single slice of a body part on a radiograph, the so-called tomography, a technique that remained the pillar of radiology until the late 1970s [32]. However, the availability of computers and transverse axial scanning resulted in the development of CT by Godfrey Hounsfield and Allan McLeod Cormack [36]. Since then, multi-slice CT has revolutionized cross-sectional imaging with scanning time down to a single breath hold today. Isotropic voxels allow for two-dimensional reformatting and thus production of high quality threedimensional images that are particularly useful when assessing complex bony abnormalities of the CVJ. However, the disadvantage of CT is increased radiation dose to patients with its ever more prevalent sequelae, particularly in pediatric population [4]. Although myelography with CT allows for excellent neural structure delineation and skeletal correlation, it has been largely replaced by MRI technology credited to Paul Lauterbur and Sir Peter Mansfield who were awarded Nobel Prize in 2003, albeit some controversy surrounds the award [40]. MRI is excellent in evaluation of neural, ligamentous, and disk structures. Sagittal images become particularly useful in CVJ, evaluation of alignment, and assessment of various craniometric lines and angles.

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_3, © Springer-Verlag Berlin Heidelberg 2011

23

24

3  Special Radiology

Upright films (plain radiography or even MRI) can be a useful adjunct to evaluation of any spinal pathology that may change under loading conditions of the erect human body [13, 47]. Under appropriate supervision, plain films, CT, and MRI can be obtained in dynamic positions when concern for instability exists, thus providing further information about soft tissue and bony dynamics of the CVJ. Multiple parameters based on static and dynamic films have been developed to grade clinical status or instability of the CVJ. Vascular anatomy of the CVJ can be evaluated by multiple modalities. Angiography later improved by digitalized subtraction was the standard since its development by Egas Moniz in 1927 [43]. However, with CT and MR angiography available, only very specific situations [51] would require the patient to be subjected to the potential risks of conventional angiography nowadays.

3.1 Radiographic Data Analysis Advances in neuroimaging allow for excellent ­visualization of the CVJ and UCS. Therefore, an effective diagnostic approach is necessary to prevent unnecessary secondary neurological sequelae (e.g., spinal cord injury from a missed traumatic atlantooccipital dislocation). Plain radiography remains the mainstay of early evaluation of the cervical spine in majority of hospitals. One has to be aware that lateral cervical spine films are centered on the C3 vertebra; therefore, the visualization of CVJ may be skewed by the oblique nature of the atlanto-occipital joints relative to the x-ray beams. Coned down views of the UCS or lateral skull films can clarify this relationship a

[8]. Although plain films have been surpassed by the quality of other modalities in this region, there are multiple indirect signs on plain films, pointing to craniocervical junction abnormalities. These include prevertebral soft tissue swelling, lack of override of mastoid processes of the odontoid tip, and disruption of UCS laminae [31]. Once seen, further evaluation is warranted. Over the years, numerous craniometric parameters (lines, planes, angles, and relationships) have been described and tested in practice to assess radiographic alignment of the structures of CVJ (Table 3.1). Specific osseous anatomical landmarks need to be visualized on imaging to derive those lines/angles to be able to detect and quantify an abnormality. Those include nasion, tuberculum sellae, basion (anterior margin of FM), opisthion (posterior margin of FM), posterior hard palate edge, anterior and posterior arches of atlas, odontoid process, and C2 vertebral body (Fig. 3.1).

3.1.1 Basal/Clival Parameters The Welcher basal angle can be measured on plain films or mid-sagittal CT or MRI images and is formed by the nasion-tuberculum and tuberculum basion lines [45]. It is a modification to the angle of Landzert, which represents an angle between planum sphenoidale and the clivus. Welcher basal angle averages 132° and should always be less than 140° [10]. Lanzert’s angle changes during fetal development and then averages 113.9° in adults [16, 55]. Both angles abnormally increase when the skull base is flat as in platybasia with or without basilar impression (Fig. 3.2).

b

Fig. 3.1  Relevant anatomical landmarks needed to be ­visualized for analysis of CVJ craniometric parameters: 1 nasion, 2 tuberculum sella, 3 basion (anterior margin of FM), 4 opisthion ­(posterior margin of FM), 5 posterior hard palate edge, 6 ­anterior

and 7 posterior arches of atlas, 8 odontoid process, and 9 C2 vertebral body. (a) lateral plain radiogram. (b) CT in sagittal plane. (c) MRI in sagittal plane

25

3.1  Radiographic Data Analysis Table 3.1  Craniometric parameters Line/angle Film

Anatomic landmarks/relationship

Normal value

Pathology

Wackenheim line

L

Posterior surface of clivus and dens

Tangent to the dens and behind it, may cross its posterior third

BI

Chamberlain line

L

Hard palate to opisthion and dens

Dens protrudes <5 mm above it

BI

McRae line (FM line)

L

Basion to opisthion and dens

Dens below line

BI

McGregor line

L

Hard palate to basiocciput and dens

Dens protrudes <7 mm above it

BI

Ranawat line/criterion

L

Distance in mid-dens coronal plane, between transverse axis of atlas and midpoint of C2 pedicle

>15 mm in males >13 mm in females

BI

Redlund-Johnell and Pettersson distance

L

McGregor line to midpoint of inferior C2 endplate

>34 mm in males >29 mm in females

BI

Fischgold and Metzger

AP

On AP films. Line between the mastoid tips and dens

Dens below the line

BI

Klaus posterior fossa height index

L

Perpendicular distance between dens and tuberculum/internal occipital protuberance line

>30 mm

BI

Welcher basal angle

L

Nasion to tuberculum line and tuberculum to basion line

<140°

Platybasia

Clivus-canal angle

L

Wackenheim line and posterior axial line

150°–180°

BI

Platybasia

Platybasia

Cervico-medullary angle

L

Measured on MRI, anterior medullary line and cervical spinal cord line

>135°

BI Cranial settling

Basion-dental interval (BDI)

L

Distance of basion to dens tip

<12 mm

AOD

Basion-axial interval (BAI)

L

Perpendicular distance from basion to posterior axial line

<12 mm

AOD

Atlanto-dental interval (ADI)

L

Distance dens to anterior C1 arch

<3 mm in adults <5 mm in children

AAI

Posterior ADI (PADI) Space available for cord

L

Distance dens to posterior C1 arch

>18 mm no neurology <14 mm always compromised

Neurological compromise in AAI

Atlanto-occipital joint axis angle

AP

On AP films. Angle between two lines parallel to AO joints

124°–127°

Condylar hypoplasia

Condyle-C1 interval (occipito-atlantal joint gap)

AP or L

Distance C1 lateral mass and occipital condyle

AOD

Clark Station

L

Atlantoaxial vertical relationship. Anterior atlas ring relative to axis height divided into equal vertical thirds

<2 mm adults <5 mm children Ring of atlas opposite the upper third of axis

BI

26

3  Special Radiology

Fig.  3.2  Welcher basal angle (nasion-tuberculum-basion angle). (a) Normal angle as measured on plain lateral radiograph. (b) platybasia measured on 3D CT sagittal reconstruction

Fig. 3.3  The Wackenheim clivus line in a case of atlas settling and odontoid process slight invagination

3.1.2 Craniocervical Parameters The Wackenheim clivus line is drawn along the posterior surface of clivus and extended inferiorly [52] (Fig. 3.3). The line delineates the relationship between clivus and the odontoid process. In a normal CVJ alignment, the clivus baseline falls behind the odontoid tip

Fig. 3.4  Clivus-canal (craniovertebral) angle

or may cross the posterior third. Any greater degree of odontoid transaction is seen in basilar invagination or AAD. The angle formed by Wackenheim line and posterior axial line represents the clivus-canal (craniovertebral) angle (Fig.  3.4). This angle varies between 150° and 180° depending on flexion/extension position and may represent spinal cord compression in situations of it being less than 150° [50]. The clivus-canal angle

27

3.1  Radiographic Data Analysis

is important in platybasia/basilar impression description as it is not unusual for this angle to become rather acute (even 90°) sometimes without abnormal violation of the above described Wackenheim clivus baseline. Due to common difficulties with visualization of bony structures required for the measurement of clivus-canal angle, particularly after decompressive procedures and the indirect nature of bony parameters attempting to represent neurological compromise, Bundschuh et  al. described cervico-medullary angle (CMA) in order to directly demonstrate the degree of brainstem and spinal cord compression in basilar invagination or cranial settling in rheumatoid patients [5]. They compared those values with 50 normal patients and concluded that all patients with CMA of less than 135° had evidence of brainstem compression, cervical myelopathy, or C2 nerve root pain. The angle is measured at the intersection of lines drawn parallel to ventral surfaces of the medulla and upper cervical cord on MRI (Fig.  3.5). Multiple publications since then have correlated the CMA with clinical syndrome as well as therapeutic indications and outcomes [39, 54]. However, the largest study defining the normal values of CMA found it to be approximately 158° (139°–175° range) when evaluating 200 patients’ MRI without craniovertebral anomalies [53]. Furthermore, the authors also found the CMA to have an excellent inter- and intra-observer reliability. Similar to Wackenheim clivus baseline, most of the original craniometric lines for detection of basilar

invagination or cranial settling in rheumatoid patients were originally based on plain radiographic assessment. The use of CT makes those measurements easier as bony structures can be easily visualized in the midsagittal plane. Historically, the most useful parameters measured on lateral films include Chamberlain, McGregor, and McRae’s lines (Fig. 3.6) as well as the Klaus index. Other criteria are listed in Table 3.1. (Clark station, RedlundJohnell and Pettersson, Ranawat, Fischgold and Metzger). All of these are less applicable to traumatic situations where distraction, rather than settling, occurs. Basilar invagination or impression can be diagnosed on plain radiographs when the odontoid tip protrudes at least 2 mm above the Chamberlain line: hard palate to opisthion line [6]; 4.5 mm above the McGregor line: posterosuperior aspect of the hard palate to the most caudal point on the mid-occipital curve line [28]; or any distance above the McRae line: basion to opisthion [29] (Fig. 3.6). Redlund-Johnell criterion represents the distance between McGregor line and inferior C2 vertebral body measured through its midpoint [38]. Distance of less than 34 mm in males or 29 mm in females is indicative of basilar invagination. Similarly, according to the Ranawat criterion, the distance between the center of the second cervical pedicle and the transverse axis of atlas along the odontoid process needs to be less than 15 mm in males and 13 mm in females to indicate invagination [37]. On AP transoral plain film, one can draw a line connecting the mastoid tips (FischgoldMetzger line) and if the odontoid protrudes above it, it again indicates basilar invagination [3, 30]. Using any criterion that requires exact odontoid process visualization in rheumatoid patients may be difficult. Therefore, criteria that utilized other landmarks than the odontoid process were developed as mentioned above. Indeed, the hard palate was seen in 93% of patients, followed

McRae Chamberlain

Fig. 3.5  Cervicomedullary angle as measured on MRI

Fig. 3.6  Schematic drawing of basal lines

McGregor

28

by the atlas (88%) but the odontoid tip in only 34% of patients in a study of 131 rheumatoid patients [41]. The authors went on to determine the sensitivity and specificity of the above-mentioned criteria in diagnosing basilar invagination and found that individually, none of them reached sensitivity or negative predictive value (NPV) greater than 90%, not even those that did not require odontoid visualization (i.e., RedlundJohnell and Ranawat). However, when combining Clark station, the Redlund-Johnell, and the Ranawat criteria to plain film assessment, the combined sensitivity and negative predictive value reached 94% and 91%, respectively. Practically, they recommended further imaging if basilar invagination was suspected by the combined criteria in rheumatoid patients. Using the combined criteria, 55% patients could be spared of unnecessary MRI evaluation in their opinion. The Klaus height index measures the perpendicular height from the tip of the odontoid process to the line connecting tuberculum sellae and the internal occipital protuberance [24]. This measurement can be more accurate if obtained from an MRI as the tentorium is clearly visible. If the posterior fossa height is less than 30 mm, it suggests platybasia and basilar invagination. All criteria/lines mentioned above utilized some sort of relationship between the skull and the cervical spine. However, Clark station of the atlas uses the vertical atlantoaxial relationship to predict presence of basilar invagination [7]. It divides the axis into three equal parts on sagittal film, and if the anterior ring of atlas crosses the middle or caudal thirds, basilar invagination exists. In fact, this method had 83% sensitivity and 85% NPV in diagnosing the condition in the abovementioned study by Riew et al. The vertical atlantoaxial index also describes the relationship between the first two cervical vertebrae and was actually developed with the understanding that a subgroup of patients with basilar invagination may, indeed, have a problem of the atlantoaxial joints rather than atlanto-occipital ones [26]. Their subgroup of patients was treated with a posterior atlantoaxial joint release, distraction, graft, or cage placement and fixation and thus avoiding transoral surgery. In anteroposterior view, the atlantooccipital joint axis angle formed by lines parallel to those joints is useful in detecting occipital condyle hypoplasia, with normal values ranging 124°–127° and increasing to even over 180° with the condition [45]. The majority of the above-mentioned parameters apply to non-traumatic situations and mostly to ­rheumatoid patients with CVJ involvement. The

3  Special Radiology

pathophysiology of the process results in cranial settling/basilar invagination. However, in the setting of trauma, highly sensitive and specific criteria are essential to detect atlanto-occipital and atlantoaxial dislocations and instabilities that are commonly missed. These injuries result in distraction, translation, or rotation between the individual components of the CVJ, making the above criteria obsolete for their use in most traumatic scenarios. However, multiple indices and ratios have been described to alert physicians to usually subtle and often missed signs of injury at the CVJ. Starting with survivors of atlanto-occipital dissociations, the most commonly used craniometric parameters include Harris’ measurements, Powers ratio, Lee’s lines, and occipital condyle displacement. Harris et al. derived the so called “Rule of Twelve” in a normal adult spine (Fig. 3.7). He used the ­basion-dental interval (BDI) previously described by Wholey et  al. [57] and basion-axial (BAI) interval (distance between basion and posterior C2 tangent) to describe normal values in adults and children [17]. It was found, that in adults, the CVJ is intact if BDI and BAI are less than 12 mm. The BDI cannot be applied to children in whom the dens ossification has not been completed. Harris et al. then applied those rules retrospectively to analyze plain radiographs of 37 patients previously diagnosed with AOD on the basis of Lee X-lines, Powers ratio, and BDI/BAI [18]. In all 23 patients with frank AOD, the BDI and BAI were both greater than 12  mm. Neither Powers ratio nor Lee X-lines could be applied to 46% of patients due to anatomical variants. In the remaining 20 patients, Powers ratio detected the correct type of atlantooccipital subluxation in only 60% and Lee X-lines in only 20% of cases. The types of atlanto-occipital subluxations identified were purely anterior, purely distracted, combined anterior and distracted, and purely posterior. The in specificity inferior Powers ratio is calculated as a distance between basion and posterior C1 arch divided by the distance between anterior C1 arch and opisthion [35]. It is considered normal when the value is less than 1 (Fig. 3.7). However, it will remain normal in purely vertical distraction and posterior types of AOD, thus resulting in delay in diagnosis. Lee X-lines [27] were applied to 12 cases of AOD as well as 100 normal patients and compared with Powers ratio, Wholey’s BDI, the Dublin method, and direct atlanto-occipital joint measurement. The Lee X-line method was the most accurate on plain radiograph assessment and correct in 75% of cases; the other methods were accurate

29

3.1  Radiographic Data Analysis Fig. 3.7  Different measurements of CVJ and UCS alignment

BAI – basion-posterior axial line interval BDI – basion-dental interval

Basion

B

Anterior arch. of C1

A

ADI – anterior atlanto-dental interval

O C

Opisthion

Posterior arch. of C1

PADI – posterior atlanto-dental interval

in 50% or less. Ultimately, all cases were confirmed by 2 mm cut CT with reconstructions. The X-line method is considered abnormal if the basion-axis spinolaminar junction line does not intersect C2 and if the opisthionaxis line does not cross C1. According to Wholey et al., the direct measurement of the width of the ­atlantooccipital joints (condyle-C1 interval – CCI, atlantooccipital interval – AOI) should be less than 2 mm in adults and 5 mm in children [57].

3.1.3 Atlanto-Axial Parameters The atlanto-axial relationship can be described by means of an anterior atlantodental interval (AADI, predental space) on plain radiographs [20]. Closely related is the posterior atlantodental interval (PADI) (Fig. 3.7). Both have been defined validated on dynamic plain films of patients with rheumatoid arthritis, developmental anomalies, and/or chronic posttraumatic instabilities but traumatic atlantoaxial instability is relatively uncommon and thus no such validation exists in the setting of acute trauma [2]. The AADI should be less than 3  mm in adults and less than 5 mm in children on plain dynamic films. Posterior ADI of 14 mm or less in combination with abnormal AADI warrants an MRI evaluation, even in the absence of neurological signs and symptoms. The craniocervical relationships were thus restudied in the era of multi-slice CT and the previously mentioned

plain radiograph parameters challenged. Rojas et al. studied 200 patients with a multidetector row CT (MDCT) and re-examined Harris’ measurements, Powers ratio, as well as atlantodental and atlanto-occipital intervals [42]. They found significant variances from the previously accepted normal values. Firstly, they found no differences when comparing men and women. Secondly, the Rule of Twelve described above was valid but on CT those values were much tighter for BDI (<8.5 mm) but unreliable for BAI. Powers ratio was less than 0.9 when measured on MDCT, thus not significantly different. Pang et al. found the normal CCI to be less than 2 mm in 89 children [34] and recommended surgical treatment in patients with greater than 4 mm widening of CCI. The differences in craniometric values on CT and plain radiographs were probably a result of magnification, landmark visualization difficulties, and patient positioning. Similarly, when craniometric parameters for basilar invagination were validated on MRI, there were significant differences when compared to plain radiographic criteria [48]. Historically, one of the most commonly used indicators of atlantoaxial instability secondary to transverse ligament (TAL) injury was the rule of Spence [46], which states that if the combined overhang of the C1 lateral masses over C2 adds up to 7 mm or more, the TAL is likely disrupted (Fig. 2, Chap. 10). One has to be aware of the fact that the original study was done on cadaveric specimens and therefore, the number is actually greater than 8.1 mm when assessed on plain films [19]. Spence’s actual number of 6.9 mm can be resurrected in the era

30

of CT with direct measurement of the displacement on coronal reconstructions, although a direct assessment of TAL integrity can always be confirmed on MRI [9].

3.2 Dynamic Imaging Several mobile CVJ and UCS abnormalities require testing under loading, i.e., in dynamic conditions of flexion and extension, or rotation. These include, for example, reducible cranial settling, AA instability, os odontoideum, or fracture malunions. Just as for static films, several radiographic indices have been ­developed to describe dynamic relationships of structures of the CVJ. Abe et  al. proposed an atlantoaxial instability index, defined as a change of rate of space available for the spinal cord between flexion and extension [1]. They recommended surgical intervention for all patients with neurological deficit, or instability index greater than 20% or maximum space for cord of less than 14  mm. This is useful for sagittal instability but, for example, in patients with os odontoideum, the instability is multidirectional. Therefore, Watanabe attempted to define a sagittal plane rotation angle as well as the above-mentioned indices in 37 patients with AAI on dynamic tomograms and concluded that not only the instability index of more than 40% but also rotation angle of more than 20° is predictive of cord signs and

3  Special Radiology

symptoms [56]. The extent of instability as measure by any dynamic means in an awake patient can be an underestimate of that seen during anesthesia. The previously mentioned vertical atlantoaxial instability index (VAAII) as described by Kulkarni and Goel is a good indicator of vertical instability. Goel et al. successfully used the VAAII to predict the feasibility of a pure vertical instability reduction and fixation without transoral decompression in cases of vertically mobile instability [14]. Although the majority of atlantoaxial pathologies mentioned so far affect the complex in vertical or anteroposterior or coronal planes, the primary movement within the AA complex is that of rotation. Atlantoaxial rotatory fixation (AARF) represents a pathological phenomenon that gained a standardized definition based on dynamic imaging of the CVJ only in recent years. Pang et al. were exemplary in defining the norm of atlantoaxial rotation in normal subjects using dynamic CT scans in rotation [34]. They defined the dynamic relationship of C1 on C2 throughout the turn and found very little variance from the mean. When C1 rotates on C2, it crosses it at 0°, moves alone until 23° then moves faster than C2 up to 65° and then they move in unison beyond that angle with a maximal separation angle of 43°. They used a three-position CT to generate a physiological motion curve (Fig. 3.8) according to which they classified AARF into three subtypes based on a diagnostic paradigm [33]. Type I AARF represented a fixed deformity whereas

Fig. 3.8  Atlantoaxial rotatory fixation (AARF) is defined on the basis of a normal three position CT evaluation that generates a physiological curve of the atlantoaxial relationship change in rotation

31

3.3  Vascular Imaging

Fig. 3.9  Dynamic MRI. Neural compression that would otherwise not be obvious can be visualized on MRI in flexion and extension (a–d)

Type III was the least severe type with C1 crossing C2 axis with rotation to contralateral side. More recently, MRI evaluation of the CVJ was done in flexion and extension to further visualize soft tissue compression or changes in instabilities of the region. Epstein et al. was the first to use the term “dynamic” MRI [12] when she documented cord compression in flexion/extension MRI although the dynamic plain films failed to show osseous instability (Fig.  3.9). Gupta et  al. studied 25 patients with suspected CVJ abnormalities due to various reasons with MRI and CT performed in neutral, flexion, and extension positions. They found that dynamic MRI detected every cord compression under loading that was not apparent in neutral position [15]. Other authors have also utilized dynamic MRI to assess patients with RA and suspected atlantoaxial instability [22] and correlated it to

findings on plain radiographs. Krodel et al. even argued that dynamic MRI plays a role in surgical planning when treating patients with rheumatoid-related upper cervical instabilities [25]. Three out of his 11 patients did not demonstrate reduction of cord compression on dynamic MRI and thus required removal of the synovial proliferative tissue around the dens.

3.3 Vascular Imaging Conventional cerebral angiography, CTA, and MRA have been used to assess vasculature at the CVJ. CTA and MRA, in particular, have revolutionized preoperative evaluation of the vascular tree at the CVJ and obviated the need for invasive angiography (Fig. 3.10). With

Fig. 3.10  Vascular evaluation of CVJ vessels. (a) CTA is particularly useful in delineating the relationship between VA and bone. (b) MRA

32

3  Special Radiology

respect to surgical reconstruction techniques, one is particularly interested in anatomical variants of the vertebral arteries and their relationship to bony anatomy. The details of VA anatomy, its variants and relationships are discussed in detail in Chap. 1. Nonetheless, we would like to mention dynamic angiography that is performed in rare cases of dynamic vertebral artery compression at the CVJ due to bony spurs, AAI or cranial settling. Janeway et  al. described positional VA

occlusion with head turning in a patient with basilar impression diagnosed on dynamic angiography already in 1966 [21]. Our case (Fig.  3.11) demonstrates an example of V3 segment injury due to bone spur compression of the VA in rotation in a 16-year-old boy with recurrent posterior circulation strokes. Dynamic angiogram demonstrated vessel occlusion with head rotation. He was successfully treated by C1 partial laminectomy. Similar cases are described in the literature such as that

Fig.  3.11  Sixteen year old boy with recurrent posterior fossa strokes as seen on MRI (a). Complete VA occlusion on left head turn (b). CTA demonstrating posterior C1 arch bone spur

(c). Dynamic intraoperative angiogram with complete release of compression in rotation (d) (courtesy of T. Abruzzo MD, Mayfield Clinic, Cincinnati)

33

3.4  Our Preference

of a 34-year-old man with recurrent embolic strokes and vertebrobasilar insufficiency secondary to dynamic VA compression during head rotation [49]. He was also successfully treated by osseous decompression. Dynamic angiogram is a rare investigation required for only specific problems at the CVJ; however, it may be essential to diagnosis and management in those rare cases. Dumas et al. further assessed the use of dynamic MRA and CTA to evaluate VA compression during head rotation in healthy volunteers as well as an actual symptomatic patient [11]. However, these non-invasive, dynamic vascular imaging techniques have not been validated further.

in clinical situations. A combination of multiple parameters relevant to the clinical scenario should always be applied which can lead to further, more detailed evaluation if abnormal. In non-traumatic cases with disease processes that result in cranial settling, basilar impression/invagination, or platybasia, we find the following parameters the most useful: Wackenheim basal clival line, cervicomedullary and basal angles, McRae and Chamberlain lines, and the BDI. We select a combination of multiple parameters in questionable cases and sometimes we are forced to use a particular parameter due to limited visualization of structures at the CVJ. For all of them, mid-sagittal CT reconstructions have certainly made the process significantly easier.

3.4 Our Preference 3.4.2 Traumatic Cases Radiographic evaluation of disorders of CVJ is not complex but requires knowledge of patient’s clinical details to guide an appropriate selection of modalities for a particular clinical scenario. Rather than using all imaging modalities for everyone, we like to divide the patients in traumatic, degenerative/inflammatory, and neoplastic scenarios.

3.4.1 Developmental/Degenerative/ Inflammatory The majority of rheumatoid patients belong to this group. Most of them are referred based on clinical history of the disease with associated neck pain and/or neurological symptoms and have had plain radiographs obtained prior to the referral. We proceed with obtaining standard neutral, non-contrast MRI images, particularly in patients with neurological compromise. If bony relationships remain questionable on plain radiographs, we obtain a thin cut CT with sagittal, coronal, and threedimensional images. This is mandated prior to any surgical intervention at the CVJ. These images are also supplemented by dynamic flexion-extension lateral plain radiographs. The CT images as well as plain radiographs allow for direct electronic measurement of the earlier mentioned craniometric parameters for disease severity assessment as well as surgical planning. It is clear that no single parameter should be used in isolation for evaluation of the alignment of CVJ structures due to individual anatomical variations and differences

For patients with traumatic situations, basilar impression parameters are not very useful as our biggest concerns are: (1) distraction injuries of the occipito-cervical junction, (2) atlantoaxial instability secondary to bony or ligamentous injuries, (3) coronal alignment in injuries of atlas, and (4) sagittal alignment of axis/dens. At our institution, all victims of traumatic injuries are transported with spinal precautions and with a hard collar, until radiographically and clinically cleared. In majority of emergently admitted trauma victims with suspicion of spinal injury (particularly if polytraumatized), we perform directly helical CT with appropriate reconstructions. Those referred from other hospitals or patients with isolated UCS trauma without neurological deficit are standardly assessed by plain radiography first. An MRI evaluation is often indicated in the acute setting unless partial spinal cord injury exists or an injury remains questionable based on mechanism of injury, clinical syndrome, and CT evaluation (Fig. 3.12).

3.4.3 Neoplastic Conditions In patients with suspected neoplastic disease of the CVJ, a contrast enhanced MRI is added to the above modalities. As tumors (metastatic or primary) commonly result in bony erosion and instability, relevant indices described above still apply and thus require a careful evaluation of bony structures by CT, stability by dynamic films, and neural compression by MRI. Occasionally, with

34

3  Special Radiology

Fig.  3.12  MRI depicting posttraumatic extra medullary ­intraspinal hematomas. (a, b) clinically silent posttraumatic ­subdural hematoma of CVJ. (c, d) initially clinically silent but

later decompensated epidural midcervical posttraumatic hematoma in another patient

invasive tumors, vascular invasion/occlusion needs to be evaluated by CTA/MRA or DSA. In conclusion, in the modern era of continuous improvement of imaging modalities, many craniometric parameters have become obsolete. However, their knowledge is important for the CVJ surgeon as they define “normal” relationships of the complex joints of the CVJ and UCS.

References   1. Abe, H., Tsuru, M., Mitsumori, K.: Atlanto-axial dislocation–instability index and indications for surgery (author’s transl). No Shinkei Geka 4, 57–72 (1976)   2. Bono, C.M., Vaccaro, A.R., Fehlings, M., et al.: Measurement techniques for upper cervical spine injuries: consensus statement of the Spine Trauma Study Group. Spine (Phila Pa 1976) 32, 593–600 (2007)

References   3. Boudin, G., Fischgold, H., Pepin, B., et al.: Syringomyelitic syndrome and associated complex malformations, especially of the atlanto-occipital joint and of the cervical vertebrae. Rev Neurol (Paris) 87, 347–352 (1952)   4. Brenner, D., Elliston, C., Hall, E., et al.: Estimated risks of radiation-induced fatal cancer from pediatric CT. AJR Am J Roentgenol 176, 289–296 (2001)   5. Bundschuh, C., Modic, M.T., Kearney, F., et al.: Rheumatoid arthritis of the cervical spine: surface-coil MR imaging. AJR Am J Roentgenol 151, 181–187 (1988)   6. Chamberlain, W.E.: Basilar impression (platybasia): A bizarre developmental anomaly of the occipital bone and upper cervical spine with striking and misleading neurologic manifestations. Yale J Biol Med 11, 487–496 (1939)   7. Clark, C.R., Goetz, D.D., Menezes, A.H.: Arthrodesis of the cervical spine in rheumatoid arthritis. J Bone Joint Surg Am 71, 381–392 (1989)   8. Deliganis, A.V., Mann, F.A., Grady, M.S.: Rapid diagnosis and treatment of a traumatic atlantooccipital dissociation. AJR Am J Roentgenol 171, 986 (1998)   9. Dickman, C.A., Greene, K.A., Sonntag, V.K.: Injuries involving the transverse atlantal ligament: classification and treatment guidelines based upon experience with 39 injuries. Neurosurgery 38, 44–50 (1996) 10. Dolan, K.D.: Cervicobasilar relationships. Radiol Clin North Am 15, 155–166 (1977) 11. Dumas, J.L., Salama, J., Dreyfus, P., et al.: Magnetic resonance angiographic analysis of atlanto-axial rotation: anatomic bases of compression of the vertebral arteries. Surg Radiol Anat 18, 303–313 (1996) 12. Epstein, N.E., Hyman, R.A., Epstein, J.A., et al.: “Dynamic” MRI scanning of the cervical spine. Spine (Phila Pa 1976) 13, 937–938 (1988) 13. Gilbert, J.W., Wheeler, G.R., Lingreen, R.A., et  al.: Open stand-up MRI: a new instrument for positional neuroimaging. J Spinal Disord Tech 19, 151–154 (2006) 14. Goel, A., Shah, A., Rajan, S.: Vertical mobile and reducible atlantoaxial dislocation. Clinical article. J Neurosurg Spine 11, 9–14 (2009) 15. Gupta, V., Khandelwal, N., Mathuria, S.N., et al.: Dynamic magnetic resonance imaging evaluation of CVJ abnormalities. J Comput Assist Tomogr 31, 354–359 (2007) 16. Guyot, L., Richard, O., Adalian, P., et al.: An anthropometric study of relationships between the clival angle and craniofacial measurements in adult human skulls. Surg Radiol Anat 28, 559–563 (2006) 17. Harris Jr., J.H., Carson, G.C., Wagner, L.K.: Radiologic diagnosis of traumatic occipitovertebral dissociation: 1. Normal occipitovertebral relationships on lateral radiographs of supine subjects. AJR Am J Roentgenol 162, 881–886 (1994) 18. Harris Jr., J.H., Carson, G.C., Wagner, L.K., et al.: Radiologic diagnosis of traumatic occipitovertebral dissociation: 2. Comparison of three methods of detecting occipitovertebral relationships on lateral radiographs of supine subjects. AJR Am J Roentgenol 162, 887–892 (1994) 19. Heller, J.G., Viroslav, S., Hudson, T.: Jefferson fractures: the role of magnification artifact in assessing transverse ligament integrity. J Spinal Disord 6, 392–396 (1993) 20. Hinck, V.C., Hopkins, C.E.: Measurement of the atlantodental interval in the adult. Am J Roentgenol Radium Ther Nucl Med 84, 945–951 (1960)

35 21. Janeway, R., Toole, J.F., Leinbach, L.B., et  al.: Vertebral artery obstruction with basilar impression. An intermittent phenomenon related to head turning. Arch Neurol 15, 211– 214 (1966) 22. Karhu, J.O., Parkkola, R.K., Koskinen, S.K.: Evaluation of flexion/extension of the upper cervical spine in patients with rheumatoid arthritis: an MRI study with a dedicated positioning device compared to conventional radiographs. Acta Radiol 46, 55–66 (2005) 23. Kevles, B.: Body imaging. How doctors learned to peer beneath our skin to see what might be wrong without using surgery. Newsweek 130, 74–76 (1997) 24. Klaus, E.: Roentgen diagnosis of platybasia & basilar impression; additional experience with a new method of examination. Fortschr Geb Rontgenstr Nuklearmed 86, 460– 469 (1957) 25. Krodel, A., Refior, H.J., Westermann, S.: The importance of functional magnetic resonance imaging (MRI) in the ­planning of stabilizing operations on the cervical spine in rheumatoid patients. Arch Orthop Trauma Surg 109, 30–33 (1990) 26. Kulkarni, A.G., Goel, A.H.: Vertical atlantoaxial index: a new craniovertebral radiographic index. J Spinal Disord Tech 21, 4–10 (2008) 27. Lee, C., Woodring, J.H., Goldstein, S.J., et al.: Evaluation of traumatic atlantooccipital dislocations. AJNR Am J Neuroradiol 8, 19–26 (1987) 28. McGregor, M.: The significance of certain measurements of the skull in the diagnosis of basilar impression. Br J Radiol 21, 171–181 (1948) 29. McRae, D.L., Barnum, A.S.: Occipitalization of the atlas. AJR Am J Roentgenol 70, 23 (1953) 30. Metzger, J., Fischgold, H., Appel, L.: Special radiographical examinations in fractures of the base of the skull. Ann Med Leg Criminol Police Sci Toxicol 42, 5–12 (1962) 31. Monu, J., Bohrer, S.P., Howard, G.: Some upper cervical spine norms. Spine (Phila Pa 1976) 12, 515–519 (1987) 32. Oliva, L.: Alessandro Vallebona (1899–1987). Radiol Med 76, 127129 (1988) 33. Pang, D.: Atlantoaxial rotatory fixation. Neurosurgery 66, A161–A183 (2010) 34. Pang, D., Nemzek, W.R., Zovickian, J.: Atlanto-occipital dislocation: part 1 – normal occipital condyle-C1 interval in 89 children. Neurosurgery 61, 514–521 (2007). discussion 521 35. Powers, B., Miller, M.D., Kramer, R.S., et  al.: Traumatic anterior atlanto-occipital dislocation. Neurosurgery 4, 12–17 (1979) 36. Raju, T.N.: The Nobel chronicles. 1979: Allan MacLeod Cormack (b 1924); and Sir Godfrey Newbold Hounsfield (b 1919). Lancet 354, 1653 (1999) 37. Ranawat, C.S., O’Leary, P., Pellicci, P., et al.: Cervical spine fusion in rheumatoid arthritis. J Bone Joint Surg Am 61, 1003–1010 (1979) 38. Redlund-Johnell, I., Pettersson, H.: Radiographic measurements of the cranio-vertebral region. Designed for evaluation of abnormalities in rheumatoid arthritis. Acta Radiol Diagn (Stockh) 25, 23–28 (1984) 39. Reijnierse, M., Bloem, J.L., Dijkmans, B.A., et al.: The cervical spine in rheumatoid arthritis: relationship between neurologic signs and morphology of MR imaging and radiographs. Skeletal Radiol 25, 113–118 (1996)

36 40. Riederer, S.J.: MR imaging: its development and the recent Nobel Prize. Radiology 231, 628–631 (2004) 41. Riew, K.D., Hilibrand, A.S., Palumbo, M.A., et  al.: Diagnosing basilar invagination in the rheumatoid patient. The reliability of radiographic criteria. J Bone Joint Surg Am 83-A, 194–200 (2001) 42. Rojas, C.A., Bertozzi, J.C., Martinez, C.R., et  al.: Reassessment of the craniocervical junction: normal values on CT. AJNR Am J Neuroradiol 28, 1819–1823 (2007) 43. Sassard, R., O’Leary, J.P.: Egas Moniz: pioneer of cerebral angiography. Am Surg 64, 1116–1117 (1998) 44. Scher, A., Vambeck, V.: An approach to the radiological examination of the cervico-dorsal junction following injury. Clin Radiol 28, 243–246 (1977) 45. Smoker, W.R.: CVJ: normal anatomy, craniometry, and congenital anomalies. Radiographics 14, 255–277 (1994) 46. Spence Jr., K.F., Decker, S., Sell, K.W.: Bursting atlantal fracture associated with rupture of the transverse ligament. J Bone Joint Surg Am 52, 543–549 (1970) 47. Suzuki, F., Fukami, T., Tsuji, A., et  al.: Discrepancies of MRI findings between recumbent and upright positions in atlantoaxial lesion. Report of two cases. Eur Spine J 17(Suppl 2), S304–S307 (2008) 48. Tassanawipas, A., Mokkhavesa, S., Chatchavong, S., et al.: Magnetic resonance imaging study of the craniocervical junction. J Orthop Surg (Hong Kong) 13, 228–231 (2005) 49. Tominaga, T., Takahashi, T., Shimizu, H., et al.: Rotational vertebral artery occlusion from occipital bone anomaly: a rare cause of embolic stroke. Case report. J Neurosurg 97, 1456–1459 (2002)

3  Special Radiology 50. van Gilder, J.C., Menezes, A.H., Dolan, K.A.: Radiology of the CVJ and its abnormalities, pp. 29–68. Futura Mount Kisco, New York (1987) 51. Vates, G.E., Wang, K.C., Bonovich, D., et al.: Bow hunter stroke caused by cervical disc herniation. Case report. J Neurosurg 96, 90–93 (2002) 52. Wackenheim, A.: Radiologic diagnosis of congenital forms, intermittent forms and progressive forms of stenosis of the spinal canal at the level of the atlas. Acta Radiol Diagn (Stockh) 9, 759–768 (1969) 53. Wang, C., Wang, S.: Letter to the Editor concerning “The single transoral approach for Os odontoideum with irreducible atlantoaxial dislocation” by Wang X, Fan CY, Liu ZH, Eur Spine J. 2009 Jul 14. Eur Spine J 19, 502–504 (2009). author reply 505–507 54. Wang, C., Yan, M., Zhou, H.T., et  al.: Open reduction of irreducible atlantoaxial dislocation by transoral anterior atlantoaxial release and posterior internal fixation. Spine (Phila Pa 1976) 31, E306–313 (2006) 55. Watanabe, I.S., Madeira, M.C., Watanabe, I.S., et  al.: The clivus-sphenoidale angle in children (postnatal flexion of the cranial base). Rev Bras Pesqui Med Biol 10, 325–329 (1977) 56. Watanabe, M., Toyama, Y., Fujimura, Y.: Atlantoaxial instability in os odontoideum with myelopathy. Spine (Phila Pa 1976) 21, 1435–1439 (1996) 57. Wholey, M.H., Bruwer, A.J., Baker Jr., H.L.: The lateral roentgenogram of the neck; with comments on the atlantoodontoid-basion relationship. Radiology 71, 350–356 (1958)

Section Principles of Reconstruction Techniques

II

4

surgical approaches P. Suchomel, J. Hradil, and R. Fricˇ

In the majority of procedures in CVJ, the surgical field should lie slightly higher or at the level of right cardiac atrium. The operating table should allow position changes in up and down directions in case of uncontrollable bleeding or, conversely, to prevent possible venous air embolism. CVJ region, intentionally set higher than heart, represents predisposition to venous air embolism, and adequate precautions such as central venous line and transesophageal echocardiography monitoring should be considered. On the contrary, the lower position of the surgical field may predispose to increased venous bleeding from epidural venous plexuses and veins surrounding C2-roots. Majority of surgical procedures in UCS have to be performed under guidance with fluoroscopy. X-ray visibility of target bony structures must not be compromised by operating table or any other hardware. Surgical position should not interfere with anesthesiological equipment, especially with reinforced tubes securing airways. To retain preoperative stability but also to reduce possible UCS deformation, the head is often fixated in a Mayfield three-point clamp (Fig. 4.1). When performing a rigid fixation of the head, appropriate position of patient’s body has to be considered, particularly in cases where the body can counteract by its weight. When a controlled axial skeletal traction is needed, it is better to use a halo ring or other freely adaptable skull fixation clamp than the table-fixed Mayfield clamp. Last but not the least, the surgical position should allow the surgeon

P. Suchomel () and J. Hradil Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic R. Fricˇ Department of Neurosurgery, Rikshospitalet, Oslo University Hospital, Sognsvannsveien 20, 0027 Oslo, Norway

Fig.  4.1  Preparing of the patient for posterior UCS surgery. Notice the attached IOM electrodes

to operate in an ergonomic and physically comfortable position. The overall setting of the operation theater should enable the use of extensive surgical armamentarium, surgical microscope, and other devices that are commonly used in contemporary surgical techniques (C-arm, navigation workstation, electrophysiological monitoring etc.). Many surgical approaches to CVJ and UCS area have been used and myriad access variants used mostly by neurosurgeons to reach tumors, vascular anomalies, and other CVJ pathologies have been reported. Most of those intended to decompress the neural structures are however not suitable for reconstruction of the spine and CVJ. In the ­following text only those approaches suitable for spinal procedures will be described in detail.

4.1 Posterior Midline Approach This is the most traditional, the simplest, and truly the most often used approach to the region of UCS and CVJ. It is suitable for simple decompression of the

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_4, © Springer-Verlag Berlin Heidelberg 2011

39

40

Fig.  4.2  Artistic drawing of structures visible from posterior midline approach

neural structures (in trauma, Chiari malformation, etc.) and is also commonly used to access the spinal cord and ­structures in posterior fossa. The posterior midline approach (Fig. 4.2) is, in principle, the easiest and safest approach for different techniques of ­posterior fixation (see Chap. 6) and therefore, it is most ­frequently used for UCS reconstruction and/or stabilization.

4.1.1 Surgical Technique Patient positioning largely depends on the type of the surgery. The specific types of surgery require specific positions. To limit the risk of venous bleeding, so called “landing Concorde” setting with the lowered position of the body and lower limbs flexed in knees can be of advantage. The “braking horse” position with straight neck and flexed UCS performed under lateral fluoroscopy enables an adjustment of the correct angle for C1-2 transarticular screw fixation. Nevertheless, no general rules can be given and surgeon’s individual preference and experience often plays the most important role. The posterior midline incision typically extends from the inion to the spinous process of C3. The surgeon advances along the nuchal ligament, preferably using a monopolar electrocautery. The dissection of the upper portion of the wound does not vary from that in cranial procedures, detaching nuchal insertions to the posterior skull base to achieve a proper exposure of the occipital squama. The external occipital protuberance is identified. SNL and INL muscle attachments are ­dissected and retracted laterally using bended self-retaining (“posterior fossa”) retractors. Continuing along the midline

4  surgical approaches

caudally, the spinous process of C2 is, usually, clearly palpable. Subperiostal exposure of lateral walls of C2 lamina allows the introduction of the second angled retractor caudally, e.g., opposite to the first one. The posterior atlantal tubercle is the most important anatomical landmark. Its muscle attachments are sharply cut off and subperiostal dissection continues laterally, first along the inferior border of C1 lamina. This is a very important step to avoid potential injury to VA in C1 posterior groove, especially if bone ponticuli are present as described in Chaps. 1 and 6. The extent of the exposure and dissection of muscular attachments should be limited only to expose the desired target structure. Muscle connections to the spinous process of the C2 are biomechanically important. However, for most open procedures and surgical techniques in this area, the dissection of muscular detachments is unavoidable.

4.2 Posterior Paramedian Approach This route is used mostly in minimally invasive procedures in the subaxial spine. There are not many indications for this approach in the region of the UCS. However, as minimally invasive techniques become more common, lateral paramedian incisions can be used for screw introduction via tubular retractors, particularly in case of percutaneous surgical techniques (Figs. 7.6 and 7.7; Chap. 7).

4.3 Lateral Approaches There are several versions and numerous modifications of lateral and posterolateral approaches such as farlateral or extreme-lateral. In neurosurgery, far-lateral approach stands for a low suboccipital approach that extends up to the occipital condyle and atlas, but it does not include removal of these structures [32]. Its caudal extensions merge with procedures designed specifically for the UCS.

4.3.1 Posterolateral Approaches These approaches are primarily designed for decompressive procedures and tumor resections. The surgery

4.3  Lateral Approaches

should not destabilize the UCS. The anatomy is complex and the course of the VA represents a major obstacle. Different surgical techniques have been described in detail, most of them introduced by neurosurgeons [3, 9, 22, 23, 54, 61, 69]. However, once the surgical dissection affects the natural stability of the spine, the reconstruction may be troublesome. Resection of 50% of occipital condyle may increase the flexion/extension movement by 153%, lateral bending by 41%, and rotation by 28% [73]. All posterolateral reconstruction techniques involve unilateral occipitocervical fixation, which does not provide sufficient primary stability. On the other hand, a substantially better functional outcome may be achieved when the C0-C1 joints are adequately preserved [66]. Because of the issues mentioned above, we do not use posterolateral approach as a primary route to reconstruct the UCS.

4.3.2 Lateral Approach for C1-C2 Transarticular Fixation There are several early notes describing lateral approach to the region of UCS. Henry (1957) used a sternomastoid eversion in order to reach important structures of the UCS [31]. Whitesides (1966) described an approach designed for the UCS fusion [76]. Barbour (1973) gave description of a technique of transarticular C1-C2 fixation and reported it being used since 1956. As he noted, it is necessary to perform this approach bilaterally, because unilateral fixation is not sufficient [8]. DuToit and Blignaut employed Barbour’s technique in 1973 [20] and found it quite difficult to place the screws accurately. Several modifications were therefore suggested [20, 62].

4.3.2.1 Surgical Technique The patient is positioned supine. For the purpose of the approach, the head is turned away; however, it must be realigned into the neutral position before fixation. Original Barbour’s technique was very straightforward. The incision was an oblique line starting at the anterior border of the mastoid process, passing down over the palpable transverse process of C1 and running a little further behind the angle of the mandible [8]. Anterior margin of the insertion of sternocleidomastoid muscle was identified. Advancing medially,

41

the surgeon cleared fascial cover and exposed the transverse process of atlas. The accessory nerve, which passes in a posteroinferior direction, was displaced, the transverse process was partially resected, and paravertebral muscles cleared to expose the anterior aspect of the C1/C2 joint. The head was then realigned to the neutral position before C1-C2 transarticular screw fixation was performed. The author was concerned about too posterior dissection as it could easily lead to injury of the VA or surrounding venous plexuses. After detailed cadaver studies, Du Toit suggested several improvements. Transverse incision over the base of the mastoid with extension downward and curving anteriorly along a neck crease was preferred for better cosmetic result. Ear lobe was retracted ­anteriorly and detachment of the sternocleidomastoid muscle from its cranial insertion was suggested for better exposure. Roy-Camille [58] tried to avoid the muscle ­detachment and inserted screws from a retro-SCM approach. The authors admitted, nevertheless, frequent ­collisions of the drill and the trajectory of the screws with the mastoid process. In the original Du Toit’s setting, the transverse process of the C1 was exposed by advancement just in front of the anterior margin of the sternocleidomastoid muscle. Care should be taken to avoid injury to the greater auricular nerve and external jugular vein during the exposure of the anterior border of the sternocleidomastoid muscle. These structures are crossing obliquely in upward and forward directions. The accessory nerve is located 2–3 fingerbreadths below the tip of the mastoid, running in posteroinferior direction and it is hardly ever encountered during the dissection. Posterior belly of the digastric muscle is retracted anterosuperiorly in order to fully expose the tip of the transverse process of the atlas. The occipital artery and vein pass directly across the tip of the transverse process and should be mobilized and retracted. Prevertebral fascia covering the process is incised and subperiostally dissected along the anterior border of the transverse process up to the lateral mass of the atlas. Finally, the antero-lateral aspect of the lateral mass is identified and cleared for insertion of the screws. The head is realigned to the neutral position; a proper adjustment of the joint has to be achieved and secured by temporal fixation with a Kirschner wire passing through the C1/C2 joint. Drilling proceeds with an angle of 25° below the horizontal plane and 10° behind the coronal plane

42

Fig. 4.3  Schematic picture of C1-2 transarticular screw introduced from lateral approach

(Fig. 4.3). The authors used a special guiding device to achieve proper angles and to protect surrounding structures while drilling. Twenty degrees in posterior direction was considered as the maximum safe angle to avoid injury to the content of spinal canal. Navicular bone screw was inserted and the joint was curetted and packed with bone chips before final tightening. The same procedure had to be performed contralaterally in order to achieve a solid C1-C2 fixation.

4.3.2.2 Our Preference We do not use the previously described approach for UCS reconstruction because of its anatomical complexity and necessity for bilateral dissection. However, the practical knowledge of the lateral approach can surely be useful in certain indications, such as in case of radical tumor removal.

4.4 High Anterolateral Approach High anterolateral cervical approach is derived from the subaxial access described to reach C3-T1 spine [64]. DeAndrade and McNab [16] developed the cranial extension of standard approach to reach C1-2 area. McAfee et  al. suggested transecting the digastric

4  surgical approaches

muscle and resecting the submandibular gland in order to enlarge the previously described access [45]. In order to directly expose the atlantoaxial joints for intraarticular cartilage debridement and to allow the perpendicular transarticular C2 to C1 anterior screw introduction, Vacaro et al. performed extensive rightsided high cervical approach accompanied by additional smaller incision on the left side [71]. Several authors have advocated this approach (either mono or ­bilateral) for treatment of different pathologies [55, 63]. However, lesions extending over the C1 level are treated rarely [35]. Recently, this approach has been successfully used in endoscopic resection of the odontoid [77]. The anterior aspect of the UCS can also be approached by a route passing behind the carotid artery [31, 75]. As compared to the previously described “prearterial” route, it is less straightforward. Although the connecting arteries and veins can be spared, the medial dislocation of the neurovascular bundle can be difficult and this approach did not gain wide popularity.

4.4.1 Surgical Technique In case of simple decompression of the UCS, the head can be slightly rotated to the contralateral side in order to achieve a better exposure. However, neutral head position should be maintained when subsequent fusion is planned. The skin incision is located either submandibularly or vertically along the sternocleidomastoid muscle (STCM). Platysma is divided along its fibers, the anterior border of STCM is dissected, and the muscle is retracted laterally. Then the anterior surface of the spine can be approached between the neurovascular bundle medially and the pharyngeal wall laterally. The parotid gland containing facial nerve tree is retracted cranially, the digastric muscle is transected, and the facial vein and external carotid artery branches are ligated. Division of descending loop of ansa cervicalis from n. XII. allows cranial dislocation of hypoglossal nerve. Finally, the prevertebral fascia is sharply cut and the longus colli muscle subperiostally is exposed with the help of bipolar electrocautery. Adequate release of the longus colli muscle is necessary for a safe anchorage of wound retractors beneath the muscle. Elevation of pharyngeal wall in a cranial direction enables direct visibility of anterior C1 arch.

43

4.5  Transoral Approach

4.4.2 Our Preference The greatest advantage of high anterolateral cervical approach is its anatomical similarity to the anterior approach to subaxial cervical spine, which most of the spine surgeons are very familiar with, performing it on a daily basis. Furthermore, no potentially infected cavity is opened during this approach and the use of metal implants is therefore safe. Difficulty can be encountered in case of low position of the mandible or immobile degenerative spine. Postoperative swallowing difficulties are frequent, though most often temporary. At our institution, the high anterolateral approach is most frequently used in cases of UCS injury. The graft and anterior plate is typically used in dislocated hangman’s fracture to fixate C2/3 segment (Figs. 12.14 and 12.18, Chap. 12). When performing the odontoid screw fixation, we do not need a large exposure and oblique approach starting at the C4/5 level suits well. Nearly all pathologies of C2 vertebra can be treated by this retropharyngeal approach, including palliative resection of tumors, evacuation of inflammatory tissue, and/ or C1-2 or even C0-1-2 transarticular fixation in case of instability (Figs.  14.2 and 14.3, Chap. 14). Major drawbacks of this approach are the oblique view of the

Fig. 4.4  Artistic drawing depicting the limited oblique visibility of UCS during high anterolateral cervical approach

spine where estimation of the midline may be difficult (Fig.  4.4), impossibility to expose the clival region, and limited radicality in tumor resection. The exposure achieved by this approach is frequently compromised by the mandibular angle.

4.5 Transoral Approach Transoral surgery is defined as a procedure carried out through the oral cavity to gain access to anterior midline structures of CVJ and UCS. Under normal conditions, the surgeon should be able to expose the area from the lower rim of clivus cranially to the level of disk space C2-3 caudally. The simple transoral approach can be extended upwards by transsection of the maxilla or downwards by splitting the mandible. Currently, minimally invasive and endoscopic techniques are increasingly used.

4.5.1 Transoral-Traspharyngeal Approach The first documented transoral (TO) surgery was performed by Kanavel (1919) who removed a bullet located between anterior arch of atlas and a skull base [39]. Scoville and Sherman (1952) studied the transoral route on cadavers and recommended its clinical use for approach to the rim of FM [60]. Southwick and Robinson successfully performed transoral removal of C2 osteoma and evacuation of an abscess [64]. The first series of patients surgically treated by transoral route was presented by authors from Hong Kong when Fang and Ong (1962) reported six cases of posttraumatic C1-2 dislocations and inflammatory process in CVJ, respectively. Mullan et al. used the transoral route for removal of tumors in CVJ [51]. Sukoff et al. were the first who reported a successful transoral decompression of the spinal canal in a case of myelopathy caused by rheumatoid disease [68]. Numerous papers dealing with this problematic issue were published later [17, 33, 44, 56, 65]; most of them case reports and small series of patients. The main obstacles for a wider acceptance of the transoral approach were: (1) need for special instruments, (2) poor illumination, and (3) depth of the surgical field.

44

The risk of infection was a big concern particularly in cases when the subarachnoid space had to be opened. The initial enthusiasm vanished after reports of high frequency of meningitis and CSF fistulas, the complications with possible catastrophic sequelae. Some other causes of high morbidity/mortality were also reported [36, 38]. Only few surgeons continued to develop the technique of transoral approach, trying to reduce the frequency of complications [13–15, 28, 47, 48]. Crockard published his results from more than 350 transoral procedures in 1993 [13–15] and later shared the lessons learned from his vast experience [14]. He defined the pathologies indicated for transoral approach, described his technique, and emphasized possible risks. Frequency of infectious complications was reduced to less than 3% in his series. Hadley et al. reduced the perioperative mortality to zero [28]. Transoral approach is often necessary in case of irreducible CVJ deformity where anterior pressure to neural structures is present [48]. This situation may be caused by a wide variety of disorders. From the historical perspective, the majority of cases were patients suffering from rheumatoid arthritis where compression caused by rheumatoid pannus was further enhanced by posterior displacement of the odontoid. This indication has become less frequent today because of studies proving that atlantoaxial fusion alone can prevent not only the vertical migration of the odontoid [26] but also reduce the size of pannus (or even lead to its disappearance) [27, 49, 80, 82]. The other frequent indications for transoral route include developmental or acquired deformities where the anterior pressure cannot be reduced by simple reduction; namely, an infection of the odontoid and its surroundings can be a reason for TO intervention in order to evacuate the pus and debride the infected tissue. Also, tumors can be biopsied or resected via transoral route. Transoral decompression can induce significant multidirectional instability of UCS and/or CVJ. This applies particularly for cases with pre-existing partial instability due to the disease itself, complete odontoidectomy, or transsection of the atlantal ligament [18, 19]. Similarly, resection of the anterior arch of atlas may influence the translational and particularly the vertical stability of the CVJ [52]. The atlas loses its anterior tension band. Due to its wedge-shaped profile, the lateral masses separate horizontally under the

4  surgical approaches

vertical load caused purely by head’s weight and/or rotations in the C1-C2 joint. Such mechanism is probably often responsible for development and progression of basilar impression. This condition can occur after simple transoral odontoidectomy without fusion, but even after vertically unstable posterior fixation techniques such as with Lugue type rods fixed with wires [53]. Precautions of atlas settling are twofold: first, as recommended by Spetzler [65], not to resect the anterior arch completely. Second, modern stable posterior fixation constructs (plate/rod and screws) have to be used. There were attempts to stabilize the UCS anteriorly with plates in one session surgery [30, 41]. However, the biomechanical insufficiency of this fixation [40] combined with the risk of hardware infection speak in favor of UCS stabilization from posterior approach. Rare cases of complex surgeries, namely tumor resections, result in total destabilization of CVJ and require a reconstruction using a complex 360° fixation [57, 67]. Current opinion on transoral procedures favors direct extradural decompression which, however, is inappropriate for intradural pathology. Watertight dural closure is still an issue as potential risk of CSF leakage and consequent infection is unacceptably high.

4.5.1.1 Anatomical Background The transoral approach to the midline is generally very safe as there are no important structures interposed (Fig.  4.5). Nevertheless, a detailed knowledge of anatomy is mandatory, particularly when facing anatomical variations during surgery. The pharyngeal mucosa, constrictor muscles, prevertebral fascia, and anterior longitudinal ligament are overlying the target area. The thickness of posterior pharyngeal wall is approximately 4 mm above the level of C1 tubercle and 6  mm above the level of the lateral masses and central part of C2 [1]. Most of the authors use the midline splitting; however, the use of mucosal flap has also been recommended [59]. Midline approach to anterior aspect of UCS is safe. Once leaving the midline as in flap technique, we have to keep in mind that important structures must be protected by using a subperiostal dissection. Cranially, care has to be taken not to injure the XII nerve at its exit from the base of condyle and the jugular foramen. Variant position of carotid artery (deformed by subbasal kinking or coiling) can be very

45

4.5  Transoral Approach

4.5.1.2 Surgical Technique

Fig.  4.5  Artistic drawing of structures visible during simple transoral approach

treacherous (Fig.  6.8, Chap. 6). Anterior tubercle of the atlas is considered to be a crucial point for safe dissection. The attachments of longus colli and longus capitis muscles are less strong than the attachment of anterior longitudinal ligament here. Anterior arch of the atlas is approximately 30 mm long, 15  mm high, and 6  mm thick. Usual working space created by resection of the anterior arch is a little smaller, reaching about 12–15 mm. Synovial joint with fluid filled capsule can sometimes be encountered behind the arch. The odontoid process is usually 20 mm (15–25.4 mm) long and slightly tilted posteriorly. The diameter at its base (“waist”) is approximately 9 mm (7.8–14.1 mm) and the maximal diameter is 11  mm (8.4–14.1  mm). The expected distance of vertebral artery from the midline is approximately 25 mm at the level of C1 and 11–15 mm at the level of C2. Detachment of more or less damaged allar and ­apical ligaments is necessary to release the odontoid process. The crucial ligament, tectorial membrane, and dura mater are located behind the resected odontoid. While reaching the clival ridge one has to be aware of  venous sinus at its margin and possible venous bleeding.

Neurosurgeons tend to perform a decompression of neural structures first, followed by stabilization of the CVJ, while orthopedic surgeons usually prefer to work in reverse order. The extent of adequate exposure of anterior surface of the CVJ and USC very much depends on preoperative imaging and the extent of the pathological process to be treated. As emphasized above, all relevant radiological investigations should be performed before planning the TO procedure. Potential fixation points for the UCS stabilization have to be defined before the surgery also. Simple transoral procedure can be performed only if the patient’s orifice can be opened wide enough to allow insertion of the instruments. Minimal opening must be more than 2–3 cm [14, 48]. This is of special importance in RA patients in whom the motion of mandibular joints is often limited. The oral cavity must be free of infection, including possible dental focuses which have to be sanated. Prophylactic antibiotics should always be administered. Corticosteroids are given in order to prevent soft tissue swelling and potential secondary damage to the spinal cord. The majority of surgeons recommend intraoperative electrophysiological monitoring (IOM). Patients are often intubated awake with the help of fiberoptic guidance. Tracheostomy is performed only in very complex surgeries with expected difficult postoperative course [50]. Some surgeons fixate the head into three-point Mayfield type clamp, others use the “horse shoe” support only. Oral cavity should be disinfected very carefully. Different types of transoral distractors are available to open the mouth and to push down the tongue. The soft palate is often hindering the view. It can either be split in the midline, leaving the uvula on one side [48], or retracted up into epipharynx with the help of a stitch attached to a rubber tube inserted transnasally [14, 28]. The incision site is usually infiltrated with a mixture of local anesthetics and adrenalin. The optimal way of mucosal incision is still being debated. A better visibility of target structures, more lateral exposure, and easier wound closure advocate for use of broadbased mucosa-muscle flap [59]. This so called U-flap may be based either cranially [37] or caudally [43]. On the other hand, U-flaps do not allow vertical extension of the wound and more extensive lateral distraction of the wound is often needed. Wound healing can be

46

troublesome especially if ischemized during the surgery. Currently, most authors recommend a simple midline incision starting at the anterior tubercle of atlas [13–15, 28, 47, 48]. Lateral fluoroscopy is used to determine the target part of the bone and also for final estimation of the extent of resection. Filling the resection cavity with contrast medium can be helpful. Most surgeons sit behind patient’s head while operating. Use of microsurgical techniques is mandatory. The maximal safe lateral extent of subperiostal dissection is considered to be 40 mm at the level of C1 and 30 mm at the level of the base of C2 [1]. The anterior arch of the atlas is usually removed by high-speed drill and borders of the odontoid are identified. The egg-shell type resection of the odontoid is then performed starting either at its tip or its base. Posterior remnants of the odontoid are removed last. If dura has to be opened, watertight suture is supplemented with patch and glue. Postoperative lumbar drainage seems to be mandatory if the risk of CSF leakage and meningitis shall be minimized. Duration of external lumbar drainage recommended varies from only 4 days [28] to 10 days or more, as practiced by most authors [13–15, 47, 48]. Final wound closure is recommended to be performed in two layers, e.g., muscle and mucosa layer, but successful results after closure in just one layer has also been described. Postoperatively, the gastric tube is used for feeding for approximately 10 days. Depending on the type and the extent of surgery, some patients are left intubated for 2–3 days following the surgery. If indicated, posterior stabilization may be performed either in the same session or it may be postponed to a later time, while the patient is wearing a halo vest fixation in the meantime.

4.5.2 Extended Transoral Approaches Under normal anatomical circumstances, the classical transoral approach (with or without splitting of the soft palate) allows the surgeon to reach the lower edge of clivus in a cranial direction and the C2/3 disk space in a caudal direction. The lower third of clivus and part of anterior C3 body surface can sometimes be exposed in patients with large orifice. This variation may certainly be helpful in some patients, but it is wise to be prepared for a situation when the approach has to be extended. When extending the simple transoral route to more complex approaches to UCS and CVJ, several factors

4  surgical approaches

have to be taken into account, including patient’s specific anatomical variations, size, location, and biology of the lesion, whether en bloc or piecemeal resection is intended, and last but not the least, the surgeon’s own preference, most often given by his/her familiarity with the procedure. Anatomical studies documented that cranial or caudal extensions of transoral approach not only increase the surgical field in vertical direction but also make the operative field more superficial [81]. On the other hand, a need for more extensive dissection and potential surgical injury to structures such as dental system possesses an increased risk of complications and/or worse functional outcome.

4.5.2.1 Transoral – Transmaxillar Approach By this technique, the dissection plane is extended cranially, which is necessary for surgical targets located above the level of the hard palate. It is the case in some primary or secondary basilar invaginations/impressions or in platybasia where obtuse-angled clivus anatomically elevates the FM cranially. Tumors with cranial extension, typically chordomas, may require a surgical route providing the approach to intact part of the clivus so that a radical resection can be achieved (Fig. 7.1, Chap. 7, Fig. 20.13, Chap. 20). Transfacial approaches had been used in maxillofacial surgery historically, but only in later years as a route to the upper two-thirds of the clivus [4]. To reach the clivus, Archer et al. [6] modified Le Fort I osteotomy (called as “Cheever’s operation”). Crockard [34] developed a complex technique to get an approach to lower clivus by adding a midline maxillary splitting to Le Fort type maxillotomy and called the procedure an “open-door maxilotomy”. In the largest published series of patients operated on by transoral route (with a remarkable number of developmental deformities), the maxillary extension of the approach was reported to be necessary in approximately 3% of cases [10].

4.5.2.2 Transoral – Transmandibular Approach This approach is only seldom used in adult patients in cases where a lesion, most frequently a tumor, extends from C2 downwards below the level of vertebra C3. Other indication is a limited opening of the mouth (less than 2  cm) in patients where no other approach is

4.5  Transoral Approach

47

possible. The transmandibular route is very old although it was originally developed for treatment of oropharyngeal malignancies; its adaptation for upper cervical spine surgery dates back to the early 1980s [5, 17]. Some of the authors use medial glossotomy [78, 81], some avoid it [12, 24, 25], but this decision is usually based on the character of the lesion and require careful preoperative radiological analysis. There are recent reports of this approach in children with complex developmental anomalies [11].

4.5.2.3 Our Preference We tend to perform decompression first, followed by stabilization procedure. Patients treated by transoral route suffer from lesions causing an anterior compression, which cannot be reduced by traction or posterior fixation alone. The indication becomes emergent in the presence or imminent danger of neurological deficit. In specific cases, transoral biopsy is the simplest way to obtain samples for biopsy (tumors, infection). All available standard radiological workouts are always performed. MRI is mandatory and CTA can give an important information regarding the course of vertebral and carotid arteries in selected cases. Careful inspection of the mouth, extent of orifice opening, and mobility of cervical spine is always examined before transoral surgery. Specific investigations can be performed, such as simulation of reachable areas under fluoroscopy in an awake patient. This is performed with a blunt metal rod after the dorsal part of the tongue and oropharynx is locally anesthetized. Although this procedure can be uncomfortable for the patient, it is the most objective and valuable way of estimating the achievable extension of the approach in a specific patient (Fig.  4.6). Under fluoroscopy, possible asymptomatic hyperextension of UCS can also be tested preoperatively. We find this method safer and more reliable than complicated calculations. IOM is almost always a necessity in UCS surgery; ventral aspect of the medulla lies close to the surgical field and MEPs are therefore more useful than SSEPs. Electrophysiology gives the surgeon vital information and confidence in certain situations. It is also useful to control the effect of anterior spinal decompression and it allows a safe rotation of the patient prior to posterior fusion. Airways are secured via orotracheal tube. Tracheotomy is reserved for complex surgeries or cases with high probability of prolonged respirator-

Fig. 4.6  Preoperative fluoroscopical testing of transorally reachable areas in awake patient

assisted ventilation postoperatively. We use a local surface application of naphazoline in order to achieve vasoconstriction in nasal and oral mucosa, and local cortisone ointment to prevent swelling of lips. The antibiotic prophylaxis according to hospital’s policy is administered before anesthesia. The transoral intubation with reinforced tube is performed with or without fiberoptical guidance. In patients with marked compression and/or neurologic deficit, awake intubation is preferred. Although it is better tolerated in awake patients, we do not use nasotracheal intubation as the tube would cross the surgical field and possibly limit lateral extension of the field, as well as increase the risk of infection. The head is positioned onto “horseshoe” pads for cases of simple decompression. Mayfield clamp is always used when stabilization is planned in the same session (Fig. 4.7). The positioning of the head is very important. Slight extension of the cervical spine may enhance the surgical access to the clival edge. Possible craniocaudal extension of TO approach is tested again using a metal rod and a fluoroscope prior to fixation in the final position. Fluoroscopical visibility of anterior atlantal tubercle and other important anatomical landmarks have to be checked. Simple laryngoscope can be helpful when disinfecting the oral cavity thoroughly. The incisives are protected with silicon rubber half-tube. In patients with irregular dentition, it might be difficult to anchor the distraction frame cranially. Polymethacrylate template of upper jaw made preoperatively can make the maxillary anchorage easier. After the first disinfection and circumferential

48

4  surgical approaches

Fig. 4.9  Final setup for microsurgical transoral odontoidectomy Fig. 4.7  Head fixed to Mayfield holder with inserted Crockard’s mouth distractor. Notice the attached DRA for navigational system

wrapping, we insert Crockard’s transoral distraction frame (Codman™). The insertion and stability of the frame is crucial for uneventful surgery and it is frequently secured by a submandibular support of its lower part. The tongue must not be interposed between the teeth and the lower blade of the distractor. The endotracheal tube is located laterally (on the right side in our setup) under the large caudal lingual distractor blade. The uvula is stitched to a rubber tubing inserted through nostril to epipharynx, and then everted cranially (Fig. 4.8). Additional isolated spatulas for reverting the soft palate are also available in the set. It is possible to split the soft palate in order

Fig.  4.8  Soft palate reverted to nasopharynx by traction of transnasally introduced cannula stitched to uvula

to get a better view cranially, no more than 1 cm can be achieved by this maneuver, however. If performed anyway, the uvula has to be left on one side. We always try to avoid any incision in soft palate as there is a high risk of postoperative velopalatal insufficiency causing rhinolalia and nasal alimentary regurgitation. We have never drilled away any part of the hard palate in our series of TO-treated patients. When necessary, opendoor maxillotomy offers a more convenient option. After final positioning of Crockard’s frame, we disinfect the surgical field again and perform the final wrapping (Fig.  4.9). The anterior tubercle of the atlas is palpated to localize the midline. Uvula may be used as an optional orientation point in case of rotatory dislocation of atlas. A vertically oriented longitudinal knife scratch is performed on the surface of retropharyngeal mucosa at the level of anterior atlantal tubercle and local anesthetic with adrenalin is applied. The injection distorts natural anatomy and marking with a knife scratch keeps surgeon’s eye on the midline. Infiltrated mucosa is incised above the tubercle and all the way to the bone. This incision avoids any anatomical structures of importance. The anterior tubercle of C1 must be dissected free as the crucial landmark. It is of utmost importance to be sure that the atlas is not rotated as this may substantially change the position of the tubercle! The muscle attachments but namely the firm attachment of anterior longitudinal ligament should be cut sharply off the tubercle with the sharp long-shaft knife. The anterior arch of the atlas is exposed bilaterally by subperiostal dissection (around 1  cm to both sides; VA is normally located more than 2 cm from the midline). We continue to dissect sharply caudally to the body of C2. In the exposure designed for the purpose

4.5  Transoral Approach

of odontoidectomy, the disk C2/3 represents the most caudal landmark. Its position may be verified with a dissector and palpation or fluoroscopy. If necessary, it is usually possible to expose approximately the upper half of C3 body caudally. Lateral extent of the exposure at the level of C2 is limited by the position of VA (10–15 mm from the midline at this level). Cranial dissection, in particular, should be performed carefully as only the anterior atlanto-occipital membrane protects the dura laterally between atlas and clivus. The clival edge has to be identified before exposed subperiostally. Caudal and cranial ridge of the atlas arch can then be dissected with a thin periostal elevator. The anterior C1 arch can then be removed easily either macroscopically with a rongeur (Fig.  4.10) or drilled out under microscope. The interlaminar distance of 12–15  mm allows full anterior exposure of the odontoid. If necessary, the approach can be safely extended laterally to the lateral masses of C1. Further steps depend on whether dens deformity caused by underlying pathology is present or not. If well demarcated, the apical ligaments (e.g., apical and allar) are sharply cut and resected before the odontoid is cut at its base so that PLL may be reached (Fig. 4.11a, b, c). The odontoid peg is then mobilized from its tip while elevated with a flat bone hook (Caspar’s osteophyte hook) until it breaks (Fig. 4.11d). The free fragment can thus be removed en bloc. The odontoid is sometimes poorly delimited or dislocated too deep or even behind clivus. In such

Fig. 4.10  Anterior arch of atlas grasped by rongeur before its removal

49

situations, we use an egg-shell milling with high-speed drilling through the tissue until the opposite cortical bone is reached. We often start the removal at the tip and continue caudally to avoid free movement of the apical fragment during drilling. The transversal atlantal ligament can be seen behind the peg finally. It is loosened in cases of RA or developmental deformities and can be then removed. However, it should be left in place if it is strong and does not cause compression of the spinal cord. Leaving the strong ligament in place helps to resist eventual distracting forces on atlas and its vertical movement. Removal of the odontoid is often satisfactory for adequate decompression. It is not necessary to remove the soft tissue in patients with RA or odontoid pseudoarthrosis if we plan a posterior stabilization later during the same surgery. Previous recommendation is not applicable in patients with developmental anomalies, infectious bone damage, and/or tumors. Depending on the type of procedure and pathology, the standard transoral approach should be appropriately modified: the atlas arch partially spared, the clivus edge cut out, the whole anterior axis removed (with or without cage or prosthesis replacing C2 body) etc. The approach can be extended cranially (maxillotomy) and/or caudally (mandibulotomy), but such an extension should better be a planned step rather than a result of decision made intraoperatively. The subarachnoid space has to be opened very exceptionally, mainly in case of tumors (clival chordoma). Whenever possible, we prefer the far lateral approach to remove midline intradural pathologies of CVJ. Watertight suture with use of dural substitutes is necessary. It might be technically difficult and adhesives and self-adhering patches are often applied. Use of external lumbar drainage is mandatory in cases of intradural procedure. However, the pressure of cerebrospinal fluid should be maintained positive, so that the content of oral cavity including potentially pathogenic microbial flora is not prompted to migrate intradurally as a result of negative pressure gradient. Under normal circumstances, we leave external CSF drainage for 7–10 days postoperatively. The decompressive procedure is finished by closing the pharyngeal wall in one or two layers. Nasogastric feeding tube is inserted at the end of surgery and left in place for 5–7 days. Most of our transoral surgeries involved a primarily unstable situations or such a situation was created by the decompression performed. Stabilization is therefore often necessary and we tend to perform it in a

50

4  surgical approaches

Fig.  4.11  Transoral odontoidectomy. (a) Demarking of the odontoid process. (b) High speed drill undercutting of the base. (c) Drilling of the odontoid base controlled by fluoroscopy. (d) Outward braking of the peg

single session, e.g., immediately after the closure of the transoral wound. It is of advantage to attach the Mayfield’s head clamp in a way that allows head fixation in both prone and supine positions without any need for replacement of the skull pins (Fig. 4.1). The motor-evoked responses must be carefully monitored and should not alter during repositioning of the patient. There are only few distinct situations where the extension of the classical TO approach makes sense, as example when the border of pathological process cannot be reached and/or the mouth cannot be opened wide enough to insert instruments. A need for cranial extension is more frequent. Maxillotomy may be used in adult patients with tumors (chordoma, sarcoma, chondroma, osteoblastoma, etc.) and/or with congenital/ acquired deformities causing flattening of the skull base (Fig. 7.1, Chap. 7, Fig. 20.13, Chap. 20). Mandibulotomy

without splitting of the tongue can be justified if the mouth cannot be opened enough. Most of tumors of the UCS are of metastatic origin and radical extirpation is seldom possible. For palliative tumor resection and stabilization, the high anterolateral approach is sufficient in the majority of cases. We always invite the maxillofacial surgeon to perform the extension of the approach.

4.5.3 Minimally Invasive Approaches to Retropharyngeal UCS Endoscopic techniques were brought to the area of UCS by neurosurgeons familiar with endoscopic and image-guided surgery of the brain. Veres et  al. used navigation-based technique for transoral surgeries in three patients. By using a halo vest during

51

References

a preoperative scanning, he elegantly overcame the problem of the shift due to mobility. The resulting accuracy was reported to be 1.5–3 mm [72]. Other authors have confirmed positive experience with image guidance for TO surgery [70, 74]. The main problem of image-guided techniques in spine surgery – the accurate registration of mobile vertebras as landmarks – has been solved by the use of fluoroscopy for in situ registration. These technologies represent a futuristic reality that is available already today, and they are helpful particularly in complex surgical cases. Endoscopy-assisted surgery was originally introduced to increase the visibility as well as illumination of the surgical field, but also to avoid soft palate split during standard TO approaches [21]. Increasing popularity of endoscopy and experience from using it in pituitary surgery allowed neurosurgeons to extend its use to the surgery of the clivus and the region of CVJ. Based on cadaveric study of Alfieri et al. [2], Kasam et al. resected the odontoid process in a 73-year-old woman, using a transnasally introduced endoscope [42]. They used binostril approach. The caudally based mucosal flap was located above the level of soft palate. Introduction of adapted long instruments was image guided and controlled with fluoroscopy. Their patient (suffering from RA) did well after the surgery and the extent of odontoidectomy was nicely documented by CT. The authors called the approach as “expanded endonasal” (EEA). Similarly, Hansen et al. performed a transnasal decompression in a patient with basilar invagination [29]. Wu et al. [79] performed endoscopic removal of odontoid in three patients (RA in two, trauma in one) using only one nostril and self-retaining holders. Transclival trajectory allows sparing of C1 arch, thus reducing the risk of extensive destabilization. To avoid infectious complications of transoral or transnasal surgery (transcavital approach), Wolinsky et al. [77] developed endoscopic transcervical imageguided odontoidectomy (ETO). The procedure was successfully performed in three patients with basilar impression, with only one complication (CSF leakage). The endoscope was introduced through adapted tubular retractor (METRx™, Medtronic) to the base of C2 in a similar fashion as screws for odontoid fixation. The anatomical landmarks were registered using C-arm fluoroscopy and following surgical steps were performed under image guidance. The authors were able to decompress CVJ by gradual drilling under endoscopic control. ETO technique was later used at the

same institution for successful treatment of four pediatric patients suffering from basilar invagination and cranial settling [46]. Although ETO and EEA are really minimally invasive techniques, they cannot be regarded as pure endoscopy because the endoscope is introduced via a tubular system or nostrils, parallel with drill and suction. Recently, Baird et  al. [7] compared all three endoscopic techniques in a cadaver study. Evaluating findings from nine cadavers, the authors found the average distance to the surgical target to be similar for all three methods – 94 mm by endonasal, 102 mm by transoral, and 100 mm by transcervical route. However, the approach angles necessary to reach the target structures differed significantly. The authors concluded that endoscopic transoral approach allows exposure of the largest surgical field. The transcervical route, certainly suitable for resection of the odontoid, does not allow safe resection of the lower clivus.

4.5.3.1 Our Preference The philosophy of minimally invasive approaches performed either from small incisions with the help of image guidance or endoscopically assisted clearly stand opposite to the techniques aiming at maximal exposure, represented by transpalatopharyngeal route with medial mandibuloglossotomy. Although we do not have personal experience with endoscopy in surgery of UCS and CVJ at our institution, we feel that there is a place for it in selected indications. Further development of minimally invasive techniques based on virtual image guidance can definitely be expected.

References   1. Ai, F., Yin, Q., Wang, Z., et al.: Applied anatomy of transoral atlantoaxial reduction plate internal fixation. Spine (Phila Pa 1976) 31, 128–132 (2006)   2. Alfieri, A., Jho, H.D., Tschabitscher, M.: Endoscopic endonasal approach to the ventral cranio-cervical junction: anatomical study. Acta Neurochir (Wien) 144, 219–225 (2002). discussion 225   3. al-Mefty, O., Borba, L.A., Aoki, N., et al.: The transcondylar approach to extradural nonneoplastic lesions of the craniovertebral junction. J Neurosurg 84, 1–6 (1996)   4. Alonso, W.A., Black, P., Connor, G.H., et  al.: Transoral transpalatal approach for resection of clival chordoma. Laryngoscope 81, 1626–1631 (1971)

52   5. Arbit, E., Patterson Jr., R.H.: Combined transoral and median labiomandibular glossotomy approach to the upper cervical spine. Neurosurgery 8, 672–674 (1981)   6. Archer, D.J., Young, S., Uttley, D.: Basilar aneurysms: a new transclival approach via maxillotomy. J Neurosurg 67, 54–58 (1987)   7. Baird, C.J., Conway, J.E., Sciubba, D.M., et al.: Radiographic and anatomic basis of endoscopic anterior craniocervical decompression: a comparison of endonasal, transoral, and transcervical approaches. Neurosurgery 65, 158–163 (2009). discussion 163–154   8. Barbour, J.R.: Screw fixation in fracture of the odontoid process. S Aust Clin 5, 20–24 (1971)   9. Bertalanffy, H., Seeger, W.: The dorsolateral, suboccipital, transcondylar approach to the lower clivus and anterior portion of the craniocervical junction. Neurosurgery 29, 815– 821 (1991) 10. Bhangoo, R.S., Crockard, H.A.: Transmaxillary anterior decompressions in patients with severe basilar impression. Clin Orthop Relat Res 359, 115–125 (1999) 11. Brookes, J.T., Smith, R.J., Menezes, A.H., et  al.: Median labiomandibular glossotomy approach to the craniocervical region. Childs Nerv Syst 24, 1195–1201 (2008) 12. Cocke Jr., E.W., Robertson, J.H., Robertson, J.T., et al.: The extended maxillotomy and subtotal maxillectomy for excision of skull base tumors. Arch Otolaryngol Head Neck Surg 116, 92–104 (1990) 13. Crockard, H.A.: Transoral approach to intra/extradural tumors. In: Sekhar, L.N., Janecka, I.P. (eds.) Surgery of cranial base tumors, pp. 225–234. Raven, New York (1993) 14. Crockard, H.A.: Transoral surgery: some lessons learned. Br J Neurosurg 9, 283–293 (1995) 15. Crockard, H.A., Johnston, F.: Development of transoral approaches to lesions of the skull base and craniocervical junction. Neurosurg Q 3, 61–82 (1993) 16. de Andrade, J.R., Macnab, I.: Anterior occipito-cervical fusion using an extra-pharyngeal exposure. J Bone Joint Surg Am 51, 1621–1626 (1969) 17. Delgado, T.E., Garrido, E., Harwick, R.D.: Labiomandibular, transoral approach to chordomas in the clivus and upper cervical spine. Neurosurgery 8, 675–679 (1981) 18. Dickman, C.A., Crawford, N.R., Brantley, A.G., et  al.: Biomechanical effects of transoral odontoidectomy. Neurosurgery 36, 1146–1152 (1995). discussion 1152–1143 19. Dickman, C.A., Locantro, J., Fessler, R.G.: The influence of transoral odontoid resection on stability of the craniovertebral junction. J Neurosurg 77, 525–530 (1992) 20. Du Toit, G.: Lateral atlanto-axial arthrodesis. A screw fixation technique. S Afr J Surg 14, 9–12 (1976) 21. Frempong-Boadu, A.K., Faunce, W.A., Fessler, R.G.: Endoscopically assisted transoral-transpharyngeal approach to the craniovertebral junction. Neurosurgery 51, S60–S66 (2002) 22. George, B., Laurian, C.: Surgical approach to the whole length of the vertebral artery with special reference to the third portion. Acta Neurochir (Wien) 51, 259–272 (1980) 23. Gilsbach, J.M., Eggert, H.R., Seeger, W.: The dorsolateral approach in ventrolateral craniospinal lesions. In: Voth, D., Glees, P. (eds.) Diseases in the cranio-cervical junction. Anatomical and pathological aspects and detailed clinical accounts, pp. 359–364. Gruyter, Berlin, New York (1987)

4  surgical approaches 24. Grime, P.D., Haskell, R., Robertson, I., et  al.: Transfacial access for neurosurgical procedures: an extended role for the maxillofacial surgeon. I. The upper cervical spine and clivus. Int J Oral Maxillofac Surg 20, 285–290 (1991) 25. Grime, P.D., Haskell, R., Robertson, I., et  al.: Transfacial access for neurosurgical procedures: an extended role for the maxillofacial surgeon. II. Middle cranial fossa, infratemporal fossa and pterygoid space. Int J Oral Maxillofac Surg 20, 291–295 (1991) 26. Grob, D.: Atlantoaxial immobilization in rheumatoid arthritis: a prophylactic procedure? Eur Spine J 9, 404–409 (2000) 27. Grob, D., Wursch, R., Grauer, W., et al.: Atlantoaxial fusion and retrodental pannus in rheumatoid arthritis. Spine (Phila Pa 1976) 22, 1580–1583 (1997). discussion 1584 28. Hadley, M.N., Spetzler, R.F., Sonntag, V.K.: The transoral approach to the superior cervical spine. A review of 53 cases of extradural cervicomedullary compression. J Neurosurg 71, 16–23 (1989) 29. Hansen, M.A., da Cruz, M.J., Owler, B.K.: Endoscopic transnasal decompression for management of basilar invagination in osteogenesis imperfecta. J Neurosurg Spine 9, 354–357 (2008) 30. Harms, J., Schmelze, R., Stolze, D.: Osteosynthesen im occipito-cervikalen Übergang vom transoralen Zugang aus, XVII SICOT World Congress Abstracts. Demeter Verlag, Munich (1987) 31. Henry, A.K.: Extensile exposure, pp. 58–74. E & S Livingstone, Edinburgh and London (1957) 32. Heros, R.C.: Lateral suboccipital approach for vertebral and vertebrobasilar artery lesions. J Neurosurg 64, 559–562 (1986) 33. Hitchcock, E., Cowie, R.: Transoral-transclival clipping of a midline vertebral artery aneurysm. J Neurol Neurosurg Psychiatry 46, 446–448 (1983) 34. James, D., Crockard, H.A.: Surgical access to the base of skull and upper cervical spine by extended maxillotomy. Neurosurgery 29, 411–416 (1991) 35. Jiang, L.S., Shen, L., Wang, W., et al.: Posterior atlantoaxial dislocation without fracture and neurologic deficit: a case report and the review of literature. Eur Spine J 28, 28 (2009) 36. Jones, R.E., Bucholz, R.W., Schaefer, S.D., et al.: Cervical osteomyelitis complicating transpharyngeal gunshot wounds to the neck. J Trauma 19, 630–634 (1979) 37. Jones, D.C., Hayter, J.P.: The superiorly based pharyngeal flap: a modification of the transoral approach to the upper cervical spine. Br J Oral Maxillofac Surg 35, 368–369 (1997) 38. Jones, D.C., Hayter, J.P., Vaughan, E.D., et al.: Oropharyngeal morbidity following transoral approaches to the upper cervical spine. Int J Oral Maxillofac Surg 27, 295–298 (1998) 39. Kanavel, A.B.: Bullet locked between atlas and the base of the skull: Technique for removal through the mouth. Surg Clin 1, 361–366 (1919) 40. Kandziora, F., Kerschbaumer, F., Starker, M., et  al.: Biomechanical assessment of transoral plate fixation for atlantoaxial instability. Spine (Phila Pa 1976) 25, 1555–1561 (2000) 41. Kandziora, F., Mittlmeier, T., Kerschbaumer, F.: Stagerelated surgery for cervical spine instability in rheumatoid arthritis. Eur Spine J 8, 371–381 (1999) 42. Kassam, A.B., Snyderman, C., Gardner, P., et  al.: The expanded endonasal approach: a fully endoscopic transnasal

References approach and resection of the odontoid process: technical case report. Neurosurgery 57, E213 (2005). discussion E213 43. Kingdom, T.T., Nockels, R.P., Kaplan, M.J.: Transoraltranspharyngeal approach to the craniocervical junction. Otolaryngol Head Neck Surg 113, 393–400 (1995) 44. Lee, S.T., Fairholm, D.J.: Transoral anterior decompression for treatment of unreducible atlantoaxial dislocations. Surg Neurol 23, 244–248 (1985) 45. McAfee, P.C., Bohlman, H.H., Riley Jr., L.H., et  al.: The anterior retropharyngeal approach to the upper part of the cervical spine. J Bone Joint Surg Am 69, 1371–1383 (1987) 46. McGirt, M.J., Attenello, F.J., Sciubba, D.M., et  al.: Endoscopic transcervical odontoidectomy for pediatric basilar invagination and cranial settling. Report of 4 cases. J Neurosurg Pediatr 1, 337–342 (2008) 47. Menezes, A.H.: Anterior approaches to the craniocervical junction. Clin Neurosurg 37, 756–769 (1989) 48. Menezes, A.H., VanGilder, J.C.: Transoral-transpharyngeal approach to the anterior craniocervical junction. Ten-year experience with 72 patients. J Neurosurg 69, 895–903 (1988) 49. Milbrink, J., Nyman, R.: Posterior stabilization of the cervical spine in rheumatoid arthritis: clinical results and magnetic resonance imaging correlation. J Spinal Disord 3, 308–315 (1990) 50. Mouchaty, H., Perrini, P., Conti, R., et al.: Craniovertebral junction lesions: our experience with the transoral surgical approach. Eur Spine J 18(Suppl 1), 13–19 (2009) 51. Mullan, S., Naunton, R., Hekmat-Panah, J., et al.: The use of an anterior approach to ventrally placed tumors in the foramen magnum and vertebral column. J Neurosurg 24, 536– 543 (1966) 52. Naderi, S., Crawford, N.R., Melton, M.S., et  al.: Biomechanical analysis of cranial settling after transoral odontoidectomy. Neurosurg Focus 6, e7 (1999) 53. Naderi, S., Pamir, M.N.: Further cranial settling of the upper cervical spine following odontoidectomy. Report of two cases. J Neurosurg 95, 246–249 (2001) 54. Oya, S., Tsutsumi, K., Shigeno, T., et  al.: Posterolateral odontoidectomy for irreducible atlantoaxial dislocation: a technical case report. Spine J 4, 591–594 (2004) 55. Park, S.H., Sung, J.K., Lee, S.H., et al.: High anterior cervical approach to the upper cervical spine. Surg Neurol 68, 519–524 (2007). discussion 524 56. Pasztor, E., Vajda, J., Piffko, P., et al.: Transoral surgery for craniocervical space-occupying processes. J Neurosurg 60, 276–281 (1984) 57. Rhines, L.D., Fourney, D.R., Siadati, A., et  al.: En bloc resection of multilevel cervical chordoma with C-2 involvement. Case report and description of operative technique. J Neurosurg Spine 2, 199–205 (2005) 58. Roy-Camille, R., de la Caffiniére, J.H., Saillant, G.: Les traumatismes du rachis cervical superieur C1-C2. Masson et Cie, Paris (1973) 59. Ruf, M., Melcher, R., Harms, J.: Transoral reduction and osteosynthesis C1 as a function-preserving option in the treatment of unstable Jefferson fractures. Spine (Phila Pa 1976) 29, 823–827 (2004) 60. Scoville, W.B., Sherman, I.J.: Platybasia, report of 10 cases with comments on familial tendency, a special diagnostic sign, and the end results of operation. Ann Surg 133, 496– 502 (1951)

53 61. Sen, C.N., Sekhar, L.N.: An extreme lateral approach to intradural lesions of the cervical spine and foramen magnum. Neurosurgery 27, 197–204 (1990) 62. Simmons, E.H., Du Toit Jr., G.: Lateral atlantoaxial arthrodesis. Orthop Clin North Am 9, 1101–1114 (1978) 63. Skaf, G.S., Sabbagh, A.S., Hadi, U.: The advantages of submandibular gland resection in anterior retropharyngeal approach to the upper cervical spine. Eur Spine J 16, 469– 477 (2007) 64. Southwick, W.O., Robinson, R.A.: Surgical approaches to the vertebral bodies in the cervical and lumbar regions. J Bone Joint Surg Am 39-A, 631–644 (1957) 65. Spetzler, R.F., Selman, W.R., Nash Jr., C.L., et al.: Transoral microsurgical odontoid resection and spinal cord monitoring. Spine (Phila Pa 1976 4, 506–510 (1979) 66. Steinmetz, M.P., Mroz, T.E., Benzel, E.C.: Craniovertebral junction: biomechanical considerations. Neurosurgery 66, A7–A12 (2010) 67. Suchomel, P., Buchvald, P., Barsa, P., et  al.: Single-stage total C-2 intralesional spondylectomy for chordoma with three-column reconstruction. Technical note. J Neurosurg Spine 6, 611–618 (2007) 68. Sukoff, M.H., Kadin, M.M., Moran, T.: Transoral decompression for myelopathy caused by rheumatoid arthritis of the cervical spine. Case report. J Neurosurg 37, 493–497 (1972) 69. Ture, U., Pamir, M.N.: Extreme lateral-transatlas approach for resection of the dens of the axis. J Neurosurg 96, 73–82 (2002) 70. Ugur, H.C., Kahilogullari, G., Attar, A., et al.: Neuronavigationassisted transoral-transpharyngeal approach for basilar invagination – two case reports. Neurol Med Chir (Tokyo) 46, 306–308 (2006) 71. Vaccaro, A.R., Lehman, A.P., Ahlgren, B.D., et al.: Anterior C1-C2 screw fixation and bony fusion through an anterior retropharyngeal approach. Orthopedics 22, 1165–1170 (1999) 72. Veres, R., Bago, A., Fedorcsak, I.: Early experiences with image-guided transoral surgery for the pathologies of the upper cervical spine. Spine (Phila Pa 1976 26, 1385–1388 (2001) 73. Vishteh, A.G., Crawford, N.R., Melton, M.S., et al.: Stability of the craniovertebral junction after unilateral occipital condyle resection: a biomechanical study. J Neurosurg 90, 91–98 (1999) 74. Vougioukas, V.I., Hubbe, U., Schipper, J., et al.: Navigated transoral approach to the cranial base and the craniocervical junction: technical note. Neurosurgery 52, 247–250 (2003). discussion 251 75. Whitesides, TEj: Lateral retropharyngeal approach to the upper cervical spine. In: Sherk, H.H., Dunn, E.J., Eismont, F.J., et al. (eds.) The cervical spine, pp. 796–804. Lippincott, Philadelphia (1989) 76. Whitesides, T.E., Kelly, R.P.: Lateral approach to the upper cervical spine for anterior fusion. South Med J 59, 879–883 (1966) 77. Wolinsky, J.P., Sciubba, D.M., Suk, I., et  al.: Endoscopic image-guided odontoidectomy for decompression of basilar invagination via a standard anterior cervical approach. Technical note. J Neurosurg Spine 6, 184–191 (2007) 78. Wood, D.E., Good, T.L., Hahn, J., et al.: Decompression of the brain stem and superior cervical spine for congenital/

54 acquired craniovertebral invagination: an interdisciplinary approach. Laryngoscope 100, 926–931 (1990) 79. Wu, J.C., Huang, W.C., Cheng, H., et al.: Endoscopic transnasal transclival odontoidectomy: a new approach to decompression: technical case report. Neurosurgery 63, ONSE92–ONSE94 (2008). discussion ONSE94 80. Young, W.F., Boyko, O.: Magnetic resonance imaging confirmation of resolution of periodontoid pannus formation following C1/C2 posterior transarticular screw fixation. J Clin Neurosci 9, 434–436 (2002)

4  surgical approaches 81. Youssef, A.S., Guiot, B., Black, K., et al.: Modifications of the transoral approach to the craniovertebral junction: anatomic study and clinical correlations. Neurosurgery 62, 145–154 (2008). discussion 145–154 82. Zygmunt, S., Saveland, H., Brattstrom, H., et al.: Reduction of rheumatoid periodontoid pannus following posterior occipito-cervical fusion visualised by magnetic resonance imaging. Br J Neurosurg 2, 315–320 (1988)

5

Basic Principles of Reconstruction Techniques O. Choutka and P. Suchomel

The craniovertebral junction (CVJ) is a mechanical part of the spine that offers the most significant amount of mobility when compared to other segments, particularly in flexion, extension, and rotation. Under physiological circumstances, stability and mobility of the CVJ is facilitated by unique morphology of the upper cervical vertebrae that form the levers in motion that are restrained by ligaments and facilitated by surrounding local and distant muscles. The atlanto-occipital and atlantoaxial joints act as pivots of the complex motion. Pathological processes, such as arthritides and tumors, as well as surgical decompressive procedures can result in profound violation of the balanced construct and thus cause instability, loss of function, pain, and neurological compromise. Prior to embarking on any potentially destabilizing procedure at CVJ, a surgeon should have a plan for reconstruction that will stabilize appropriately. The unique nature of the upper cervical spine (UCS) vertebrae offers an opportunity for not only stabilizing rigid constructs but also for a number of direct osteosynthetic designs that preserve motion of the segment. In general, constructs of the CVJ involve those designed for ventral approaches and posterior instrumentation, or both. Basic biomechanical principles and forces generated by any implant must be respected and understood when instrumenting the UCS and are ­covered elsewhere in the book. The biomechanical ­properties of the CVJ must be either

O. Choutka Department of Neurosurgery, University of Cincinnati College of Medicine, 231 Albert Sabin Way, Cincinnati, OH 45267-0515, USA P. Suchomel Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St.10, 46063 Liberec, Czech Republic

matched or appropriately counteracted by any construct that is to stabilize and maintain motion and appropriate alignment. Several basic reconstruction techniques are ­discussed in this section but the reader should refer to Chap. 2 for ­biomechanical principles and Chap. 6 for specific reconstructions.

5.1 Defect/Instability/Decompression Most of the axial rotation (60%) and some of the ­flexion-extension (40%) and lateral bending of the head occur in the UCS (C0-C2) [14, 30, 39]. The highly ­specialized anatomy and osteoligamentous integrity ­provides for a relatively paradoxical kinetic profile with  loose enough arrangement to allow for the ­above-mentioned range of motion but tight enough to prevent injury to spinal cord, nerves, and vertebral arteries. When the integrity is interrupted for any reason (trauma, tumor, inflammation, and degeneration or iatrogenic decompression), instability ensues. White and Panjabi defined clinical instability as “the loss of the ability of spine under physiologic loads to maintain relationships between vertebrae in such a way that there is neither initial nor subsequent damage to the spinal cord or nerve roots, and in addition, there is neither development of incapacitating deformity nor severe pain” [40]. The biomechanical profile of various types of instability affecting the UCS is described in Chaps. 2 and 3. Irrespective of the etiology of ­mechanical ­instability (acute vs. chronic) of the UCS, one has to be aware of the biomechanics involved in the development of the condition in order to be able to determine if a ­surgical construct is necessary to restore stability and balance to the region, and if so, what kind of construct can withhold the forces involved while fusion is taking place.

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_5, © Springer-Verlag Berlin Heidelberg 2011

55

56

5.2 Construct Design The concept of surgical spinal stabilization for a fracture was introduced in 1891 by Hadra [19] when he operated on a fracture dislocation of a cervical spine in a child showing progressive cord deterioration, using wires wrapped around the spinous processes to stabilize the vertebral sector. This was a new concept for the time of “holding broken or diseased bones together” and was widely adapted until 1911 when spinal fusion was first described as a treatment for Pott’s Disease [3,  23]. Albee and Hibbs worked independently on patients with Pott’s disease to design a method of spine fusion. Albee used tibial cortical autograft and Hibbs used spinous processes to produce surgical fusion. Spinal constructs since then have changed significantly but the concept of fixation and fusion remains the mainstay of treatment of spinal instabilities, including the UCS and CVJ. Screws, cages, and wiring/cable techniques are all used in this region and must be able to withhold the main forces with OA and AA joints, flexion/extension, and rotation, respectively. Multiple basic principles have been described for application of metallic implants in the UCS. Most commonly used implants at the CVJ include both anterior and posterior techniques and each individual clinical scenario determines the most appropriate approach. Anterior implants include direct osteosynthetic screw of the dens (lag) or cage constructs after odontoidectomy/corpectomy/spondylectomy or atlantoaxial transarticular screws. Posterior constructs can include occipitocervical and atlantoaxial fixations through means of various screw and rod/plate constructs or direct osteosynthetic screws or wire/cable techniques (Chap. 6). Buttressing, tension band, and neutralization principles usually apply to the constructs of the CVJ [2] with the primary goal being immediate rigid fixation so that favorable environment for bone fusion is created.

5  Basic Principles of Reconstruction Techniques

for use in cervical trauma patients [8]. His plating technique has gained wide popularity in the subaxial spine in particular and offered immediate stability without the use of external orthosis. The stability of an anterior plate, however, is dependent on screw purchase within the vertebral bone. The initial Caspar design was thought to be dependent on bicortical screw purchase to prevent screw toggling. However, this requirement was deemed unnecessary once locking screw plates were developed and unicortical screws were sufficient [6, 27]. Plate and screw constructs have been used in orthopedic trauma management of various long and short bone fractures well before the use in spine surgery and follow the tenets put forth by the AO  group in late 1950s [15]. The use of plate and screw constructs in the ventral UCS is limited to C2-3 fusion when done for treatment of hangman’s fracture [37], one can argue that certain anterior cage constructs also follow buttressing principle when used in combination with anterior screws such as demonstrated by the anterior clival-C3 construct in our patient with C2 chordoma resection [36] (Fig. 5.1). Posteriorly, buttressing principle is applied with use of plating systems with either screws or wires. Lateral mass screw and plate construct offers similar rigidity in

5.2.1 Plate and Screw Constructs in the CVJ Buttressing implants prevent axial deformity and are placed on the side of load application [2]. An example includes anterior cervical plate as developed by Caspar

Fig.  5.1  Sagittal CT demonstrating a complicated UCS cage/ screw construct after C2 spondylectomy. Anterior cage does not only serve as the major load-bearing apparatus for the anterior column but through its attachment to the clivus and C3 vertebral body acts as a buttress

57

5.2  Construct Design

biomechanical in vitro studies to other types of posterior cervical constructs [12]. Even further, the authors found no significant difference in rigidity between combined anteroposterior and posterior only constructs. It needs to be reiterated that the strength of a construct is more dependent on the screw than the plate itself. Magerl lateral mass screw technique is biomechanically superior to the Roy-Camille method, probably due largely to the difference in screw length and trajectory [10]. Posterior cervical plate constructs are a safe construct with low complication rate [21]. Nonetheless, failure can be related to either the plate or the screw. Screw bone interface (surface area) will influence the pullout strength of any given, but a tapered or conical screw configuration does not alter the pullout strength [17]. Wellman et al. examined the safety and complications associated with lateral mass screws in their 43 patients [38]. Although, a proponent of bicortical screw purchase in lateral mass screws to increase pullout strength, he concluded that bicortical screw purchase did not offer decrease biomechanical failure rate (none in his series) and therefore, was not worth the potential neurovascular risk. On the other hand, Heller et al. argued that engaging the far cortex increases the pullout strength by 28% [20] with less than 2% risk of radiculopathy [21]. The issue of bicortical vs. monocortical screw purchase is a common discussion point when it comes to the instrumentation of UCS and, perhaps, the key to differentiation between a bicortical screw and bicortical penetrating screw needs to be made (Fig. 5.2). The former offers the potential biomechanical advantages of a bicortical purchase whereas the latter may, in addition, increase the chance of neurovascular injury. It may be sensible to weigh the risk/benefit ratio in each clinical scenario and consider the true need for bicortical screw purchase in a good quality bone versus not achieving far cortex purchase in osteoporotic bone. The bicortical screw discussion surrounds also occipital screw [22] and odontoid screw [35] placement (indirectly with one vs. two screw conflict). Anterior odontoid screw fixation is a well-accepted method of direct, compressive, and osteosynthetic construct that has evolved over time with multiple variations all resulting in fracture line apposition, alignment and compression, thus creating a favorable bone-healing condition [1, 29]. All modifications of anterior direct osteosynthetic odontoid screw, such as fully threaded screw [7], cannulated K-wire guided

Fig. 5.2  Monocortical screws are at risk of toggling. Bicortical screw purchase increase the pullout strength in many constructs. Penetrating bicortical screws may, however, increase the risk of neurovascular injury. Nearly bicortical screw purchase may present an alternative

screw [1], and double-threaded screw [9, 25] utilize the same principle of fracture reduction, alignment, and compression. However, even a good reduction is not always feasible with anterior odontoid screw as up to 19% of cases end up malaligned [1]. The compressive lag screw effect can be achieved through a differential thread design, proximal overdrilling or a standard lag screw design (Figs. 5.3 and 5.4). When using a fully threaded screw, without overdrilling, the lag principle does not apply and the screw simply becomes a ­neutralizing/stabilizing one. This obviously creates different bone-healing conditions than compressive constructs. Neutralization does still provide stability through minimization of torsional bending and shearing but may undergo indirect bone healing (i.e., formation of a callus rather than going through tissue differentiation and resorption of bone surface [32]. Similar compressive, lag principle is applied to certain posterior techniques. Direct osteosynthesis of hangman’s fracture as described by Judet [26] is a classic example (Fig. 12.19, Chap. 12). We have applied this principle to a patient with a unilateral C1 lateral mass fracture with fracture displacement (Figs.  7.5– 7.8, Chap. 7). A cannulated lag screw over a K-wire was used to successfully reduce and fix this fracture under CT guidance.

58

5  Basic Principles of Reconstruction Techniques

Fig. 5.3  Lag screw principle – fully threaded screw with the need of proximal overdrilling

Fig. 5.4  Lag screw principle – partially thread screw without necessity of proximal overdrilling

Other unique posterior screw constructs include atlantoaxial transarticular screws as initially described by Magerl, in 1979 [28]. The pronounced rotator and translational instability present with Type II odontoid fractures or TAL injury results in decreased fusion rates when the initial posterior wiring techniques alone

were used. Therefore, any form of rigid fixation at this level was going to enhance fusion success rates (Chap. 6). The transarticular screw resulted in immediate fixation that allowed for posterior fusion of any kind (Gallie, Sonntag, Brooks) to take place. This took away the new, abnormal axis and amount of rotation

59

5.2  Construct Design Fig. 5.5  For adequate stability, atlantoaxial screws need to be placed sufficiently deep within C1 lateral mass (tricortical) or even through the far cortex of C1 (quadricortical). Increased risk of neurovascular injury with long screws exists

and translation seen after isolated posterior wire/cable constructs [13], offered significant shear shielding and decreased pseudoarthrosis rate [18]. The atlantoaxial screws can offer a neutralizing and compressive effect on the AA complex. Although, theoretically, they do not need to offer any kind of compression across the joint as the point of fusion is distant from the joint (i.e., posterior if C1 arch remains intact), the strength of the transarticular screw can be enhanced by quadricortical rather than tricortical screw purchase (Fig.  5.5). Quadricortical screw purchase obviously carries a risk of neurovascular injury anterior to the C1 as described in Chap. 6. Quadricortical screw purchase can be intentional for screw pull out strength and toggle avoidance as ­mentioned above, or can serve as a rescue option. C2 pedicle screws are a good example. The technique (Chap. 6) and anatomical reasoning (Chap. 1) are ­discussed elsewhere; however, a small C2 pedicle can make placement of a 3.5  mm screw impossible. Exceptionally, as an alternative, a more medial trajectory through the lateral spinal canal can provide a robust screw anchor through four cortices (Fig. 5.6).

Fig. 5.6  C2 quadricortical pedicle screws

5.2.2 Anterior Structural Constructs The anterior column reconstruction techniques at the CVJ, just as elsewhere in the spine, must result in restoration of a stable load-bearing column, maintenance of appropriate height, and sagittal alignment for a long enough period, so that bony integration and fusion can take place. The eventually biologically integrated construct will be replaced by living bone and become obsolete. The available constructs include tricortical autografts or allografts and synthetic cages. The lack of vertebral body at the C1 level and the unique biomechanical profile of the UCS make structural anterior constructs much less common. Unlike in the subaxial spine, axial loading forces are transmitted from the head to C2 via the occipital condyles, C1 lateral masses, atlantoaxial joints in the middle column, and then C2/3 disk space. Therefore, from a biomechanical viewpoint, an isolated anterior column construct at the CVJ (except C2/3 interbody cages/

60

5  Basic Principles of Reconstruction Techniques

grafts) does not make sense and most likely would fail without posterior support. Nonetheless, they can be used for stability restoration in defects created by treatment of neoplastic, inflammatory, or infectious lesions in combination with posterior techniques. The presence of a robust anterior weight-bearing column with a posterior tension band (occipitocervical fusion) represents a tension band principle that allows dynamic compression of the anterior column and thereby encourages fusion [4]. The difficulty of anterior column reconstruction at the CVJ is the relative lack of sufficient anchors at the superior end of the construct (i.e., clivus or C1 attachment) and the forces applied by the head. Some authors have attempted to replace anterior elements of C2 ­vertebra with a specific C2 prosthesis [24] that utilizes a load-bearing interbody device with buttress-platelike attachment principles. The authors gradually ­developed an implant that respects the loading force distribution of the head from a two-column system of the C0-C1-C2 segment to the three-column one ­present in the ­subaxial spine. When C2 corpectomy or vertebrectomy is undertaken, a clear reconstructive plan must be present. In our opinion, two main strategies

exist: (1) reconstruction is going to include both anterior and posterior instrumentation from the occiput/ clivus to subaxial spine as shown in Fig.  5.1 where anterior cage was anchored into clivus and middle column support was created through inter-facet cages between C1 and C3 or (2) decreasing the forces transmitted to the ­construct from the large lever arm of the head by a structural attachment to C1 both anteriorly and posteriorly (Fig. 5.7) and thus excluding the C0-C1 segment from the construct and allowing for a shorter period of postoperative immobilization [33]. Without anterior column reconstruction at the C2 level, a bridge fixation principle needs to be applied [4] with posterior constructs. An increased stress transfer and thus minimized fatigue failure can be achieved by creation of a posterior construct with multiple points of fixation. At the CVJ, that means extension of the construct to the lower cervical spine [16] over segments not involved in the pathological process. And, without the prospect of any anterior support over time, it also means a likely fatigue failure of the construct, which would need to be able to endure three million loading cycles to survive one year after insertion [4]. Long term, a posterior fusion

Fig.  5.7  Combined atlanto-cervical construct utilizing an ­interbody cage anteriorly and bicortically anchored screws posteriorly used to reconstruct a C2 spondylectomy defect. Compare to Fig. 5.1. This construct is off-loaded by exclusion

of the ­normal C0-C1 segment. (a) Sagittal CT reconstruction. (b) Axial images (courtesy of R. Bohinski, MD, Mayfield Clinic, Cincinnati, OH)

61

5.3  Fracture Healing/Bone Fusion

with instrumentation cannot compensate for a complete defect in the ­anterior column. Irrespective of the construct created, the basic ­principle of sufficiently rigid immobilization of the involved segment must be achieved in order for a bone fusion to take place. Ideally, the construct design allows for both anterior and posterior (or lateral) with sufficient mediation of bone growth, as solid, stable bony fusion in an anatomically aligned and balanced CVJ is the ultimate goal of any construct created.

5.3 Fracture Healing/Bone Fusion Although, historically, spinal instrumentation [19] preceded attempts at fusion [3, 23], it is clear that the two need to go hand in hand. Frequently, a spinal ­reconstruction cannot be considered successful without the presence of bone fusion. Fractures heal by means of an indirect bone healing [32] that involves sequential steps of tissue differentiation, resorption of surfaces of the fracture, uniting of the fracture fragments by callus, and internal remodeling [31]. Therefore, with any fracture, there is an attempt to reduce it, align it, and maintain it in reduced position until bone healing is complete (i.e., biological fixation has taken place). On the other hand, internal fixation produces a stable, rigid construct until bone healing occurs (direct bone healing). Direct healing occurs under compressive conditions and skips the intermediate steps of the indirect process and proceeds directly (not necessarily faster) to internal remodeling of the Haversian system. Therefore, direct healing does not result in callus formation. If this is visible after an internal fixation, it is understood that the stability did not reach the intended levels [32]. However, a completely rigid fixation of a fracture gap results in lack of mechanical induction of callus formation (strain theory) [11]. These basic orthopedic concepts, derived from fractures of long bones, hold value when assessing healing of fractures at the CVJ that were treated by means of an internal fixation with direct osteosynthesis (e.g., odontoid screw, C2 compressive pedicle screws). Vascular supply to the fracture site also plays a role in healing [34]. Unlike fractures, fusion techniques at the CVJ (Brooks, Gallie, Sonntag, anterior cage construct)

require deposition of a new bone in intersegmental locations that are not biologically structured for bone formation. Although instrumentation significantly improves fusion rates as demonstrated by addition of atlantoaxial screws to posterior fusion techniques in atlantoaxial instability treatment [18], failures of ­long-term stability occur. Fusion rates are dependent on multiple local, host’s, technical, and environmental factors. The internal fixation factors affecting healing are discussed above. Graft properties (osteo-induction, -conduction and –genecity) and type (autograft, allograft, xenograft), mechanical stability, and graft site preparation are among some of the local factors determining fusion occurrence [5]. Nicotine, osteoporosis, hormonal imbalance, and certain pharmacotherapy are well-known interferers of bone fusion. An organized effort has been made to supplement bone graft materials with fusion enhancers, such as growth factors or electrical stimulation [5]. Discussion of those factors is, however, beyond the scope of this chapter.

5.3.1 Our Preference Constructs at the CVJ are complex and basic biomechanical as well as biological principles need to be applied to each individual scenario. During anterior decompressive procedures at the UCS, we favor the use of perimesh cage reconstruction fashioned in such a manner as to be utilized as an anterior plate also. This then fulfills the role of a buttress plate, interbody loadbearing construct with packed autograft to facilitate anterior column stability, minimize subsidence, and allow anterior fusion. We believe bicortical screw purchase is important in design of a stable construct and we aim for all hardware to be well anchored. We do not use a cannulated odontoid lag screw construct over a K-wire with the understanding that it could be inadvertently advanced with catastrophic results (Fig. 5.8). Economic restraints only allow selective use of bone morphogenic enhancers of fusion in our practice but all patients are screened for any potential factors that would influence adequate healing.

62

5  Basic Principles of Reconstruction Techniques

Fig. 5.8  Lag screw introduction along the K-wire. Note that if odontoid apex is drilled through the wire become free (not fixed) presenting a danger of its cranial dislocation during screw purchase

References   1. Aebi, M., Etter, C., Coscia, M.: Fractures of the odontoid process. Treatment with anterior screw fixation. Spine (Phila Pa 1976) 14, 1065–1070 (1989)   2. Aebi, M., Thalgott, J.S., Webb, J.K.: Priniciples of surgical stabilization. AO ASIF principles in spine surgery, pp. 5–12. Springer, New York (1998)   3. Albee, F.H.: Transplantation of portion of the tibia into the spine for Pott’s disease. JAMA 57, 885 (1911)   4. Arlet, V., Datta, J.C.: Upper cervical spine. In: Aebi, M., Arlet, V., Webb, J.K. (eds.) AO spine manual: principles and techniques, vol. I, pp. 265–288. Thieme Verlag, New York, Stuttgart (2007)   5. Babat, L.B., Boden, S.D.: Biology of spine fusion. In: Benzel, E.C. (ed.) Spine surgery, techniques, complication avoidance, and management, pp. 169–177. Elsevier, Philadelphia (2005)   6. Benzel, E.C.: Upper cervical and occipitocervical arthrodesis. In: Benzel, E.C. (ed.) Spine surgery, techniques, complication avoidance, and management, pp. 329–340. Elsevier, Philadelphia (2005)

  7. Bohler, J.: Anterior stabilization for acute fractures and nonunions of the dens. J Bone Joint Surg Am 64, 18–27 (1982)   8. Caspar, W., Barbier, D.D., Klara, P.M.: Anterior cervical fusion and Caspar plate stabilization for cervical trauma. Neurosurgery 25, 491–502 (1989)   9. Chang, K.W., Liu, Y.W., Cheng, P.G., et  al.: One Herbert double-threaded compression screw fixation of displaced type II odontoid fractures. J Spinal Disord 7, 62–69 (1994) 10. Choueka, J., Spivak, J.M., Kummer, F.J., et al.: Flexion failure of posterior cervical lateral mass screws. Influence of insertion technique and position. Spine (Phila Pa 1976) 21, 462–468 (1996) 11. Claes, L.E., Heigele, C.A.: Magnitudes of local stress and strain along bony surfaces predict the course and type of fracture healing. J Biomech 32, 255–266 (1999) 12. Coe, J.D., Warden, K.E., Sutterlin 3rd, C.E., et al.: Biomechanical evaluation of cervical spinal stabilization methods in a human cadaveric model. Spine (Phila Pa 1976) 14, 1122–1131 (1989) 13. Dickman, C.A., Crawford, N.R., Paramore, C.G.: Bio­ mechanical characteristics of C1-2 cable fixations. J Neurosurg 85, 316–322 (1996)

References 14. Dvorak, J., Panjabi, M.M., Novotny, J.E., et al.: In vivo flexion/extension of the cervical spine. J Orthop Res 9, 824–834 (1991) 15. Egol, K.A., Kubiak, E.N., Fulkerson, E., et al.: Biomechanics of locked plates and screws. J Orthop Trauma 18, 488–493 (2004) 16. Fourney, D.R., York, J.E., Cohen, Z.R., et al.: Management of atlantoaxial metastases with posterior occipitocervical stabilization. J Neurosurg 98, 165–170 (2003) 17. Griffith, S.L., Zogbi, S.W., Guyer, R.D., et al.: Biomechanical comparison of anterior instrumentation for the cervical spine. J Spinal Disord 8, 429–438 (1995) 18. Grob, D., Jeanneret, B., Aebi, M., et al.: Atlanto-axial fusion with transarticular screw fixation. J Bone Joint Surg Br 73, 972–976 (1991) 19. Hadra, B.E.: Wiring of the vertebrae as a means of immobilization in fractures and Pott’s disease. Med Times Reg 22, 423 (1891) 20. Heller, J.G., Estes, B.T., Zaouali, M., et al.: Biomechanical study of screws in the lateral masses: variables affecting pull-out resistance. J Bone Joint Surg Am 78, 1315–1321 (1996) 21. Heller, J.G., Silcox 3rd, D.H., Sutterlin 3rd, C.E.: Compli­ cations of posterior cervical plating. Spine (Phila Pa 1976) 20, 2442–2448 (1995) 22. Heywood, A.W., Learmonth, I.D., Thomas, M.: Internal fixation for occipito-cervical fusion. J Bone Joint Surg Br 70, 708–711 (1988) 23. Hibbs, R.A.: An operation for progressive spinal deformities. NY State J Med 93, 1013–1016 (1911) 24. Jeszenszky, D., Fekete, T.F., Melcher, R., et al.: C2 prosthesis: anterior upper cervical fixation device to reconstruct the second cervical vertebra. Eur Spine J 16, 1695–1700 (2007) 25. Knoringer, P.: Osteosynthesis of injuries and rheumatic or congenital instabilities of the upper cervical spine using double-threaded screws. Neurosurg Rev 15, 275–283 (1992) 26. Leconte, P.: Fracture et luxation des deux premieres vertebres cervicales. In: Judet, R. (ed.) Luxation Congenitale de la Hanche. Fractures du Cou-de-pied Rachis Cervical. Actualites de Chirurgie Orthopedique de l’Hospital Raymond-Poincare, vol. 3, pp. 147–166. Masson et Cie, Paris (1964)

63 27. Lehmann, W., Briem, D., Blauth, M., et al.: Biomechanical comparison of anterior cervical spine locked and unlocked plate-fixation systems. Eur Spine J 14, 243–249 (2005) 28. Magerl, F., Seemann, P.S.: Stable posterior fusion of the atlas and axis by transarticular screw fixation. In: Kehr, P., Weidner, A. (eds.) Cervical spine, pp. 322–327. Springer, Wien (1987) 29. Montesano, P.X., Anderson, P.A., Schlehr, F., et al.: Odontoid fractures treated by anterior odontoid screw fixation. Spine (Phila Pa 1976) 16, S33–S37 (1991) 30. Penning, L., Wilmink, J.T.: Rotation of the cervical spine. A CT study in normal subjects. Spine (Phila Pa 1976) 12, 732– 738 (1987) 31. Perren, S.M.: Physical and biological aspects of fracture healing with special reference to internal fixation. Clin Orthop Relat Res 138, 175–196 (1979) 32. Perren, S.M.: Evolution of the internal fixation of long bone fractures. The scientific basis of biological internal fixation: choosing a new balance between stability and biology. J Bone Joint Surg Br 84, 1093–1110 (2002) 33. Piper, J.G., Menezes, A.H.: Management strategies for tumors of the axis vertebra. J Neurosurg 84, 543–551 (1996) 34. Schiff, D.C., Parke, W.W.: The arterial supply of the odontoid process. J Bone Joint Surg Am 55, 1450–1456 (1973) 35. Stulik, J., Suchomel, P., Lukas, R., et al.: Primary osteosynthesis of the odontoid process: a multicenter study. Acta Chir Orthop Traumatol Cech 69, 141–148 (2002) 36. Suchomel, P., Buchvald, P., Barsa, P., et  al.: Single-stage total C-2 intralesional spondylectomy for chordoma with three-column reconstruction. Technical note. J Neurosurg Spine 6, 611–618 (2007) 37. Tuite, G.F., Papadopoulos, S.M., Sonntag, V.K.: Caspar plate fixation for the treatment of complex hangman’s fractures. Neurosurgery 30, 761–764 (1992). discussion 764-765 38. Wellman, B.J., Follett, K.A., Traynelis, V.C.: Complications of posterior articular mass plate fixation of the subaxial cervical spine in 43 consecutive patients. Spine (Phila Pa 1976) 23, 193–200 (1998) 39. Werne, S.: Studies in spontaneous atlas dislocation. Acta Orthop Scand Suppl 23, 1–150 (1957) 40. White, A.A., Panjabi, M.M.: Clinical biomechanics of the spine. Lippincot, Philadelphia (1990)

6

Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction P. Suchomel and O. Choutka

A thorough knowledge of anatomy is necessary when instrumenting the upper cervical spine. Because of the frequency of anomalies of the bone and neurovascular structures in this region, multiple diagnostic studies are usually required for a comprehensive evaluation and surgical planning. The use of simple axial CT imaging is insufficient. In the majority of cases, using CT reconstructions and MRI is essential. At times, other imaging modalities such as CT angiography or CT myelogram may be required. Generally, for a safe 3.5 mm screw purchase, the diameter of available bone surrounding the screw should exceed 5  mm if visual and/or fluoroscopic control is used [153, 232]. When using image-guided navigation, the diameter of available bone should be at least 4 mm [19]. Some authors primarily prefer the use of 4 mm screws [9, 234], and thus the available bone amount in the plane perpendicular to the axis of screw trajectory should be adapted by adding a minimum of 1  mm to the previously ­mentioned dimensions. If, in exceptional cases, direct real-time visualization is used (prioperative CT or isofluoroscopy), the outer diameter of the bone can be the same diameter as the screw. Some experienced surgeons will accept the core diameter of screw being

P. Suchomel Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic O. Choutka Department of Neurosurgery, University of Cincinnati College of Medicine, 231 Albert Sabin Way, Cincinnati, OH 45267-0515, USA

slightly smaller than the diameter of pedicle, arch, or isthmus in the belief that the slight cortex “blow-out” is not dangerous. In our opinion, this philosophy can be accepted only if there are no other options and if the target structure has an appropriate “guiding tunnel” of cancellous bone surrounded by cortex. The other extreme possibility is to intentionally go out of the bone (e.g., out of the pedicle) when the standard purchase can endanger vitally ­important structures. In such a situation, one can select tri- or quadri-cortical purchase involving the spinal canal and/or extravertebral space. Despite numerous techniques of fixation described in the literature using different types of very sophisticated constructs manufactured from state-of-the-art materials, these technical developments are only supportive tools facilitating the correct environment for bony fusion and healing. The preparation of fusion surface and the use of osteoinductive and osteoconductive biomaterials are of paramount importance. Only a stringent, independent evaluation of fusion result can confirm the validity of one’s own work. Certainly, in some situations (e.g., the elderly), radiographic and functional stability without evident bony fusion can be enough. Strictly, the term “fusion” should not apply to a situation when there is lack of movement on dynamic radiographs. There should also be evidence of bony mass bridging the fused segment (best documented on CT) without any radiolucency surrounding the hardware. First, we shall describe occiput, atlas, and axis one by one as anchoring structures and then the potential fusion constructs of craniovertebral junction(CVJ) and upper cervical spine (UCS).

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_6, © Springer-Verlag Berlin Heidelberg 2011

65

66

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

6.1 Occipital Bone as Anchoring Structure 6.1.1 Occipital Squama In any type of CVJ instability, the head has to be included in the stabilization construct. The occipital bone therefore is always involved and most frequently the occipital squama is used as a cranial anchor. Historically, pure onlay bone grafting was used [169] and then various wiring techniques were utilized to fix either bone ­strut grafts [78, 235] or polymethacrylate inlays [160] between occiput and UCS and thus stabilize the CVJ. Despite mandatory use of external supports (halo or Minerva), previous techniques often fail in the long term. The application of contoured rods, loops, and frames which are fixed to intact posterior spinal elements and doubled holes in the occiput started the semirigid era of craniovertebral fixation [183]. To increase the solidity and reliability of CVJ fixation, screws ­connected to plates, both locked and unlocked, were used in the early 1990s [84, 143, 199, 207]. These constructs demonstrated much higher rigidity and substantially increased the fusion rate. This decreased the necessity of rigid external fixation; however, the fixed design of the plates often dictated the position of screw to suboptimal locations. Additionally, a straight line concordant with UCS fixating points and the plated was essential. This was disadvantageous, especially when the lateral placement of screws in the thin part of occipital bone often leads to loosening or breakage. Currently, a

modular screw – rod fixating systems are available [1, 115, 178]. They provide flexibility to place occipital screws in the area of thickest bone independently of the positions of the spine screws. The majority of occipital plates have movable and multiaxial U-shaped fixating heads that enable variable positioning of the contoured rods to the cervical polyaxial screws. Such variable rigid constructs allow shorter segments of fixation preserving more motion segments. When instrumenting the occipital squama, a thorough knowledge of anatomy and any patient variation is necessary. Evaluating the bone thickness in the planned screw locations and the intracranial position of venous sinuses is paramount.

6.1.1.1 Anatomical Background The occipital squama bone thickness is the greatest at the external occipital protuberance (EOP) and decreases in a radial distribution (Fig. 1.1, Chap. 1) [55, 250] The superior nuchal line does not reflect the internal position of transverse sinus accurately; the relation of the confluence of sinuses to EOP is more consistent [189]; therefore, our screw position should be approximately 1 cm below it and not more than 2  cm, laterally. As was described in the prior chapter on anatomy, we can expect the bone thickness in the EOP to be 15 mm in males and 12 mm in females on an average (Fig. 6.1a). The other “safe area” is relatively thin strips of bone that extends caudally from the EOP and is a reflection of the internal occipital crest. The thinnest bone is directly above the cerebellar hemispheres inferior to INL (Fig. 6.1b). b

Fig. 6.1  Normal thickness of occipital bone visible on CT sagittal reconstructions documented in a male patient. (a) Midsagittal scan with bone thickness between 10–15 mm. (b) Less than 4 mm thin bone over cerebellar hemisphere in the same patient

67

6.1  Occipital Bone as Anchoring Structure

6.1.1.2 Surgical Technique Usually, the posterior midline skin incision starts 1 cm above the inion and continues splitting the nuchal ligament inferior to the desired level of cervical spine. Subperiostal dissection with or without sparing the EOP muscular attachments exposes the external anatomical landmarks, the SNL, INL, and the edge of foramen magnum. Palpating the C2 spinous process and critically, the C1 posterior tubercle helps to identify the posterior FM rim. Depending on technique chosen, the extent of occipital bone exposure is defined. In the case of in line lateral plates, which are in continuity with lateral masses of spine, more lateral dissection is required, whereas midline fixation only needs limited exposure. Reviewing the anatomical landmarks, preoperative radiographs, and CT, the holes corresponding to plate are drilled. On an average, 12  mm screw purchase in midline is safe. If lateral screw location is chosen then the necessary bone thickness for screw purchase should be at least 6 mm but preferably, 8 mm. Bicortical screw purchase is probably not necessary in occipital area. This is supported by Zipnick’s paper, which demonstrates that the outer cortex contributes 45% of total occipital bone thickness whereas the inner one is providing only 10% [250]. However, Haher et al. [90] found that bicortical pullout strength was 50% greater than unicortical. Because the strength of screw fixation is proportional to thickness of the bone, the EOP (inion) and caudal midline are ideal for screw placement [177, 189]. Whichever type of screw introduction is selected, the holes should be tapped as the bone can be very hard and the screw may break during tightening. Four to six millimeters diameter occipital screws are used to fix the plates. Although the length of screws is often established preoperatively, lateral fluoroscopy is recommended to ­double-check the desired length, to achieve correct perpendicular screw angle and to verify the full contact between the plate and the bone. Final tightening is done in controlled fashion with the torque wrench.

Comparing these landmarks with preoperative CT and plain X-ray, we select the appropriate position of the occipital plate on the occipital squama inferiorly to SNL. The final position of the occipital plate is also influenced by the availability of skin coverage above it. Sometimes, we have to localize the plate more caudally to avoid the potential of erosion over the plate. Holding the plate in the planned location, we mark the planned drill entry points with high speed burr or awl. Then with a chisel or high speed reamer we prepare the surface of the bone to accept the plate (eventually contoured) without any air gaps in the interface. The plate is held in the final position and the final holes are drilled with the safety stop drill guide. We always start with the deepest screw to firmly attach the plate for further drilling. If bicortical screw placement is attempted, then one has to be aware of dural and/or sinus injury. Potential sinus injury has to be treated properly. The head must not be above the level of right cardiac atrium to avoid air embolism. It is much better to err on the side of venous bleeding than a dry field that is sucking air. In such cases, the “sucking” hole must be plugged with cottonoid and flooded with water. The surgical position of the patient has to be changed immediately (Trendelenburg) and a shorter screw, away from the sinus is placed. Violation of the sinuses can cause not only bleeding but also much more dangerous venous sinus thrombosis. Simple dural penetration is frequently seen with CSF leak plugged with the screw introduction. However, such a “minor complication” can injure surface vessels of the cerebellum causing subdural or epidural hematoma. In conclusion, we can say that the best and the safest anchorage can be achieved in the midline below the EOTP and that bone thickness less than 6 mm is not sufficient for firm screw purchase. We believe that the safest approach to this procedure is ensured by precise preoperative evaluation of occipital bone by thin cut CT (Fig. 6.2).

6.1.1.3 Our Preference

6.1.2 Occipital Condyles

Radiological evaluation prior to any occipital bone screw introduction is essential to achieve firm, bicortical screw placement while avoiding the venous sinuses. We prefer to dissect the occipital bone subperiostally without any midline muscular attachment left in place and then to define the external anatomical landmarks.

In the case of thin or missing (after craniectomy) occipital bone, the occipital condyles can be used as an anchoring structure. Also, a monosegmental transarticular atlantooccipital (atlantocondylar) screw fixation can be used to fix AOD, especially if it is mild and reduced. Other indications include the

68

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

a

b

c

Fig. 6.2  Extreme midline thickness of occipital bone in another female patient. (a) Midsagittal CT reconstruction depicting more than 20  mm of available bone. (b) Perfect monocortical

screw anchorage in the same patient (“tooth-brush appearance”). (c) Axial CT reconstruction in the same patient

desire to fortify a multipoint construct involving the condyle screws and/or failed previous occipitocervical fusion. Currently, the posterior approach using either transcondylar or C1-0 transarticular screws is most frequently discussed [81, 228, 229, 245]. Nevertheless, anterior transcondylar screw purchase has also been described [52, 53]. As was depicted in the anatomy chapter, the occipital condyle is normally twice as long as it is wide and is a medially oriented structure. For practical purposes we can calculate the length as approximately 25 mm, the width around 10 mm, and the height as 10 mm. Because of its variability only CT scan can accurately show its shape, orientation, mass, and relationship to neighboring structures [165]. Since it is a part of occipital bone, cranial image guidance can be used not only to model the ideal screw trajectory, but also to directly guide the instruments. Most of the published

anatomical works show the safety of using 3.5  mm screws for transcondylar purchase [133, 229, 245].

6.1.2.1 Posterior Transarticular Atlantocondylar Screw (Fig. 6.3) Dieter Grob from Zurich was the first who published the posterior transarticular C1-0 screw fixation combined with a Y-shaped C2-occipital plate in a patient with a failed previous wire fixation of AOD [81]. Gonzalez et  al. [75] studied the feasibility of C1-0 transarticular screws on cadaveric model. They found that for atlanto-occipital screw fixation the same stiffness as occipitocervical constructs in all directions, with the exception of flexion-extension. Their recommendation was to supplement this technique with a posterior buttress. Similarly to Magerl’s technique it

6.1  Occipital Bone as Anchoring Structure

Fig. 6.3  Schematic drawing of posterior atlantocondylar screw fixation

can be strengthened by a Gallie type of graft [75]. The same group of authors later published a report of a patient with AOD treated using this method [60] and also in two cases of posttraumatic instability where they performed a combined double level transarticular procedure (C1-0 and C2-1), simultaneously [76]. Yan et al. in their works first defined the ideal entry point for atlantocondylar fixation and then analyzed 20 dry specimens and CT reconstructions of 30 healthy volunteers. They also conducted a simulation surgery on another 12 fresh cadavers to establish the safety angle ranges and length of screws [245, 246]. As a safe angle of introduction, they established 53.3° (SD = 3.4°) in sagittal plane and the medial inclination 20° (SD=2.6°) in the axial plane. The appropriate length of transarticular C1-0 screw was between 24 and 34  mm. However, always, one has to consider the individual patient specific anatomy as well as the possible discrepancy in such values obtained in, generally, smaller South-Eastern Chinese population.

69

anatomical works by Uribe et al. [229] and LaMarca et al. [133]. The position of vertebral artery (VA) in the C1 arch posterior groove does not usually affect the condyle screw purchase as it is most frequently located more caudally [133]. From an anatomical point of view, the natural borders for screw placement are as follows: rostrally, the hypoglossal canal; rostrolaterally, the emissary vein; caudally, the occipitoatlantal joint; and medially, the foramen magnum. The condylar emissary vein can be of importance as a major drainage vein in cases of jugular bulb occlusion (tumors) or in congenital anomalies but it can be sacrificed under normal conditions [13, 28]. LaMarca et  al. [133] described the ideal transcondylar screw trajectory after analyzing thin sliced CT in a 3D navigational station. They found that it was feasible to achieve safe screw purchase in all the 12 cadaveric condyles studied with a safety rim of bone surrounding the screw larger than 1.5 mm in all cases. The average SAS on the posterior condylar wall was 5–8 mm rostrocaudally and 5–9 mm mediolaterally. Uribe et  al. [229] studying the feasibility of transcondylar screw purchase on six silicone injected cadaver heads determined the condylar entry point (CEP) to be 4–5 mm laterally from posteromedial edge of the condyle, anatomically defined as approximately 1–2  mm above the joint fissure. The base of condyle (connection to occipital bone) was used. The pilot hole was made by an awl and the trajectory of drill was 5°

6.1.2.2 Posterior Transcondylar Screw (Fig. 6.4) The idea to use the transcondylar screw purchase as a new point of fixation was described independently in

Fig.  6.4  Schematic drawing of posterior transcondylar screw purchase

70

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

upward and tilted 17° (12°–22°), medially. They used 30–32  mm long screws where the unthreaded shaft 11–13  mm long was protruding above the bone to enable the polyaxial screw head movement. They attempted bicortical screw purchase and combined anatomical guidance with fluoroscopy. All 12 operated condyles were assessed by CT afterwards. Not one screw violated the hypoglossal foramen or other important structures. The screw length inside the bone was, on an average, 22  mm (20–24  mm) and the C1 arch overhang (smooth shaft) 12 mm (11–13 mm). This suggested technique of transcondylar screw placement as a part of longer caudal construct was later successfully used in patients with odontoid type II fracture pseudoarthrosis and cranial settling [228]. Despite the fact that bicortical screw placement is stronger than unicortical, we have to be aware of the potential injury of structures located anteriorly to the condyles. Most commonly, it is the pharyngeal wall but variant carotid arteries can be anterior as well. There are some advantages of posterior condyle screw purchase. Using a polyaxial screw, the contoured rod connecting the occipital plate is not necessary and thus the risk of eventual stress rod fracture can be avoided. Also, the construct connecting the occipital bone with the UCS or subaxial spine is of low profile, and therefore, the muscular damage necessary for occipital plate placement can be decreased as well as the risk of plate erosion.

6.1.2.3 Anterior Transarticular Axial-Atlantocondylar Screw (Fig. 6.5) The original idea to fix the occipital condyle from an anterior approach similar to the technique of anterior atlantoaxial fixation came from Dvorak et al. [52, 53]. They suggested this method as a salvage procedure for those unique situations where posterior fixation is not possible, failed, or has to be fortified. Also, in some very rare situations such as after total tumor removal or in complex reconstruction due to congenital anomalies, these ideas can be utilized. The first part of their work was an anatomical study documented by a successfully treated patient with failed posterior wire and graft fusion after repetitive trauma. The second part was a biomechanical comparison of anterior fixation with posterior methods. They confirmed the superior strength of posterior transarticular

Fig. 6.5  Artistic drawing of anterior C2-C1- occipital condyle screw fixation

screw connected to suboccipital plate for all directions; however, they found comparable stability of their anterior fixation in rotation and lateral bending. Both screw methods were much more stable than posterior graft and wiring alone. They have suggested approaching the anterior surface of C2 the same way as for an odontoid screw (high oblique anterolateral approach), to identify the groove below the middle third of atlantoaxial joint and introduce the K-wires tilted 25° posteriorly and 15° laterally under biplanar fluoroscopical control. Then cannulated self-tapping screws 24–30  mm long were introduced along the wire. From the anatomical and CT studies, they concluded that the angle of introduction can vary substantially (posterior angle 15°–36°, lateral tilt 10°–20°). As a major limitation, they cite the course and volume of hypoglossal canal and the impossibility to add graft material or to abrade the joint surface to enable long-term bony fusion.

6.1.2.4 Our Preference The occipital condyle has been confirmed as a solid structure for screw anchorage. However, it has also been found that the variability of VA course and location of n.XII canal within the condyle can exclude the possibility of safe screw placement in as many as 17% of the specimens studied [52, 53]. Also, in basilar impression

71

6.2  Atlas as an Anchoring Structure

and other compound CVJ congenital anomalies, it could be unsafe or even impossible to expose the occipital condyle. Although case reports describing the successful use of either posterior transcondylar [228] and posterior atlanto-occipital transarticular screw fixation [76, 81] or anterior transarticular C1-0 screw placement [52] were reported, no large series of patients treated with condylar screw placement has been published. In our opinion, the use of the condyle as a part of construct can be more reasonable than transfixation of occipitoatlantal joint without support of bone grafting. Given the technical and anatomical difficulties, although these methods are feasible, they will likely have a limited role mainly as salvage procedures.

the length of the whole clivus is 4.5 cm (3.7–5.2) in the sagittal plane and the basilar portion of occipital bone is 3.1 cm (SD = 0.3). In occipital hypoplasia, the basilar portion may be only 1.7  cm long [134]. The thickest portion is anterior and superior and contains the cancellous bone. The thinner part is formed only by compact bone in the region of foramen magnum. Usually, the outer cortex is more solid and thicker than the inner one. For practical purposes, we can calculate that the wedge-shaped clivus is only 4  mm thick and safely reachable 10  mm on both sides from midline at the level of FM. It gradually increasing in thickness to a maximum of 22 mm at the level of pituitary fossa with a safe strip of bone 10  mm from midline along its entire course. The caudal half of the clivus can be safely resected if necessary or can serve as a screw anchorage or cage support (Fig. 6.6).

6.1.3 Clivus The upper part of the clivus belongs to sphenoid bone whereas the lower part is a basilar portion of the occipital bone. These two parts are separated by spheno-occipital synchondrosis till the age 16.5 (13–18) in males and 14.4 (12–15) in females. This represents the growing potential in correct formation of skull. In normal adults, a

6.2 Atlas as an Anchoring Structure C1 has no vertebral body, and therefore, its bone stock has a limited volume for any screw anchorage; nevertheless, the lateral masses are frequently used for screw b

Fig. 6.6  Two screws 12 mm long introduced in to the anterior lover clivus under image guidance. (a) Midsagittal CT reconstruction showing appropriate screw length. (b) Coronal plane CT reconstruction

72

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

purchase, either from anterior or posterior approach. Additionally, C1 lateral mass and laminar screws are described in the literature. To reach desired lateral mass posterior and/or anterior screw entry points, the arch is the anatomical guiding structure. When performing posterior approach, the dissection of C1 arch should stay strictly subperiostally starting from the clearly distinguishable posterior tubercle. Ebraheim recommended [57] not to extend this dissection more than 12 mm lateral from midline and be aware 8 mm from midline on superior arch. However, it is our experience that following the lower edge of arch where we do not expect any important neurovascular structures, safe microdissection is possible even to the transverse process. The inferior half of the C1 lateral masses can be exposed subperiostally under the thinnest part of posterior arch with care taken not to injure the C2 root, ganglion, and surrounding venous plexuses. Preoperative evaluation of individual VA course is critical. Careful observation of axial CT scans can show different C1 anomalies leading us eventually to further diagnostic modalities (CTA). The anterior approach to the atlas is limited to transoral exposure and/or high anterolateral access and again for any anterior approach the most important guiding point is the anterior tubercle of atlas. It is most often visible on lateral fluoroscopy, it is not surrounded by any danger and clearly localizes the midline, and thus distances to vertebral arteries and spinal cord. We have to be aware of possible slight rotation, nevertheless, the safety defined by anterior tubercle is always valid.

6.2.1 Posterior Lateral Massa Screw Fixation of the lateral mass of atlas with posteriorly placed screws was first described by Atul Goel from Mumbai, India in 1994 (first surgery 1988) as a part of C1-2 plate construct where the inferior screws were placed into the C2 pedicle [72]. This technique was later popularized by Jürgen Harms from Karlsbad, Germany who developed a more adaptable polyaxial screw rod system allowing preservation of the C2 nerve root and enabling manipulation with atlas particularly in fractures and dislocations [94]. Both authors used the middle posterior part of the C1 lateral mass below the arch as the entry point. Harms recommended reaching this area subperiostally and to make

a pilot hole in the middle of mass-arch base and thus avoid bleeding from the venous plexuses surrounding the C2 root. A straight or slightly convergent anteroposterior drilling trajectory parallel to the plane of the C1 posterior arch was advised. The screws were introduced bicortically. Goel described the entry point as the middle of the available bone area but introduced the screws monocortically, initially. In his later works, he conceded using bicortical purchase and an entry point localized in the lower facet joint surface if not enough bone is available in the posteroinferior C1 pillar, especially in children [70]. Other authors [148, 186] under influence of anatomical [225] and biomechanical works [148] suggested penetrating directly through the surface of the posterior arch just above the previously described entry point as another possibility of how to avoid bleeding, not irritate the C2 root, and prolong the screw pathway in order to increase the pullout strength. Such a strong posterior lateral mass screw anchor can also potentially limit previously necessary long occipitocervical constructs to occipitoatlantal or atlantocervical and thus avoid unnecessary fusion of adjacent segments. The originally suggested midline trajectory was criticized by Blagg et  al. [18] because of possible injury of the VA in the transverse foramen. They recommend to start more medially, drill straight anteriorly or in medial angle up to 20° and never tilt laterally during the procedure, but they concur with the original concept of following the posterior arch attachment angle in sagittal plane. Because of possible injury of structures situated in front of atlas (n.XII and ICA) another screw entry and trajectory was described by Rocha et al. [191].

6.2.1.1 Anatomical Background The space available for 3.5–4.5 mm lateral mass screw (SAS) introduction had been repeatedly studied on anatomical specimen and CT images [30, 191, 233, 241]. The posterior inferior pillar of the C1 lateral mass is bordered superiorly by the arch and inferiorly by the joint as measured by Blagg et al. [18] on 50 CT scans. They established the height of available space to be 4.6 mm (0–7 mm) and the width to be 14.9 mm (11–18 mm). Rocha et al. measured the same parameters on 20 dried atlantal specimens. They gauged electronically the mean working space height as 4.5  mm (range 4.3–6.1  mm)

73

6.2  Atlas as an Anchoring Structure

and the mean width as 9.6 mm (range 7.7–12.8 mm). To reach the aforementioned height they had to cut out the posterior arch lip in 50% of the studied specimens and they concluded that 93% of studied vertebras are able to accept screws with 4.5 mm diameter [191]. Wang analyzed 74 dried cadaveric spines and found a mean height of 3.9 mm and a width of 7.3 mm of the space for possible lateral mass screw entry [233]. All specimens could accommodate 3.5  mm screws and 97% could accept 4mm diameter screws. In 65% of cases, it was necessary to remove at least a part of the posterior arch overhang to be able to insert a 4mm screw. The possible introductory angle for safe bicortical purchase varied according to the position of entry point from 13° laterally to 45° medially. Cranial angulation without violation of C0/1 joint was acceptable until 19°. Because this technique can be very advantageous in fixation of the pediatric spine, Chamoun et al. [29] has performed a CT morphometric analysis of 76 atlases in children between 1.5–16 years old (mean 7.8 years) and found that only in one case of a 19-month-old infant one of the lateral masses was not of sufficient size to accept 3.5 mm screw. When performing the posterior approach to the atlas surgeon can find different anomalies of the VA course. Up to 15.6% of patients have a partial or total covering (arcuate foramen) of the VA artery in the groove of C1 by the “ponticulus posticus”. (Fig. 1.3, Chap. 1) [31, 104]. Lee et al. in their anatomical work analyzing 709 C1 cadaveric vertebrae found the appearance of ponticuli significantly more frequent in males (15.9%) vs. females (8.1%) [138]. Erroneous evaluation of a “too broad” a lamina can lead to VA injury [249] during arch subperiostal preparation or translaminar screw introduction. There are also anomalies of the course of the horizontal (V3) segment of the VA in 5.4% of the normal healthy population [104] and in 13% of the group of patients selected for CVJ surgery [244]. The most important and dangerous variant is the persistent first intersegmental artery. This aberrant vessel partially or totally substitutes the VA and courses below the posterior atlantal arch and thus prohibits the subarcuate approach to posterior lateral mass screw entry area. In a very large series of 1,013 patients with CT vertebral angiography, Hong found persistent first intersegmental artery on one side in 3.8% and bilaterally in 0.8% [104]. Structures adjacent to the anterior surface of lateral mass represent another risk during bicortical penetration

of drill, tap, or screw. The lumen of the internal carotid artery (ICA) is located medially to FT in more than 80% of cases. The mean distance of the ICA medial border measured from the medial border of FT is 2.78 mm on the left side and 3 mm on the right [37]. The average distance of ICA from anterior lateral mass aspect is less than 3 mm on both sides (left, 2.88 mm; right, 2.89 mm). Rotation of head due to positioning of the patient probably has no effect on any change of ICA position [38]. This close relationship can potentially be dangerous during placement of lateral mass screws, C1/2 transarticular, and/or C2 transpedicular fixation if bicortical purchase is chosen. Such a carotid artery impingement has been described after transarticular fusion [38]. The risk of ICA injury during drilling or tapping or lateral injury due to the screw contact and gradual ICA wall erosion was considered by Currier et al. as high, if the artery was more than 4 mm medial to the medial edge of FT and less than 2 mm from anterior bone surface. In moderate risk are those with the artery less than 2  mm from FT and within 2–4 mm from anterior C1 wall. They found that there was no risk when the ICA is laterally from transverse foramen and more than 6 mm away from the anterior mass aspect [37]. Currier, in his series of 50 atlases analyzed by CT with contrast, found that 12% of patients were at high risk and another 46% at moderate risk of ICA impingement at least on one side. Additionally, the position of the hypoglossal nerve in front of lateral C1 mass is of importance because of possible injury during bicortical screw purchase as described by Hong [103] and Jeanneret [116]. This could be a cause of some reports of swallowing difficulties after purely posterior procedures [86, 151]. The CN XII exits the skull via hypoglossal foramen at the base of occipital condyle with a diameter of 2–3 mm. It lies 2–3 mm laterally from the middle of the mass and courses vertically to the C1/2 joint [56].

6.2.1.2 Surgical Technique In a standard technique first described by Goel [72] and later by Harms [94] the middle subarcuate portion of the posterior lateral mass was used as a screw entry point. Goel, who performed his first procedure in 1988 in order to avoid venous bleeding from plexuses accompanying the second nerve root, and to explore widely the C1-2 joint, always cut out the C2 root with

74

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

its surroundings. This allowed to him to open the joint, distract, or manipulate the joint in C1/2 dislocations and also put bone graft or cages in between the roughened joint surfaces. Such a wide exposure enabled him also to adapt a short plate directly onto the bone. The price for this procedure was the denervation of C2 area which can be sometimes not well accepted by the patients [73, 88]. In order to avoid this “non physiological” dissection and to make the technique more versatile, Harms developed the polyaxial screw-rod modular system and introduced the screws at the midline base of posterior C1 arch. Later, the “transpedicular technique” was developed [186, 225] where the screw is introduced through the posterior arch of atlas straight forward in the lateral mass midplane. Unfortunately, this anatomical term “C1 pedicle” or “C1 pseudopedicle” or “C1 pedicle analog” incorrectly naming the posterior arch of atlas (as the atlas has no pedicles) has gained wider acceptance [30, 31, 148]. Initially, the methods of C1 lateral mass fixation [94, 152] were described with bicortical screw purchase. Currently, there is discussion if that is necessary. Cyr et  al. described no statistically significant difference in pullout strength between bicortically or monocortically introduced C1/2 transarticular screw [40]. Eck et al. found that significantly much larger force is necessary to pull out the lateral mass screws if bicortically introduced [58]. Advocates of so called “transpedicular” C1 screws argue from biomechanical work of Ma [148] demonstrating that posterior arch monocortical screws present larger pullout force than bicortically placed screws through the lateral mass in standard fashion. Nevertheless, in all the referenced studies, the pullout strength for standard monocortical screws was much larger than previously reported acceptable values for subaxial spine [119, 126] thus giving rational for surgeon’s preference to choose monocortical introduction whichever entry point is used. Certainly, bone mineral density, presence of background disease (inflammatory bone destruction -RA, pure bone quality in bone diseases or osteoporosis), and surgeon’s feeling of bone solidity play important roles in the decision if opposite cortex penetration can increase the stability of the construct. Some of the advocates of “transpedicular” method accept that 5 mm height of the arch above lateral mass and below VA is enough to accept 3.5 mm screw [30, 138]. Lee, in his series of 709 measured cadaveric atlases, found the average thickness 3.95 mm at the side

of VA groove. He stated that only 6.9% of female and 17.4% male specimens can safely receive the 3, 5 mm “transpedicular screw” and suggested as an alternative the notching technique where the inferior part of arch is drilled with 2  mm burr making a notch in which the screw shaft can be placed. Nevertheless, even this technique was not possible in 26.7% of females and 8.3% of males because the bony area of the atlas arch was less than 3 mm thick [138]. Christensen et al. analyzing 240 lateral masses of cadaveric C1 vertebrae accepted the smallest height of 4 mm and found that in19% of cases there was not enough space available for 3.5 mm screw placement [31]. In other studies [102, 148, 225], the larger portion of studied C1 posterior arches was able to accommodate screw placement and 4 mm arch thickness was considered enough to accept 3.5  mm diameter screws. This is debatable in our opinion because one can hardly imagine this without breaching the cortex in practical application. Other complications can include arch fracture and/or VA injury because of its atypical location [14] or during taping of thin arch bone [8].

6.2.1.3 Our Preference In our opinion, thin sliced CT, with 3D image reconstructions is essential for planning of C1 lateral mass placement. As in other pathologies of UCS we always procure MRI imaging as well. This can help to exclude vascular anomalies and localize the position of VA in relation to C1, as well as the ICA position anterior to the lateral masses (Fig.  6.7). The surgeon could also review if the ipsilateral FT is small or even absent on standard axial scans. If any suspicion of anomalous VA or ICA course arises from previous imaging, MRA, or CTA should be performed to elucidate its course (Fig. 6.8). With all images reviewed and analyzed the longest safe bicortical screw trajectory is planned. Based on anatomical studies, a quadratic area of 5 × 5 mm should be available in the majority of cases but in about 50% of patients the posterior arch lip overhang has to be removed to reach this working space (Fig. 6.9). Normally, the lip is resected with Kerrison rongeur or high speed drill. If the bone overhang is not removed, the subperiostal cleavage plane can be lost and the soft tissue can be violated. The amount of bleeding from injured venous plexuses surrounding the C2 root can be very serious. The elevation of the operative field can help but often we have to use haemostatic sealants and

75

6.2  Atlas as an Anchoring Structure

a

Fig. 6.7  Normal position of both carotid arteries shown on axial CT and MRI. (a) Axial CT scan after contrast media application. Notice the atypical loop of left V3 segment, VA hypoplasia on

a

b

the right side and the position of both carotid arteries. (b) Transversal MRI in case of AA subluxation with clearly seen position of both carotid arteries in front of atlas

b

Fig. 6.8  Course of arteries as shown on 3D CTA in two different patients. (a) Normal position of both carotid arteries, hypoplastic left VA. (b) Abnormal coiling of right and kinking of left ACI just in front of C1 lateral mass

temporary cottonoid pressure. If these methods do not help then we have to coagulate and/or ligate and transect the whole C2 neurovascular bundle. Once in the past we had a case of massive arterial bleeding during subperiostal approach to the subarcuate

lateral mass on the left side. Compressive packing together with a fibrin glue helped us to stop it and solve a very difficult situation. In spite of no pathological vascularization visible on postoperative angiograms, with our current knowledge we hypothesize that it could be

76

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

Fig. 6.9  Overhang of posterior arch impeding direct approach to the posterior inferior lateral mass screw entry area

persistent first intersegmental artery. This experience supports our previously mentioned complex preoperative investigation protocol. Theoretically, the previously described difficulties can be avoided if the screw is introduced directly through the arch. However, from daily practice, we know that the so called “transpedicular” drilling through the posterior arch of atlas can be very difficult. Even if an entry hole is prepared with a high speed burr, the drill can easily slip up or down and injure the

VA or C2 root bundle. It is also very complicated to hold the proper trajectory if the arch is less than 5 mm wide. In general, we feel that this technique and its modifications as a more risky option and we reserve the notching technique for specific anatomical situations. We prefer to locate the entry point in the middle of the posterior lateral mass at the top of the arcuate surface, which is palpable by a blunt probe in the base of posterior arch attachment or below it. To know the exact position of medial and lateral mass pillar borders, the thin Penfield probe is used to palpate it directly. Sometimes, the joint fissure can be visible but mostly this is unnecessary. The joint cleft can be frequently visible on lateral fluoroscopy. It is advantageous to make a small entry pilot hole with a high speed drill or awl to avoid dislocation of the drill in the beginning of pilot hole drilling. The C2 neurovascular bundle has to be covered and slightly caudally dislocated during all the work. A special set of instrument guides protecting the C2 bundle from direct contact of drill, tap, screw, etc. can be used as another option. If parallel to the posterior arch the introductory angle of the entry point should be at least 3 mm cranially from C1/2 joint for a safe 3.5–4.0 mm screw placement. Upward trajectory has to be used if the available posterior vertical working distance is smaller than normal. Using cranial trajectory inclination we have to take care not to encroach the C0-1 joint (Fig. 6.10). Most often, we use a trajectory of 10°–15° medial and 10°–15° cranial with active lateral fluoroscopy directing drilling (Fig.  6.11). Considering the

Fig.  6.10  (a) Parasagittal CT reconstructed image showing the vicinity of C1 lateral mass screw to the atlantooccipital joint, (b) coronal plane reconstruction

6.2  Atlas as an Anchoring Structure

77

a

b

Fig. 6.12  Axial CT scan of correctly bicortically introduced C1 posterior lateral mass screw

Fig. 6.11  Schematic drawing of C1 lateral mass posterior screw trajectory: The screw has partially smooth shank. (a) Screw is introduced in medial angle between 10°–15°. (b) Trajectory is tilted cranially in angle 10°–15°.

dominance of the left hemisphere in right-handed people, the author always prefers to put any hardware jeopardizing the left side vessels, last. The length of screw trajectory is defined by the position of the entry point and angle of introduction. Because of ovoid shape of the lateral mass and its medial tilt in transverse plane, (more pronounced in upper facets than in the lower) we can expect that the screw thread length will not correspond with the anatomical length of the lower facet pillar. Using the depth gauge helps us confirm or correct our measurements. The aim is to fully accommodate the screw thread within the bone of lateral mass and to have a smooth contact with C2 root, ganglion, and surrounding veins. Therefore, smooth shank screws with different thread versus shank ratio are used (often 60:40). The polyaxial screw head has to be located freely behind the posterior atlantal arch to enable its multidirectional movement and connection rod attachment. Most frequently we use 4 mm smooth shank screws 30–38 mm long. The final screw positioning is again always checked by lateral fluoroscopy. We prefer bicortical screw purchase because the pullout force is higher than in monocortical placement (Figs.  6.12. and 6.13). However, in our opinion, not

Fig. 6.13  Screw penetration of anterior C1 lateral mass shown on 3D CT

only pullout force is important; especially, in longer constructs, the vertical, horizontal, and rotational stabilities of the screw play an important role. As a simplified example, we will use a rod and brick analogy. The rod is introduced in a predrilled hole of a slightly larger diameter in a wall made from hollow bricks. The resistance is weak if only placed into the hollow portion of the brick; however, its side load resistance can be much improved when the opposite brick wall is drilled and the rod is “bicortically introduced” even though the effect on pullout effect may be negligible. The bicortical screw tip should not overrun the anterior mass surface more than one thread.

78

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

Fig. 6.14  Artist’s drawing of anterior C1/2 transarticular screw fixation according to Lesoin

The slightly medial trajectory and preoperative image analysis helps us avoid ICA and hypoglossal nerve injury.

6.2.2 Anterior C1 Lateral Mass Screw The screws are placed in the C1 lateral mass anteriorly in two modalities. First, the transarticular anterior screw fixation serves as a monosegmental C1-2 fixation. This can be eventually extended to fix the occipital condyle as well [52, 53]. The other possibility is using horizontal mass transfixation as anchorage for screws fixating either plates or mesh cages. The anterior lateral mass of the atlas was first used as anchoring structure by Lesoin et al. [141] who performed C1-2 anterior transarticular fusion in six patients (Fig. 6.14). Later, this technique was advocated by others in the case of C1-2 instability secondarily to odontoid fractures [11]. Koller et al. has suggested a different trajectory for the same purpose (Fig.  6.15) [131]. The horizontal trajectories are preferred in the case of C1 lateral mass cranial fixation of plates and mesh cages after odontoidectomy or other anterior decompressions [95, 118, 197]. Similarly, the plate used either for reduction and fixation of isolated atlas fractures [21, 195] or atlas split in deformity [108], can be fixed to the lateral C1 masses with horizontally introduced screws (Fig. 6.16).

Fig. 6.15  Schematic drawing of different trajectory suggested by Koller et  al. for anterior AA screw fixation. The screw entry point is localized at the base of anterior C2 body

6.2.2.1 Anatomical Background For proper atlantoaxial transarticular screw purchase, high anterolateral or oblique approaches are used preferentially; however, when planning horizontal screws, transoral exposure is necessary. The anatomical guidance and surgical technique for transarticular screw placement is described elsewhere in this text. The surgical anterior surface of lateral mass can be fully exposed only transorally. From anatomical works of Kandziora et al. [123] and Ai et al. [3] we can find that the maximal safe lateral exposure of anterior lateral mass is 20  mm laterally from the midline. The VA is located another 5 mm far laterally. The anterior lateral mass is trapezoid in shape having mediolaterally an average length of 15 mm. Its medial height is 9  mm and lateral one 22  mm. When calculating the screw diameter of 3.5  mm, Kandziora found a safe trapezoid zone 13.3  mm long (respecting the shape of anterior surface) with a medial height of 4.1  mm and lateral 12.9  mm. This, in fact, means that the screw entry point should be located in the middle of lateral mass. The trajectory angle of drilling should respect the lateral mass’  outward inclination. This angle was established as 20° by Kandziora et al. [123] and 12° by Ai et al. [3].

79

6.3  Axis as an Anchoring Structure

plate techniques. This may prove to be especially helpful if the plate is stable enough and gives us the opportunity to reduce kyphosis and allows avoidance of a second posterior stage surgery. Use of this technique is also applicable in anterior constructs in tumor surgery.

6.2.3 Posterior Arch of Atlas Intralaminar Screw

Fig. 6.16  Schematic drawing of anterior plate fixed to C1 by horizontal screw

6.2.2.2 Surgical Technique The anterior surface of the C1 lateral mass is exposed transorally. The atlantoaxial joint can be identified easily at its caudal border. Often difficult but possible is to palpate the cranial atlantooccipital joint. Often the anterior arch of atlas is resected, thus giving us information about the medial mass wall. The most dangerous aspect is to establish the lateral boundaries because of the vicinity of the VA. Despite the knowledge that the VA should be at least 25 mm away from midline, we have to analyze each case especially if the midline (tubercle of C1) is missing or resected. The best choice is to analyze the preoperative CTA when available; however, meticulous subperiostal microtechnique is mandatory in every case. The entry point is located in the middle of lateral mass and the drill passes in a lateral angle of 12°–20°, respecting laterally the midportion of the C1 mass on lateral fluoroscopy. We never drill bicortically because the VA located in VA groove posteriorly can be injured. Usually, the pilot hole is tapped and 3.5 mm screw monocortically introduced.

This type of screw anchorage was first used by Floyd and Grob [63] to transfix the bone graft in the cases of congenital or iatrogenic posterior arch deficiency when the sublaminar wire could not be used. They used this technique in five patients. First, the edge of arch remnant was osteotomized and then carefully drilled and tapped to approximately 10–15  mm depth along its course. The autologous rectangular unicortical graft harvested from iliac crest was transfixed by 2.7 mm screw, wedged between both C1 arch stumps and the decorticated C2 spinous process. Later, similar technique using 3.5 mm screws was recommended by Donnellan et al. [48] as part of C1-2 fixation constructs. They began the procedure with a wide posterior midline arch opening to expose the medullary core and then drilled and tapped the cavity. The screws were introduced laterally enough to minimize overlap of the polyaxial screw heads.

6.2.3.1 Our Preference From a practical standpoint, we can conclude that this technique can help in rather rare situations where the C1 lateral mass screws cannot be used and we would like to avoid the extension of the fixation to the occiput. Especially in cases of concomitant posterior arch congenital deficiency and surgical posterior arch damage, intralaminar C1 screw can be a valuable choice. The anatomical position of the VA, thinning of the proximal part of posterior arch by its groove, and possible anatomical variation have to be encountered in our surgical consideration.

6.2.2.3 Our Preference

6.3 Axis as an Anchoring Structure

Our experience with anterior fixation involving the C1 lateral mass is limited; nevertheless, we suppose that there is a place for further development of the anterior

The second cervical vertebra can be approached from all the sides. Currently, most frequently, it is exposed from posterior midline or anterior high lateral

80

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

approaches. Less often is transoral access performed and rarely used is lateral dissection. Screws placed only in the C2 vertebra are used to surgically treat some types of trauma, but more often the screws introduced into the C2 vertebra serve as a part of a longer construct. Transarticular C1-2 screws can be used as a standalone fixation to stabilize the atlantoaxial complex, but nearly all the screws introduced to C2 vertebra can be attached to longer fusion systems involving the occiput, C1, and/or the subaxial cervical spine. Their safe introduction and firm anchorage is therefore very important.

6.3.1 Pedicle Screw First of all, it is necessary to emphasize that the true anatomical pedicle is the connection between the C2 vertebral body and the posterior elements. This strong and very short structure leaves the body almost in the frontal plane (Fig.  1.4, Chap. 1). Therefore, the so called “transpedicular” screws are due to unique C2 anatomy frequently introduced not directly through the pedicle or the pedicle axis as in subaxial cervical spine but passes the true anatomical pedicle obliquely and often only partially. In fact, during modeling of the ideal trajectory on 3D navigational software, sometimes it is questionable if the screw pathway is more transisthmic or transpedicular and the final reached medial angle could be very different from those described in anatomical papers. When reading papers about C2 transpedicular screws one has to be aware because most of the authors are incorrectly labeling the pars interarticularis as “pedicle” or “pseudo-pedicle.” Robert Judet in France was the first surgeon who introduced the C2 transpedicular screw on September 19, 1962 (Ch. Mazel, personal communication). The first published description of the use of C2 transpedicular screw came from Leconte in a book about cervical spine injuries edited by Judet [137]. Judet used this technique to perform direct osteosynthesis of hangman type fractures. Unfortunately, this logical approach did not attract attention until Borne et al. [22] published a larger series on patients treated this way, 20 years later. Axis transpedicular anchorage, as a part of larger construct, was first mentioned by Roy-Camille who published a series of treated patients with C2-related

instability [192, 194]. Nowadays, this method is used not only for compressive osteosynthesis of C2 ring fractures [130, 136, 144, 218, 219, 224, 231] but especially as a part of short [14, 72, 94, 215] and long constructs [5, 201] where the C2 transpedicular screw is usually considered as the most solid anchorage. This philosophy was supported by Dmitriev et al. [47] who biomechanically tested 14 cadaveric specimens and found the largest insertional torque while introducing the screws into C2 pedicle when compared to other types of C2 screws. Also, the postfatigue pull-out strength was significantly larger than in other techniques. Although it is rarely reported up to now [5, 70], the main drawback of C2 transpedicular screw is potential arterial and/or neural injury. An often used argument that the transpedicular trajectory is less dangerous than transisthmic used in transarticular method [70, 94] is not correct in our opinion.

6.3.1.1 Anatomical Background Gupta and Goel [89] analyzing 100 cadaveric dissections recommended to introduce the screws into C2 pedicle in its upper-third with sharp medial inclination targeted to the anterior spinal midline (or anterior C1 tubercle) because the VA groove can occupy up to two-thirds of C2 lateral mass in 15% of patients. Analyzing the surgical pitfalls in his series of 160 patients, Goel [70] described 4 VA injuries during C2 pedicle drilling. Resnick et al. [186] modeling of the ideal screw trajectories on 3D and 2D thin sliced CT reformatted images found the same risk for transpedicular as for transarticular screw trajectories. More than 90% of sides of investigated vertebras offered more than 4 mm space for eventual screw purchase. They idealized the trajectory of “transpedicular” screws to nearly parasagittal angle of introduction; therefore, we suppose that the more correct final statement should be that the risk of VA injury is the same for long transisthmic screw as for transarticular screw. Yoshida et al. [248] very correctly argued that the transisthmic screw is passing above the VA lateral bending inside FT while the transpedicular screw is crossing the bone medially or superomedially to the course of VA. Such a difference hardly can be discovered and compared on 2D images. They have analyzed 3D CT images of 62 patients. The ideal trajectory was modeled in a computer navigational station. They measured the mean

81

6.3  Axis as an Anchoring Structure

space available for transarticular screw as 6.2 mm (SD = 1.4) and for transpedicular screw 6.1 mm (SD = 1.4) on 124 vertebral sides. These values did not differ significantly but they found significantly lesser maximal values for female vertebrae. The authors defined the space available for screw less than 4 mm as risky and less than 3 mm as unacceptable. In the groups of limited available space they determined if the most limiting factor was the height or width of available bone. They described 9.7% of transpedicular and 11.3% of transarticular trajectories as risky. Non acceptable bone space for transpedicular screw was found in 4% and for transarticular in another 3.2% of sides tested, respectively. The differences were not statistically significant, but often the screw trajectory judged as risky for one technique was risky for the other one also. In transarticular risky group the height of available bone was the limiting factor in 57.1 % of cases and the width in another 42.9%, whereas in the transpedicular risky group the only limit was the width of C2 pedicle in the place of VA groove. This work is the first available correctly stating that the anatomical risk of VA injury for transpedicular and/or transarticular screw purchase is on the same level. Data obtained from this paper also confirmed that for evaluation of risky situations the 3D modeling should be performed for transarticular trajectory; however, for planned transpedicular screw purchase, the axial CT could give us enough information. Measurements of Yoshida et al. were confirmed by Moftakhar et al. [162] who measured not only the bony space available for the screws but also the distance between the bone and the lumen of VA. They analyzed computer tomographic angiograms (CTA) of 106 patients and found 6.4 mm (range 2.09–13.20 mm) of osseous space available for screw at the VA groove level of C2 pedicle. The isthmus thickness was measured on an average as 5.62  mm (range 2.08–11.00  mm). The distance between the bone and the VA was interestingly more than 1.18  mm on an average in all measurements. However, one has to evaluate this value individually and patient specifically because of great variability, the range was from 0 to 4.94  mm and the distance was measured to intraluminal contrast media without calculation of the arterial wall thickness. They also did not find any correlation between VA groove/foramen size and diameter of VA. Several other anatomical studies were conducted to establish the guidelines for pedicle C2 screw placement

[55, 107, 243]. Although they studied the isthmus rather than the pedicle, they correctly measured the angles of possible screw trajectory. Xu et al. [243] stated that the average angles for a pedicle screw are 33° medially and 20°, rostrocaudally. They also estimated the possible entry point location and created a placement algorithm. However, the same group, while strictly respecting their guidelines, found an unacceptably high rate of cortical breach in their later study [55]. Standardized algorithmic approaches usually fail in C2 pedicle screw placement due to a great anatomical variability and also due to a number of possible screw trajectories. Therefore, Howington et al. [107] repeated the anatomical test on ten cadavers respecting the trajectory given by the surgical visualization of the pedicle. All their screws were correctly placed without cortical breach. The established medial inclination was on an average 35.2° and craniocaudal angle 38.8°. The notable, almost double, difference in craniocaudal angle measurements could be explained by the methodological variance in obtaining of their values. Xu et al. [243] used the line perpendicular to the axis of odontoid process as a reference whereas Howington et  al. [107] used the plane of C2 body endplate. Although some discrepancies could be found in pure anatomical works, Sciubba et  al. [201] confirmed clinically the prerequisite of Howington et al. in a large (single surgeon J.P. Wolinski) series of patients treated with C2 pedicle screws. They place C2 pedicle screws under a pure visual control with excellent clinical results. However, the same group recently published a retrospective analysis of 170 transpedicularly introduced screws (19.4% with fluoroscopical guidance) and found on the postoperative C2 pedicle focused coronal CT reconstructed images  25% of screw cortical breaches. Most of the cortical wall violations (67.4%) were lateral in FT with one recognized VA injury [5]. They also discovered that the correct screw placement is related to the surgeon’s experience. The average pedicle diameter measured on postoperative CT scans was found 6 mm in its thinnest portion, and the lateromedial angle of screw purchase was approximately 40°.

6.3.1.2 Surgical Technique Standard Technique The C2 spinous process, lamina, C2/3 facets, and lateral pars border are exposed through a standard midline

82

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

posterior approach. Depending on the extent of planned fusion, other structures such as the posterior arch of atlas, occipital bone, or subaxial spine may need to be exposed as well. In situations where an isolated pedicle screw is planned, the muscular attachments can be saved and the approach focused only to the planned screw entry point and important trajectory determining structures. Despite that, it is usually helpful to dissect out the entire pars of C2, superomedial aspect of the isthmus, and the pedicle. The dissection should be carried out in a subperiostal fashion to avoid venous bleeding and/or injury of C2 nerve root and ganglion, if they are to be preserved. Screw placement is carried out under fluoroscopic guidance or, with increasing frequency, with the use of navigational systems. Different authors describe different screw entry points along the pars interarticularis. In the majority of clinical papers, the angle of craniocaudal trajectory is related to the coronal plane crossing the odontoid process midline and the lateromedial angle related to sagittal midplane. In their publication, Borne et al. [22] advised to introduce the pedicle screw angled 20° cephalad and medially. Roy-Camille et  al. [193] proposed the entry point located in superomedial quadrant of articular process and direct the drill and screw 15° medially and cranially. In AO Spine manual [12], the geometrical middle of the pars interarticularis is recommended as a standard entry point. The trajectory of drilling and screw purchase is directed 25° cranially and 25°–35° (15°–25° in older edition), medially. Dickman et al. [42] proposed to locate the entry in the pars midline but only 2–3 mm above the lower C2 facet edge. The trajectory was suggested 20°–30° rostrally and medially. Levin et al. [142], while treating hangman’s fractures, suggests to start above the entry point for transarticular screw and to use biplanar fluoroscopy and direct visualization of internal pedicle side to guide the drill in correct angles.

of C2 lamina to direct the screw into the C2 vertebral body while respecting the medial border of isthmus and pedicle to determine the medial angulation. They did not use fluoroscopy and the key point was clear anatomical visualization of medial border of isthmus/pedicle. Postoperative, thin sliced CT done in prospective manner discovered only 15% of cortical breaches. Only two screws (2%) were evaluated as more than half of diameter cortical breakthrough. Both perforations were without clinical consequences.

6.3.1.3 Our Preference Using the C2 pedicle screw since 1993, we began with compressive osteosynthesis of fractures of the ring of axis as was originally described by Judet (Figs. 12.16 and 12.19, Chap. 12) and later we embarked on the use of this method in short C1-2 fusions (Fig. 20.11, Chap. 20) as well as in occipitocervical (Fig.  19.29, Chap. 19) and long subaxial constructs (Fig.  6.17).

Free Hand Technique Sciubba et al. [201] in their single surgeon (J.P. Wolinski) series of 55 patients treated with 100 C2 pedicle screws have done a thorough anatomical preoperative analysis of CT and MRI images to exclude those pedicles not large enough to accommodate a 3.5 mm screw. This was the case in 10% of analyzed pedicles. Their entry point was located more superiorly and laterally than usually recommended. They then used the ­cranio-caudal slope

Fig.  6.17  Lateral radiogram of circumferential cervical spine reconstruction for deformity. The transpedicular screw is a part of long posterior subaxial construct

83

6.3  Axis as an Anchoring Structure

The  comprehensive radiological workup always precedes the procedure. Plain lateral and transoral films serve us mainly for basic orientation; however, they can be helpful as a predictor of potential fluoroscopic visibility of the bony structures during the surgical procedure. Dynamic lateral radiographs can reveal potential instability and the effect of eventual UCS movement during surgical positioning of the patient. Although, axial CT images of C2 pedicle are usually sufficient to determine feasibility of a C2 pedicle screw [248], we insist on thin sliced CT images prior to any planned intervention. Currently, only three-dimensional CT imaging with software modeling of ideal screw trajectory (Fig.  6.18), can reliably demonstrate variants of bony anatomy and determine actual individual availability of bone for screw acceptance. An MRI evaluation, although used less frequently than CT, can also provide useful information in our opinion. It clearly demonstrates not only neural anatomy and pathology of interest but can also depict the exact course of vessels (especially, VA) without the need for invasive angiography. If an anomalous VA or ICA is suspected then a standard angiography or CTA (our preference) can be performed. Rarely, VA can be hypoplastic on the side of intended intervention simultaneously with a surprisingly large bony transverse foramen. This

allows for a pedicle screw placement with expected/ intended cortical breach without the risk of arterial injury (Fig. 6.19). Pre-procedural knowledge of such variant can be important if other type of fixation is not possible or fails during the procedure. We will usually plan our entry points and trajectories virtually based on preoperative imaging. Virtual planning is only avoided in situations when unexpected extension of construct to C2 is required intraoperatively.

6.3.1.4 Our Surgical Technique The direct visualization of superomedial aspect of isthmus and pedicle is of utmost importance in guiding screw trajectory without navigation. The medial border of the isthmus and lateral lamina is always exposed, and if possible, the dissection extends to involve the medial surface of the pedicle (Fig. 6.20). Such exposure gives us the correct lateromedial angle and often we find that more than 30° (usually round 35°–40°) of medial inclination is necessary to introduce the drill. Direct visualization of the medial anatomical border of the pedicle practically eradicates the risk of injury of medially located neural structures. The entry point becomes obvious once the intended trajectory given

Fig. 6.18  Navigational plans created in computer station for C2 transpedicular screws purchase in a patient with bilateral high riding vertebral arteries showing high risky trajectories on both sides. (a) Plan for right pedicle. (b) Plan for left side

84 Fig. 6.19  Discrepancy between larger FT osseous diameter and smaller VA caliber in two different patients. (a) Parasagittal CTA 2D reconstruction. (b) CTA in 3D

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

a

Fig. 6.20  Artist’s drawing of entry and exit points of C2 transpedicular screw with emphasize of the area of necessary direct visibility

by previously described anatomical landmarks is established (Fig.  6.21). It is also necessary to calculate the screw diameter to pass safely directly under the isthmus/

Fig. 6.21  Change of C2 transpedicular screw trajectory requires simultaneous change of entry point position. (a) Without entry point change the screw comes closer to dangerous structures. (b) With change of the screw entry point respecting the narrowest part of planned trajectory (fulcrum) the risk can be decreased

b

pedicle ridge superomedially from VA groove toward the C2 body without significant penetration of the cortex. The screw entry is frequently more cranial and sometimes, more lateral than the middle of the lateral mass. It is not absolutely necessary to use fluoroscopic guidance for the initial drilling, but we recommend it as the upper ridge of pedicles can frequently be seen on lateral view and parallel placement can be ensured. Fluoroscopic information is not only useful in determining the correct craniocaudal angle of the screw but also allows for an estimation of screw length. The last parameter is due to the wavy C2 anterior surface rough only. We use a high-speed drill to make a small pilot hole at the entry point. This prevents a slip of the hand drill from the desired entry and allows pure concentration on trajectory. A drill-guide covered 2.5 mm drill used for the initial drilling enables one to feel the opposite cortex and thus minimize the risk of injury of structures located anterior to C2 body. From a technical viewpoint, we feel that higher drill speed gives better feedback of cortex penetration at the front. After a ball tip probe check, the

85

6.3  Axis as an Anchoring Structure

hole is usually tapped. In our opinion, tapping is useful in preparing the hole to match the screw thread exactly and thus allowing a firm anchorage. This is important particularly in axis ring fractures where compression is desirable. On the other hand, tapping can result in cortical breach and VA injury. One should be careful, if significant venous bleeding is encountered after initial drilling as this is likely a result of injury to the venous plexus around VA. In such a situation, we do not tap the hole but simply place a self-tapping screw once the hole is checked with a blunt-tip probe. Screw length can be determined during virtual planning and confirmed during tapping with calibrated instrument and/or with a probe gauge. If instruments are not navigation enabled, there may be a slight difference between virtually planned screw length and the actual measured one. This may be due to some discrepancy between actual entry points and screw angles. Screws need to be left somewhat proud if polyaxial screw heads are used to allow their free movement. This facilitates an easier rod application. We usually end up using screws that are 26–38  mm long when part of a construct. However, when treating hangman type fractures, the compression screw is usually shorter than estimated preoperatively and only partially threaded to allow for compression of the fracture gap. It is usually possible to recognize if a shorter screw is required during the compression maneuver and make appropriate changes as necessary. Once again, fluoroscopy can be of value here. We evaluate correct screw placement and construct position by means of a postoperative CT scan (Fig. 6.22 and Fig. 6.10). If significant cortical breach is identified and concern for VA artery injury exists, we obtain a CT angiogram to confirm vessel patency and exclude possible pseudoaneurysm or fistula. Postoperative CT is an important educational tool and an audit of one’s technique, especially if not performing such constructs frequently.

Fig.  6.22  Postoperative coronal CT reconstruction showing correct intra-pedicular screw purchase

6.3.2 Long Pars Interarticularis Screw – Transisthmic Screw Screw passage through the C2 pars interarticularis was first suggested by Friedrich Magerl, Austrian orthopedic surgeon working in St. Gallen, Switzerland, as a part of transarticular screw fixation for the treatment of C1-2 instability [152]. The first surgery using a C1-2 transarticular screws was performed in 1979. Although it was originally suggested for atlantoaxial fixation, it was later also utilized as a part of longer constructs and/or simple C2 anchorage point [86, 199]. Given the potential risk of VA injury, many authors used a shorter version of this trajectory as part of longer constructs and thus avoiding the transverse foramen of C2 [41, 181, 213]. Many surgeons advocate the use of alternative fixation methods due to the potential risk of VA injury with a transarticular screw. However, we believe that the majority of VA injuries do not represent a failure of the method. They are usually a result of poor decision making or preoperative planning. A thorough preoperative anatomical analysis is absolutely essential prior to instrumentation of C2 pars and can avoid potential disasters while allowing for solid fixation.

6.3.2.1 Anatomical Background According to Magerl’s original description, the screw trajectory passes through pars interarticularis of C2 vertebra in the plane parallel to the sagittal one, reaching the midline of posterior half of C2 superior articular process. The VA groove below the anterolateral portion of C2 superior facet potentially narrows the path of both the pedicle and the pars interarticularis and may preclude screw placement. Initially, exceptionally large C2 transverse foramen was described in several case reports as a rarity. This was usually due to bony erosion by tortuous VA [34, 236]. However, with increasing interest, VA anatomical variability was found to be more common than initially thought and was described by many authors [120, 151, 153, 179, 222, 223]. The majority of those studies were done on cadaveric spines [19, 55, 153, 243] or evaluating bony structures (i.e., transverse foramina) on axial and reconstructed CT images [51, 120, 168, 179]. The true

86

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

course of the VA could not be established in such a study design. Taitz and Arensburg [222] analyzed 300 dried specimens and found 33% incidence of C2 transverse foramen erosion (21% moderate, 12% marked). Although, no VA injuries as a result of isthmic C2 screws were published before 1995, Dull et  al. [51] recognized correctly the potential hazard of an anatomical VA groove variability at this level. Good resolution 3D CT reconstructions did not exist at that time and therefore, the authors suggested obtaining images in an oblique plane to better visualize the pars. In order to visualize the potential screw path on CT images, complicated positioning of patient’s elevated torso was required in a maximally tilted CT gantry to respect the screw angle. This method was, obviously, not widely accepted. In the comments to Dull’s article, Paul R. Copper (New York) mentioned two deaths caused by misplaced screw related to VA injuries in the US. Paramore et al. [179] suggested reformatting traditional CT axial and sagittal images to visualize isthmic bone bridge 2–3  mm from the internal lateral canal border and thus evaluate its suitability for a screw. The anatomy would not be suitable for a screw in 17 of 94 (18%) patients on at least one side. In three subjects (3%), this anatomical restriction was seen bilaterally. Another five examined vertebrae were evaluated as risky but the exact anatomical measurements were not mentioned in their study. Their conclusion was that in 18–23% of patients, the transarticular C1-2 screw might be impossible or risky. There was no gender predominance. The authors suggested obtaining a preoperative MRI and/or contrast-enhanced 3D CT to trace the VA course exactly if doubts exist after regular or reformatted CT images. The “pedicle” width and height were studied in cadavers by Ebraheim’s group. In fact, they measured the isthmus and its width was 7.9–8.6 mm (female × male) and height 6.9–7.7  mm (f × m) at the level of transverse foramen [55, 243]. The anatomical nomenclature (Fig. 1.4, Chap. 1) was corrected by the same authors in their later work [54]. Similarly, Madawi et al. [151] wrongly calling pars the “pedicle” measured mean isthmus width 7.8  mm (range 3.4–12.2 mm) and height 7.9 mm (range 4.7– 12.4 mm) on 50 cadaveric specimens. They stated that transarticular screw would be hazardous in 22% of tested vertebrae. Also, the internal height of lateral mass was considered as an important restriction of screw introduction especially if the measured distance

is less than 2.1  mm. The same group also published clinical results of 61 patients treated with transisthmic screws [150]. Mostly due to anatomy distorted by disease or previous surgery, 14% of screws were placed incorrectly and VA injuries occurred in 8% of patients. In their pleasantly honest paper, they proposed that the majority of VA penetrations were due to incorrect (too low) trajectory inside the C2 vertebra. They emphasized that the anterior atlas tubercle often used as a trajectory target endpoint on lateral fluoroscopy can be lower than expected from anatomical studies. This is especially true in rheumatoid arthritis patients with settling of atlas. The other cause of an error when using anterior tubercle of C1 as the target, is its absence after transoral resection. The estimated risk of C1/2 transarticular screw misplacement after transoral odontoidectomy was as high as 55%. Jun [120] tried to establish the risk ratio of transisthmic screw using the sagittal reconstructed images of 64 healthy volunteers. He used reconstructed images of isthmus in strictly parasagittal plane 3.5 and 6 mm laterally from internal spinal canal wall. Modeling the longest trajectory line from the ridge of lower C2 facet, crossing the posterior part of superior facet, he defined the point of intersection as a distance from posterior rim of superior facet. Moving the line anterocaudally he virtually reached the transverse foramen in both (3.5 and 6 mm) parasagittal planes and measured the space available for screw (SAS). If the SAS distance was less than 3.5 mm he defined the situation as non-acceptable and if it was less than 4.5 mm as risky. Logically, the more anterior and lateral trajectories represented higher risk of transverse foramen perforation. Among 64 tested volunteers (128 sides), 4 sides were risky or unacceptable in the 3.5 mm and 21 sides in the 6 mm distant planes. Extrapolating this to per patient risk ratio this means that 6.3% of patients are in danger if the 3.5 mm distant plane is used and 32.6% if the screw is introduced in the longest parasagittal trajectory 6  mm laterally from the canal border. However, this study has some practical limits as a strictly parasagittal trajectory is not mandatory and the entry point can also be adjusted to facilitate an ideal screw trajectory. Solanki and Crockard [208] suggested transferring anatomical knowledge and preoperative CT scans to the computer aided design (CAD) program and plan the ideal trajectory before the procedure. Then they extrapolated the obtained information to the lateral fluoroscopical view. They also determined the safe

6.3  Axis as an Anchoring Structure

limits of lateromedial screw inclination as 0–14°. It is not clear how frequently or how successfully this method was used in clinical practice. To minimize VA injury, Goffin et al. [74] proposed a cheaper alternative to image guidance. They used custom-made polymer templates based on preoperative CT planning with imbedded stainless steel drill guides. This technique was only applied twice in clinical setting. Mandel et al. [153] measured 205 dried C2 vertebrae to determine the isthmus height and width at the level of transverse foramen. They found the mean isthmus width of 8.2 mm in male specimens and 7.2 mm in female vertebrae (3.9–14.7 mm). Five subjects (2.4%) had the isthmus width less than 5 mm at least on one side (i.e., not large enough to accommodate a 3.5 mm screw under standard fluoroscopic guidance) and those vertebrae were evaluated by CT to obtain further details. The obtained mean isthmus height was 8.6 mm in male and 6.9 mm in female samples (2.8–14.7 mm). Twenty-four vertebrae (11.7%) had one or both height measurements less than 5 mm. They also demonstrated significantly larger left-sided isthmi and significantly smaller dimensions in female specimens. In selected cases of repeated CT measurements, they found only 1 mm difference in comparison to values obtained by electronic caliper. They concluded that approximately 10% of population might be at risk of VA violation during the placement of an isthmic screw. Bloch et al. [19] tested image-guided placement of isthmic screws on 17 cadavers. Using the standard anatomical requirement of 5  mm of bone available for isthmic screw, they found that 20% of specimens would not be large enough to accommodate a 3.5 mm screw. They proposed that, with computer aided virtual navigation, the available isthmic bone requirement can be decreased to 4 mm. In other words, if the isthmus diameter in perpendicular plane to the ideal computermodeled trajectory is more than 4 mm, then the 3.5 mm screw can safely pass through. They thus concluded that, with the help of image guidance, the rate of isthmi not suitable for a safe screw placement can be reduced from 20% to 5.9%. As mentioned earlier, Resnick et al. [186] tested the ideal trajectories for transarticular and pedicle screws in their series of 50 standard axial CT investigations of trauma patients and 10 selected 3D CT images in patients with UCS anomaly. In fact, their attached figures demonstrate that they probably truly tested the

87

long isthmic screw trajectories rather than the actual pedicle screw. They also noted that the angle of pedicle screw is roughly parallel to the C2 spinous process. This is obviously incorrect from an anatomical point of view. There were no significant differences in 4 mm screw acceptability for both trajectories and the isthmus height was, on an average, 6.6  mm (SD = 1.8). Despite the above, some interesting facts became apparent in this study. First, it was demonstrated that different entry points and directions can be used to pass the C2 isthmus safely. Second, the accuracy of preoperative planning of the possible screw trajectory can be improved with the use of three- rather than twodimensional thin-cut CT images. Igarashi et  al. [112] found that there were differences between the two sides in 45% of 98 dried C2 specimens and that 20% of pedicles (he actually measured the isthmus) had diameter smaller than 3.5 mm. This high frequency of isthmus low profile can be explained by generally smaller Japanese population. Neo et al. [168] suggested changing the screw trajectory in cases of aberrant transverse foramen anatomy. They defined as “high-riding” VA anatomical situation where the isthmus height measured on CT sagittal reconstructions was less than 5 mm and/or the internal height of C2 lateral mass less than 2 mm. This situation occurred unilaterally in 7 (26%) of their 27 consecutive patients planned for atlantoaxial fusion. Using the most posterior and most medial possible trajectory, they successfully placed screws even in the seven mentioned patients with only two intraforaminal cortical breaches but without consequent VA injury. They have also constructed special aiming device enabling the most posterior trajectory. However, the extreme cranial angle tilt resulted in screw penetration into C0-1 joint in four of the seven with high-riding VA. Finally, they correctly concluded that the internal height of lateral mass is not as important as the isthmus width and height. The internal lateral mass height becomes important only when the VA groove is located more medially than usual. Lee et al. [139] also tried simulating a different trajectory on a computer and then used it in seven patients with high-riding VA. Mostly, their screws were placed using a more cranially located entry point and more medial trajectory directed in a flat sagittal angle. The problem with flattening of the angle is not only to avoid the VA groove but to adequately anchor the screw in lateral mass.

88

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

The MR flow void or MRA can show an anomalous course of VA; however, the bone-vascular relationship cannot be well appreciated on MR imaging. Therefore, Theodore et al. [226] used CTA to better visualize the VA course in FT in three patients. Clearly, it is not only the course of the VA but also the VA-bone occupancy ratio that are important to assess. Cacciola et al. [26] studied ten cadavers and found that VA completely filled up the bony groove only in 30% of specimens. In the majority of specimens, there was a space between the artery and bone filled by periostial tissue and venous plexus. The mean VA occupancy within the groove was 79% (range 34–100%). This very important fact was reproduced by Moftakhar et al. [162] as mentioned earlier. They performed detailed CTA ­measurements in 106 patients with stroke or trauma presentations. Their motivation to do so was the fact that the majority of previously mentioned dimensions were obtained from cadaveric specimens or evaluations of only osseous VA groove on CT scans rather than imaging the actual real size and location of VA itself (Fig. 6.19). The potential risk of neurovascular injury anterior to the C1 or C2 body is similar to those described for C1 lateral mass screws and pedicle screws if bicortical screw purchase is preferred. Jeanneret and Magerl [116] have seen temporary bilateral hypoglossal palsy, which they were not able to explain. Nonetheless, they ended up replacing one of the screws as it was 4 mm longer than desired.

6.3.2.2 Surgical Technique Friedrich Magerl [116, 152] suggested to place the screw entry point on C2 caudal articular process 2 mm laterally and 3 mm cranially from the internal edge of C2-3 facet and to follow strictly parasagittal trajectory to reach posterior or middle thirds on inferior C1 facet surface. The target point on lateral fluoroscopy was the anterior tubercle of atlas in a reduced position. He used a 2.5 mm drill bit and always tapped the pilot hole for a cortical fully threaded 3.5 mm screw. If lagging was required, they over-drilled the proximal part of the hole with a 3.5 mm drill bit. The first drill was always left in place until the second pilot hole was drilled to avoid eventual dislocation. He emphasized the importance of visibility of the upper isthmus edge up to the atlantoaxial joint capsule and always used a bone

adjunct, usually in the form of Gallie’s graft or intraarticular bony chips to enable fusion. Magerl and his followers have described most of the surgical tricks in their initial descriptions although they were later attributed to other authors [83, 116]. For example, they have initially described: the possibilities of manual atlantoaxial dislocation reductions, how to perform the intraarticular bone fusion in cases of missing C1 posterior arch, how to achieve the correct sagittal working angle by means of cranial C2 spinous process traction, and/or creating caudal stab wounds for trocars. Ongoing widespread use of Magerl’s technique led to more frequent descriptions of complications. The first VA injury caused by transisthmic screw was published by Sasso et al. [199]. Madawi et al. [151] has described 8% of VA injuries without any mortality. Wright and Lauryssen [240], in their retrospective survey analysis of 1,318 patients treated with transarticular screws, estimated the risk of VA injury 4.1% per patient or 2.2% per screw inserted with a 0.2% per patient incidence of subsequent neurological deficit and 0.1% mortality. However, the actual incidence of VA penetration might be higher because of a low survey response (25.1%). Gluf et  al. [69] later reported 2.6% incidence of VA injury and a 0.5% mortality rate as a result of transarticular screw fixation in 191 patients. Recently, the same group of authors published their series enlarged by 78 additional patients (269 total) emphasizing that increased experience and more careful radiological planning can dramatically decrease the frequency of VA injuries [61]. In this largest published series worldwide, they evaluated the risk of VA injury as 1.9% per patient and 1.2% per screw having 13.3% sides anatomically unable to accept the transarticular screw. Conversely, there are large studies available describing no VA injury [86, 154]. In the meantime, there were a lot of other studies describing the use of transarticular screw and its modification [36, 44, 69, 86, 116, 154].

6.3.2.3 Our Preference So far there is no published study substantially modifying the originally and, in our opinion, genially described technique by Magerl. The technical development since then only allowed performing this surgery more precisely, less invasively, and using hardware

6.3  Axis as an Anchoring Structure

made of better materials. The crucial step forward was the evolution of imaging techniques. The precise knowledge of anatomy of the isthmus of pars interarticularis and VA course can substantially decrease the frequency of complications if correct screw trajectory is preplanned and if those who cannot accommodate the screw are primarily excluded. The main drawback of this technique is that the screw passage cannot be controlled by naked eye and the sagittal angle is not very comfortable and sometimes even not achievable. Interestingly, there is an important difference between VA injury risk established from cadaveric anatomical studies or 2D CT reconstructions (up to 23% risky) and those where thin sliced CT with threedimensional reconstruction were used (6% of risky or unacceptable). The VA injury risk is very variable among the published studies [69, 86, 151]. This fact brings about a suspicion that approximately one half of these events are not only the technique but possibly more surgeon related. We suspect that this failure can hypothetically be caused by an incorrect angle of placement in sagittal plane or wrong preoperative planning and/or decision

89

to use this technique on unsuitable patients and/or sides. Many authors warn about the potential injury of the spinal canal contents. It needs to be stressed that no spinal cord injury has ever been described in the literature when placing pars or pedicle screws in C2. This fact can be explained by the usual direct visualization of isthmus. Also, the medial cortex can be exceptionally breached without a significant risk of dural/neural injury. Such or even quadricortical (intra-canal) screw purchase in the case of pedicle screw placement is an important bailout technique in certain situations where other methods fail or are not possible (destructive arthropathy in rheumatoid arthritis patients). It is the circular bone stock available for round-shaped screw in the narrowest place of its trajectory that we find critical (Fig. 6.23). The SAS can only be measured on 3D models in the plane perpendicular to its planned trajectory. The preoperative modeling can be done on the CT workstation or with the help of navigational machine and its software (Fig. 6.24). The SAS should be at least 5 mm in diameter to allow a safe passage of a 3.5 mm screw while using the standard technique guided by anatomical landmarks and fluoroscopic guidance. If a

Fig. 6.23  Space available for long transisthmic screw modeled on navigational computer station. (a) Model of 3.5 mm screw passage through the isthmus depicting critical SAS in coronal plane. (b) Model of 3.5 mm screw showing critical SAS in parasagittal plane

Fig. 6.24  Correct transarticular screw trajectory plan prepared on navigational workstation. (a) 3D image. (b) Parasagital projection. (c) Tilted axial section

90

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

4  mm screw is to be used then the SAS must be at least 6 mm. In cases where image guidance, real-time CT or isofluoroscopy are used, the SAS diameter can be 1 mm less. In cases where morphology makes the screw placement impossible (worst case scenario), it has to be recognized on preoperative planning and other options need to be considered. The screw trajectory can vary substantially. If the aim of the screw is to fix C1 in reducible dislocations, then it has to pass through the posterior half of C2 upper facet. In cases of irreducible anterior dislocation or settling of the atlas, the screw trajectory has to primarily respect a safe passage through the C2 pars and only secondarily consider if C1 anchorage is sufficient or even possible. Targeting of anterior tubercle of the atlas is never as important as achieving a safe C2 transisthmic trajectory. At least 5 mm of screw should pass through the bone of C1 lateral mass to achieve a firm anchorage. Change of trajectory due to dislocated C1 or high located VA groove can result in undesirable deviation of the screw. Thus, for example, the hypervertical trajectory can penetrate the C0-1 joint and in such a situation, an adequately shorter screw has to be selected. Medial or lateral tilt can be acceptable if sufficient anchorage of C1 is possible and if the screw does not leave the middle third of the lateral mass in coronal plane. An intentional decrease in cranial angle could potentially be the most dangerous. If such caudal tilt is necessary, the screw entry point needs to be adjusted more cranially; otherwise, the VA groove would be violated (Fig. 6.25). Similarly, a more lateral

trajectory can increase the risk of VA tear if the pathway is not changed appropriately according to 3D modeling. Too long a screw can injure the anterior retropharyngeal structures. Internal carotid artery, hypoglossal nerve, and the posterior pharyngeal wall can be injured if the tip of bicortical screw is too lateral and protrudes more than 5mm out of the vertebra. More freedom exists if the C1 anchorage is not intended to fix C1 but rather is a part of a longer construct. In such situations, a nearly pedicular trajectory can be selected. Alternatively, a concomitant caudal tilt of the screw with a more superior location of entry point can be advantageous in bicortically or monocortically placed isthmic screws. With respect to the VA, one has to pay attention not only to the shape and direction of pure bony canal or the groove of VA but also to the actual artery itself as it occasionally occupies only a part of the foramen/groove. This becomes obvious in cases of arterial asymmetry due to hypoplastic VA as seen on a CTA (Fig. 6.19). The other choice is to place only a short pars screw as described below.

Fig. 6.25  Schematic picture of possible trajectories of C1-2 transarticular screw demonstrating that changing of trajectory must be accompanied by the change of screw entry point. (a) Tilting of the screw in sagittal plane without change of entry point can lead

either to inadequate C1 lateral mass bone anchorage (cranial tilt) or undesirable VA vicinity. (b) Synchronous change of trajectory and screw entry point is much more relevant

6.3.2.4 Our Surgical Technique Patients treated for C1-2 instability are carefully and often fiber-optically intubated and electrophysiological monitoring electrodes are attached in the initial recumbent position if needed. Prone positioning is performed gradually holding the head in neutral position sometimes secured with a hard collar or halo vest. The

6.3  Axis as an Anchoring Structure

operating surgeon is always responsible for this maneuver and most often he is the person holding the head and directing others to rotate the patient. Three-point head fixation (e.g., Mayfield clamp) can be advantageous to allow for reduction of atlantoaxial dislocation as well as positioning the C2 vertebra appropriately for isthmic screw placement (40°–50° tilt). This maneuver flexes the head while keeping the cervical spine straight. Lateral fluoroscopy is always used during positioning of the patient and the planned trajectory angle is tested with metal probe (Fig.  6.26). Occasionally, barrel chest or fixed hyperlordosis do not allow for the correct trajectory. This is crucial and one must not be satisfied if an absolutely perfect angle cannot be achieved. It is also necessary to figure out the angle achieved by extended instruments (drill etc.) as the dorsal chest can dictate how much upward angle is possible. The other trick is the “landing Concorde” position of cervicothoracic junction. The head and UCS should be located higher than right cardiac atrium to avoid excessive venous bleeding from potentially injured plexuses around the C2 nerve root. It is necessary to emphasize repeatedly that the positioning of the patient before incision has the same substantial value for final success as a clean and anatomically clear surgical approach. We should do everything possible to increase patient’s safety, and also our comfort to be concentrated to the critical parts of the procedure. As neurosurgeons, we feel that any unnecessary bleeding results in blurring of anatomical structures and we therefore insist on meticulous hemostasis from the beginning of the procedure. We use subcutaneous local anesthetic with adrenaline (epinephrine) infiltration

Fig. 6.26  Testing of achievable trajectory angle with metal probe on lateral fluoroscopical view preoperatively. (a) Screen view. (b) Video print

91

and electrocautery to decrease bleeding from skin and subcutaneous tissue. It is important to maintain midline on deeper dissection through the nuchal ligament and the muscular fascia as this will prevent blood loss and allow for subperiostial dissection of neck muscles off the spinous processes and laminae. The C2 spinous process is often bifid, readily palpable, and is the largest one in the UCS thus allowing orientation. The other important point of orientation is the posterior tubercle of atlas. When placing the long isthmic screw, we expose the posterior arch of atlas, C2 laminae, C2/3 joint, and adequate part of occiput. If a simple C1-2 transarticular screw is to be placed with intra-articular bone fusion only, the approach can be less invasive with sparing of some of the muscles attached to the C2 spinous process. However, when atlantoaxial reduction and posterior midline graft fusion are required, the exposure needs to be extended to see all previously mentioned structures. When placing an isthmic screw, we believe it is critical to expose the whole C2 pars interarticularis and the cranial ridge of C2 isthmus. The desired sagittal angle for drilling is very often achievable only via a caudolateral stab wound incisions. The entry points most often located 2 mm laterally and 3 mm cranially from lower medial aspect of lower C2 facet are predrilled with a high-speed drill to avoid drifting of the drill. The trocar is introduced through the caudal stab incision and long drill in appropriate sleeve is passed through to reach the entry point. Lateral fluoroscopy is now brought into the field and the sagittal angle checked. We usually draw a line on the fluoroscopy display to extend the line of drilling and thus better predict the target. Frequently, the

92

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

anterior atlas tubercle is the endpoint of our drilling. This is, however, not true in dislocated situations as was emphasized earlier. Before the drilling begins, the superomedial ridge of C2 isthmus is visualized with a Penfield dissector, thus enabling the surgeon to directly monitor the lateral wall of the spinal canal. Usually, a parasagittal plane or slightly medial (up to 15°) trajectories are used. If known, the screw hole is pre-drilled on the non-dominant side first. Otherwise, we usually choose the right side first. The drilling is discontinued when the end cortex (most often, anterior C1 lateral mass) is penetrated. If no excessive bleeding is seen after the drill removal the tap is introduced through the same trocar sleeve and the hole is tapped. This is the most dangerous part of the procedure. The tap is a very sharp instrument and can injure the VA more easily than the drill and the trajectory of the original drill hole can be changed unintentionally. Because of the mentioned drawbacks, some authors do not recommend use of a tap and continue with self-tapping screw. Nevertheless, we believe that tapping can increase the strength of screw anchorage and also avoids unintentional redirection as is possible with self-tapping screw. A full-threaded or half-threaded screw is finally introduced via the same trocar. A specially developed selfretaining screw driver can be very helpful during the trocar passage of the screw. All steps of the screw placement are monitored by lateral fluoroscopy. In our series of 100 transisthmic screws, we have seen two VA injuries. It was always during tapping. One has to consider this complication whenever significant venous bleeding is encountered out of the drill hole. This can be a warning sign caused by injury of venous plexuses surrounding the VA. If the VA is violated, pulsatile arterial bleeding may be seen out of the drill hole. In such situations, a second attempt to pass the screw would be ill-advised, in our opinion. The transisthmic screw can also be introduced under image guidance. One has to consider that only C2 vertebra is registered. We only use image guidance in borderline cases where we believe that transisthmic screw is necessary and where the SAS is between 4 and 5 mm.

6.3.3 Short C2 Pars Interarticularis Screw The use of a short screw in the C2 pars interarticularis to avoid the vertebral artery was suggested by Resnick

and Benzel, in 2001 [185]. They have chosen this t­echnique in a very obese woman with odontoid ­pseudoarthrosis where the low angle of a transarticular screw was impossible and patient’s obesity made fluoroscopy guidance difficult. They used an entry point located 14 mm above the C2-3 facet line, approximately in the center of the pars. A 20 mm screw was introduced in a sagittal angle parallel to the spinous process. Unfortunately, as other authors, they called this screw a “pedicle” screw despite it being placed in the actual C2 pars in terms of anatomy. The construct was finished by a rod connection to C1 lateral mass screws bilaterally. Stokes et al. [213] described C2 “pedicle” screws placed via the same entry point as Magerl’s screw in his series of four patients (first 1999) treated by “Harms” technique. However, the trajectory was less cranial and more medial. In fact, he placed short pars screws not reaching the VA groove. Many other authors presented good clinical and morphological results using the connection between C1 lateral mass screws and short pars screws while describing them as “pedicle” ones [41, 181].

6.3.3.1 Our Preference The short screws placed monocortically in the pars represent the least stable method of C2 fixation [47]. The only real advantage is a minimal risk of VA injury. This technique can be used very seldom if all other possibilities are not possible or fail. This can be of importance, for example, in very obese patients where achieving of required angles can be difficult.

6.3.4 Laminar C2 Screws Wright [238, 239] was the first to describe the possibility of C2 laminar screw instead of a more technically demanding and risky isthmic or pedicle fixations in atlantoaxial stabilization techniques. He performed the first surgery using crosslaminar C2 screw in 2002. Independently the same method was described later by Gorek et al. [77] who used unilateral laminar screw as a salvage technique in cases where one side C2 pars/ pedicle was not large enough to accept a 3.5 mm screw. As an alternative, for cases with deeply furrowed C2 spinous processes, Sciubba et al. [200] recently developed a technique where shorter, not crossing screws

93

6.3  Axis as an Anchoring Structure

are introduced into the laminae after removal of bifid spinous process. Laminar screws currently represent a viable alternative to the other techniques using C2 vertebra as an anchoring structure. Their stability was tested by Nassos et al. [167] on cadaveric specimens and compared to other methods of C2 fixation in occipitocervical constructs (pedicle and transarticular screws). Although similar in its ability to limit motion, laminar screws tended to be weaker in lateral bending. Their long-term stability remains questionable. Parker et al. [180] analyzed a cohort of 167 patients treated with 152 C2 laminar and 161 pedicle screws in C1-C3 fusions or subaxial long constructs. On postoperative CT scans, they documented lower frequency of cortical breach with laminar screws (1.3%) than with pedicle screws (7%) but both without any clinical sequelae. During at least 1 year follow-up, they confirmed that, especially in the short UCS constructs, the crosslaminar screws remain stable. In longer subaxial constructs they found a higher frequency (6.1%) of laminar screw revisions mostly due to pull out, hardware failure, or pseudoarthrosis. Safety and feasibility of laminar C2 screws were also described by Sciubba et  al. [200]; however, 12.5% of their patients required revision a few months after surgery because of pseudoarthrosis and screw pullout. Wang [232] also reported early hardware failure in 6.6% of treated patients. Wright [239] described 100% fusion in a 1-year follow-up in his series of 20 patients treated for various pathologies with different constructs using the C2 laminar screws. He did not have any clinical complications but postoperative CT revealed 15% of cortical breach of C2 laminae. Most authors [105, 180] prefer a free-hand technique respecting anatomical landmarks with occasional help of fluoroscopy when placing laminar screws. Although, the accuracy of laminar screws is generally very good, some authors suggest more precise techniques. Nottmeier [172] placed 4mm laminar screws in eight patients using 3D isofluoroscopy image-guided technique with excellent accuracy. Lu et  al. [147] suggested the use of a patient-specific navigational template attached to C2 spinous process. This template was created according to a plastic model of C2 vertebra prepared before the procedure with the help of CAD and 3D CT image. Recently, Dmitriev et al. [47] biomechanically tested 14 cadaveric specimens and found similar insertional torque while placing the screws into C2 pedicle or lamina but significantly lower values were registered with short monocortical isthmic screws. Interestingly, the

postfatigue pull-out strength was significantly better for pedicle than for laminar or isthmic screws. The strength of laminar screw pull-out resistance can be increased by bicortical purchase as described by Jea et al. [114]. This modification of Wright’s method is also safer because the tip of the screw is visible at the distal end of its pathway. On the other hand, it is probably not possible in all C2 arch variations.

6.3.4.1 Anatomical Background C2 lamina is considered to be the largest in the upper and middle cervical spine; nonetheless, a great variability in its diameter and length is documented by anatomical works [27, 232, 242]. The average laminar angle to sagittal plane was 44.1° (range 37°–50°) when measured on 38 dried cadaveric specimens (76 C2 laminae) by Wang [232]. The laminar length available for a screw (including the width of spinous process) was, on an average, 31.6  mm (range 27.0–37.0  mm). The same author measured the cross-sectional diameter of the C2 lamina. He found that, in its thinnest portion, the average height of the lamina was 11.5 mm (range 9.0– 14.1  mm) and average thickness was 6.3  mm (range 3.6–9.1 mm). He thus found that 21% of laminae could not accommodate a 3.5 mm screw (less than 5.5 mm space available) if allowing for an extra circumferential millimeter 42% of tested sides cannot accommodate a 4 mm screw (less than 6 mm space). Cassinelli et al. [27] observed a similar C2 lamina thickness on 420 dried C2 vertebrae similar but found a gender difference with larger values obtained from male specimens. They concluded that 70.5% of laminae had thickness equal to or greater than 5 mm, 92.6% more than 4 mm, and 96.7% more than 3.5 mm. The average length of arches was lower than in previous series. Only 45.2% of laminae were able to accept 25 mm screws and only 1.9% could accommodate 30 mm screws without entering the lateral mass (internal distance measured). As others [203, 242], they found that caudal part of lamina is thicker than cranial one and recommended to aim the screw to lower laminar margin.

6.3.4.2 Surgical Technique According to Wright’s recommendation [239], after a subperiostal exposure of C2 spinous process and

94

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

laminae, a small cortical window is created on the right side of rostral part of C2 spinous process at its junction with the right lamina. The hand drill is used to drill the pilot hole along the visible left-sided C2 lamina. The trajectory should be kept slightly away from the spinal canal than the downslope of the lamina. A ball probe confirms the integrity of the intralaminar canal. A 30 × 3.5 mm screw was placed and no tap mentioned. The opposite screw is introduced in the same manner but the entry point is located more caudally in order to allow the screw intralaminar crossing. According to the author, no fluoroscopy or navigation is necessary to introduce 3.5 mm crosslaminar screws but a preoperative CT evaluation of laminae and their size is mandatory. To avoid canal content damage, the internal surface of C2 lamina can be palpated in cases of angle or thickness uncertainty. Gorek [77] advised to tap the initial part of screw hole located in the middle of the junction of spinous process and lamina for unilateral screw purchase. The appropriate screw length with endpoint located no further than the junction between pars and lamina was determined by depth gauge. Jea et al. [114] prefer the penetration of the distal cortex with a screw tip by modification of the introductory angle.

6.3.4.3 Our Preference Influenced by the mentioned anatomical works and our own experience, we argue that prior to any C2 laminar screw placement, a thin-sliced CT should be obtained to determine the thickness of C2 laminae. It also allows establishing the ratio between cortex and cancellous intralaminar bone canal and the appropriate screw length in order to avoid penetration of C2/3 joint (Fig. 6.27). In our opinion, this technique is a useful salvage procedure (Figs. 6.28 and 6.51). Laminar C2 screw technique can be helpful in cases of small pedicles or pars or erroneous attempt of pedicle or isthmus cannulation resulting in weak bone or need for a different (now risky) trajectory. It seems obvious that a correct transpedicular or transisthmic screw technique with bicortical purchase should provide stronger screw anchorage than a screw placed in the cancellous bone cavity of C2 lamina. A further potential and so far unknown risk is that of a ventral laminar breach of the screw during daily life overload. Fortunately, such complication with clinical sequelae has not yet been described. An early

postoperative CT confirms proper screw position and is highly recommended as plain radiographs will not show any ventral cortical penetration [140]. When planning various types of fixation with axis serving as anchoring structure, the preoperative CT analysis should always include a study of possible “salvage” options. Routine production of patient-specific templates will, probably, not be available to every surgeon but the use of 3D isofloroscopes connected to navigational systems seems to be the feature allowing not only a precise screw placement but also an immediate check of their real-time position.

6.3.5 Odontoid Process Screw This specific screw is used only for the treatment of odontoid process fractures and not as a part of any segmental fixating construct. It was first used by Nakanishi et al. from Japan in August 1978 [166] and independently Friedrich Magerl performed the same procedure in January 1979 [87] in Switzerland for odontoid pseudoarthrosis. He also suggested using this technique for the treatment of fractures. First documented series of 12 fractures treated this way was published by Böhler [20] with annotation that the idea came from Magerl. The pioneers described the necessity of use of two screws where the longer one is acting as a lag screw and the shorter securing the rotational stability. Numerous papers then described the success of this method in the treatment of type II and shallow type III odontoid fractures [2, 68, 128, 182, 220]. Later, especially North American authors, under influence of biomechanical works [79, 198] began to advocate single screw fixation of odontoid fractures [117, 216]. Overall fusion rate achieved by direct odontoid screw was between 85% and 95%.

6.3.5.1 Anatomical Background The approach trajectory is oblique to allow the screw passage from anteroinferior edge of C2 vertebral body to the tip of the odontoid process. There is approximately 14 mm of available space for a safe screw purchase in the midline of the base of C2 [131]. The distance to the base of the odontoid is approximately 20 mm and the process itself is another 20 mm long, on

95

6.3  Axis as an Anchoring Structure

Fig. 6.27  Modeling of SAS for C2 intralaminar screw placement on navigational workstation documenting large SAS

an average. Thus, one can calculate that the length of the lag screw should theoretically be between 36 and 44  mm. The odontoid process diameter is approximately 10 mm at its base, just enough space for two 4 mm screws under normal conditions. Certainly, there is large gender, race, and age variability, which has to be taken into account and patient-specific individual parameters have to be measured exactly (best on CT).

The odontoid tip is covered by very dense cortical bone, which has to be penetrated with a longer screw.

6.3.5.2 Surgical Technique Odontoid fractures can be treated with a direct screw compressive osteosynthesis only if reduced

96

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

Fig. 6.28  Crosslaminar screw used as salvage procedure on the side where the transpedicular and transisthmic screw introduction was not feasible. (a) Left high riding VA visible on coronal CT reconstruction. (b) Left parasagittal reconstruction showing unfeasibility of transisthmic screw. (c, d) Postoperative axial CT scan showing left C2 crosslaminar screw and Magerl’s screw on the right side

preoperatively. Therefore, patient positioning on the table allows for oblique introductory angle without dislocation of the fracture and is, indeed, a critical step in the surgical procedure. The screw angle can be a limiting factor in obese people, patients with fixed cervical kyphosis, and those with a barrel chest. The obliquity of this approach usually starting at the level of C5 and steep cranial angle allows only for a fairly limited illumination. This fact led to further development of various self-retaining and tubular retractors [96, 204]. Also, the screws come in a variety of shapes and materials: full-threaded or partially threaded; double-threaded; 3.5 or 4 mm diameter; steel, titanium alloy or PLDLA, and full or cannulated [2, 6, 9, 129]. Initially, the fullthreaded 3.5 mm stainless steel screws were used and the proximal part of the pilot hole for the lag screw was over-drilled with larger diameter drill to achieve compression [20]. Later, cannulated screws guided by a Kirschner wire were used with a short apical thread and long smooth shaft [2]. The fixation itself started with drilling followed by tapping and screw placement performed under real-time double fluoroscopy.

6.3.5.3 Our Preference There is no doubt, in our opinion, that direct compressive osteosynthesis of type II and shallow type III odontoid fractures is the most physiological method available with highest fusion rates if correctly indicated and performed. Main advantage of the direct osteosynthesis is that no other motion segment is involved in the fusion. It can be performed only if the fracture is reduced and the introductory angle is possible.

6.3.5.4 Our Surgical Technique In our experience it often took more time to position the patient on transparent table correctly, than the surgery itself. The patient’s head is fixed in halo ring (not necessary in every case) to enable free manipulation with continuous traction (Fig.  6.29). The patient is intubated without changing of the spine position (under fluoroscopical control, fiberoptically, or awake), and his/her mouth is held open with mouth distractor or simple roll

6.3  Axis as an Anchoring Structure

97

Fig. 6.29  Setup of two rectangular fluoroscopes. (a) Position of C-arms and the transparent surgical table. (b) Peroperative haloring traction

of gauze. It is very difficult and very important to set the C2 vertebra in sagittal plane correctly to achieve ­acceptable angle for screw placement and not to dislocate the fracture especially in posteriorly dislocated or oblique fractures. Positioning maneuvers and the introductory angle are checked on lateral fluoroscopy using an extended radiopaque marker (e.g., long K-wire) (Fig. 6.30). Following that, a second fluoroscopic machine is set up perpendicular to the first one to allow for a transoral view. It is usually adjusted in such a way as to visualize the tip and shaft of odontoid process and the base of C2 simultaneously on both screens (Fig.  6.31). The screens are always positioned in front of the operating surgeon and the pictures turned to depict the same orientation as the patient position (Fig. 6.32). We never start the procedure if the previously described harmony of position and radiographic visibility is not achieved. A right-sided horizontal incision is made just above the C4-5 interspace and the anterior spine is exposed through a standard, bloodless anterolateral approach.

Fig.  6.30  Preoperative fluoroscopical testing of correct angle achievability with metal probe

The prevertebral fascia is sharply cut and oblique tunnel reaching the base of C2 is created by blunt dissection. The extent of exposure can be checked by lateral

98

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

Fig. 6.31  Necessary preoperative fluoroscopical visibility. (a) Odontoid process shank and tip. (b) Unobstructed lateral C2 view

Fig.  6.32  Surgeons position with direct view on the fluoroscreens

fluoroscopy. We do  not damage the C2-3 intervertebral disk. Because of lack of illumination in the oblique approach tunnel, we use the headlight and the space is held open either with the Apfelbaum retractor or simply by the assisting surgeon using a Hohmann hook fixed to C2 body (Fig. 6.33). We do not have experience with tubular retractors that may have the advantage of being fixed to the table and having tubular lighting. Thin and sharp probe or K-wire is positioned on the expected midline of the inferior C2 ridge and its correct position is checked on both fluoroscopes. A long, straight probe can also predict the final sagittal screw angle needed. If a double-screw construct is planned (more common), two entry point holes are created approximately 4–5 mm from the midline at the C2 inferior edge with a high-speed drill (3 mm burr). If a single screw

fixation is planned, the entry point is created in the midline. We always attempt to first drill the pilot hole for the lag screw. This strategy allows for a potential second chance in case of an initially wrong trajectory that is difficult to fix by a simple change of the drill angle. The drill is covered by a drill guide with a depth gauge; however the depth measurement is usually fairly inaccurate as it is difficult to obtain a good contact of the sleeve with the oblique surface of the vertebra. Drilling is performed slowly step-by-step under continuous fluoroscopy but with high drill speeds as this provides better tactile feedback of resistance. The drill is stopped once the condense bone of the apex is completely perforated. According to anatomical works, there is, on an average, a 6 mm safety zone between the apex and the brainstem. However, maximum attention must be paid to this part of the drilling (i.e., fixed elbows to prevent unintentional “plunge”). While drilling the first hole, we already plan the position of the second screw as sufficient space needs to be maintained. The screw should not be located too far anteriorly within the C2 body as this may predispose the screw to anterior breakout. The hole is then tapped including the dens apex using the same drill guide (Fig. 6.34). We use the Apfelbaum titanium, short-threaded 4  mm screws attached to self-retaining screw driver for odontoid fixation. Thus, no overdrilling of the proximal screw hole is necessary. The screw is forcefully tightened to perforate the apex approximately by one thread. In our opinion, this is another essential part of the procedure. If the screw does not perforate the apex it can distract the fracture rather than compress it. It is possible to note any rotational instability during

6.3  Axis as an Anchoring Structure

99

Fig. 6.33  Fluoroscopical view of two possibilities of oblique surgical canal spreading. (a) With Apfelbaum’s distractor. (b) With pediatric Hohmann retractor held by assistant

Fig. 6.34  Tapping of the pilot hole including the odontoid process tip

Fig. 6.35  AP radiogram of double odontoid screw purchase

tapping or screw tightening as rotation of the odontoid becomes apparent on fluoroscopic images. In such a situation, we drill a second screw hole and the lag screw is tightened with the drill left in place. Under normal conditions, the second screw is placed in the same manner with the aim to simply cross the fracture line sufficiently

(Fig. 6.35) to fulfill the anti-rotational purpose. Finally, the wound is closed in a standard fashion. Anteriorly oblique fractures are usually considered a contraindication to the odontoid screw fixation due to the dislocation risk during screw tightening. In our ­opinion, this fracture type can also be treated by this

100

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

Currently, the use of a single versus two odontoid screws remains controversial. In our opinion, only those who are technically able to place two screws can discuss the possibility of a single screw fixation. If the fracture surface is rough (irregular) and the fracture can be well compressed, then a single screw fixation can be sufficiently strong (Fig. 6.37). We feel that the crucial point is the perforation of apical cortex to allow strong compression. Otherwise, distraction of the fracture and rotational dislocation of the dens will prevent bony healing and the screw will subsequently break (Fig. 6.38). Another possible drawback of direct odontoid screw fixation is poor bone quality in some patients preventing adequate screw anchorage.

Fig. 6.36  AP fluoroscopical view of parallel K-wire introduced to allow lag screw tightening without odontoid rotation

technique provided enough bone is available below the fracture line in the C2 body. This is the fracture that needs to be reduced preoperatively with positioning, as emphasized earlier. First, the Kirschner wire (1.5 mm) is passed to the tip of the dens without cortical perforation. Then, a pilot hole for lag screw is drilled penetrating the odontoid apex with subsequent tapping. A short-threaded screw is placed and forcefully tightened to achieve compression of the fracture while the Kirschner wire is holding the odontoid in place not allowing anterior dislocation (similarly to rotational instability) (Fig.  6.36). Finally, the wire is replaced by a short antirotational screw. We do not use cannulated screws. The main reason is that we consider apical penetration and compression of fracture very important and therefore, if a K-wire was to be in place during tapping or screw tightening, it could be inadvertently advanced into the canal with catastrophic consequences. a

Fig. 6.37  Compression needed for fracture alignment and healing. (a) Double screw purchase. (b) Single screw compression

6.3.6 Screw Introduced into C2 Body Either monocortically or bicortically introduced screws into the C2 body are used almost exclusively to fix the plate stabilizing the axis to C3 and/or lower cervical vertebras. High anterolateral cervical approach is derived from the subaxial access described to reach C3-T1 spine as described in detail in Chap. 4.

6.3.6.1 Our Preference We use this type of screw purchase only to fix the plate stabilizing subaxial cervical segments. We prefer to introduce the screws bicortically in trauma cases especially in hangman type fractures (Figs.  12.14d and 12.18, Chap. 12) or combined injuries (Figs. 14.2 and 14.3, Chap. 14). The monocortical purchase can be chosen in traumatic and degenerative disk prolapses without marked segmental instability. From technical b

101

6.4  Monosegmental Fusion Constructs

6.4.1.1 Posterior C0-1 Fixation Methods

Fig. 6.38  Tomogram of improperly tightened screw 9 months after surgery. Note distraction – pseudoarthrosis

point of view, compared to standard anterolateral ACDF, the only difficulty represents the distant upper C2 screw purchase. To enable the C2 anterior surface perpendicular drilling trajectory quite important midline dislocation is necessary. In such a case, we remove all the automatic wound retractors. This wound release allows for more easy medial pharyngeal and laryngeal dislocation. Usually, the drilling and screw purchase are performed with manual wound distraction and with instruments covered by a protection sleeve.

Grob [81] speculated that the ideal fixation should only fix the target segment. Therefore, atlanto-occipital instability or dislocation should only be stabilized through the CVJ if possible. Currently, there are two main options for posterior stabilization of atlanto-occipital joint. As described above, Grob [81] suggested performing a C1-0 transarticular fixation similar to Magerl procedure for C1-2 (Fig. 6.3). Because of biomechanical weakness in flexion, Gonzalez et al. [75] proposed the use of a bone graft between occiput and the atlas, similar to Gallie type fusion. The same group successfully performed three such posterior stabilizations [60]. Maughan et al. (Fig. 6.39) described another option for atlanto-occipital fixation in a patient with circular avulsion fracture of foramen magnum. It involved occipital plate and C1 lateral mass screws connected by a rod [156]. Bambakidis et al. [16] compared the two previously mentioned techniques with occipitoatlantal wire and rod fixation, biomechanically. They found that both screw techniques are substantially stiffer than the wirerod method. Nevertheless, they also confirmed that the transarticular screw supplemented with a buttress graft is slightly more rigid than the plate-screw-rod system.

6.4.1.2 Our Preference For mild types of AOD, isolated condyle fracture dislocations, and chronic atlanto-occipital instabilities, the short monosegmental posterior fixation is sufficient.

6.4 Monosegmental Fusion Constructs 6.4.1 Posterior Monosegmental Fusion Constructs Posterior monosegmental fusion constructs are usually strong enough to maintain stability without the need for external bracing and allow for exclusion of mobile segments not involved in the pathological movement.

Fig. 6.39  Artistic drawing of atlantooccipital fixation according to Maughan

102

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

However, for typical AOD caused by high impact trauma disconnecting the ligamentous attachments of the head to the spine, the short construct is insufficient. Such patients, if they survive, are often bedridden and ventilator dependent needing very frequent positioning and passive manipulation. Instability can have catastrophic consequences. Taking patient’s prognosis into account, a construct extended to C2 or more caudally may have negligible influence on further quality of life.

6.4.1.3 Posterior C1-2 Fixation Methods Mixter and Osgood Silk Loop In 1910, Mixter and Osgood [161] were the first to describe surgical treatment of atlantoaxial instability in a 15-year-old boy who fell from a tree. He came to be operated in Massachusetts General Hospital, 6 months after unsuccessful conservative treatment of unrecognized odontoid fracture with C1/2 dislocation. He had pain but no neurological deficit. He was first manipulated under anesthesia. Then new radiographs, including transoral projections, revealed the old odontoid fracture. A leather external orthosis was manufactured. Surgery was performed by Dr. Mixter from posterior midline approach. Benzoin-soaked, braided silk loop was passed below the C1 arch and around spinous C2 process and no bone graft was added. The patient survived and fused without complications. In this report, we can recognize some interesting facts. They performed “open mouth” films to visualize the UCS. They reviewed the available literature and concluded that in the UCS trauma cohort, the most common abnormality was atlantoaxial dislocation without fracture, followed by odontoid fractures and then fractures of C1 and C2 arches, and lateral masses.

imperative. Posterior surgical treatment was indicated whenever displaced facets were irreducible by traction, or failed to maintain reduction in external orthosis or in cases of non-reduced malunion. He stated that “recurrence of displacement can be prevented by fastening the two vertebrae together by fine steel wire passed around the laminae or spines. And the risk of late recurrence can be eliminated by bone grafts laid in the spines or on the laminae and articular facets.” So, no typical H shape graft was mentioned neither depicted on the pictures in the original paper. Other authors also describe the technique of posterior graft and wiring, even 2 years earlier [33]. Fried used the “Gallie method” of C1-2 wiring and grafting after scraping the atlantoaxial joints with failure rate of 80% [66]. The first paper describing an H-shaped onlay notched graft placed on the surface of C1 and spinous process of C2 is that of McGraw and Rusch [158]. It can be summarized that the original technique of onlay grafting is the simplest fusion method but also the least stable when compared to other techniques, especially in rotation [82]. Today, the notched onlay graft positioned on the surface of C1 and around the spinous process of C2 is called “Gallie – type graft” (Fig. 6.40). Failure (pseudoarthrosis, wire breakage, or loosening) of Gallie type fusion can be seen in up to 80% of cases if used as standalone method [66]. Fusion rate can be improved with halo bracing, but even then 25% of cases still fail [36].

Brooks and Jenkins – Wire and Graft In order to increase the stability of posterior wire and graft fusion, Brooks and Jenkins suggested interposing

Atlantoaxial Wire and Graft The addition of an H-shaped bone graft currently used as supplement to other more solid metal constructs, is usually attributed to W.E. Gallie, surgeon from Toronto. He published his algorithm in 1939, describing how to treat fractures and dislocations in cervical spine [67]. He suggested to always begin with skeletal head traction in cases of subluxation anywhere in the cervical spine, emphasizing that reduction of dislocation is

Fig. 6.40  Gallie’s type of posterior AA fixation

103

6.4  Monosegmental Fusion Constructs

Fig. 6.41  Brooks and Jenkins type of grafts and wire posterior AA fixation

two grafts laterally between the C1 and C2 laminae by wedge compression technique [24]. Beveled iliac crest autografts were fixed in place with doubled 20 gauge stainless steel sublaminar wires (Fig.  6.41). It is the need for bilateral sublaminar cable passage that has a higher potential rate of neurological or dural injury. They successfully treated 15 patients with final fusion rate of 93%. They supplemented their surgical procedure by postoperative use of either a Minerva or SOMI brace. Later works evaluating larger series using Brooks method for C1-2 fixation reported failure rate up to 30% [36, 80].

Sonntag – Wire and Graft Sonntag et al. modified the Gallie graft fusion technique in the early 1990s in an attempt to improve stability of the construct and avoid bilateral C2 sublaminar cables [46]. They decorticated the contact surfaces of C1 and C2 arches and interposed curvilinear strut graft approximately 4 cm long with caudal notch for the C2 spinous process. The graft was then fixed by a cable passing under the posterior arch of C1 and looped around a notched inferior C2 spinous process (Fig.  6.42). However, in patients treated with posterior wiring techniques only, they treated them with three months of postoperative halo immobilization. This approach achieved 97% fusion rate in their series of 36 patients.

Acrylic C1-2 Fusions Poly-methyl-methacrylate (PMMA) onlay for atlantoaxial fixation was advocated as a fast option for surgical immobilization in patients with traumatic atlantoaxial

Fig. 6.42  Sonntag’s modification of AA graft and wire fixation

instability. Kelly et al. [124] successfully treated seven patients suffering from traumatic AA instability with pure C1-3 acrylic and wire fusion. Authors reported “fusion” in all cases in “satisfactory” position during an 8–9 year follow-up. Good, long-term results were also reported in other series [50, 206]. The experienced authors recommended using screw anchors imbedded in the acrylate to increase the long-term stability [50]. However, the main objections of polymethacrylate fixation method are the heat produced during polymerization process and the inability of PMMA to bond to bone. Further, a large number of revisions due to infection, inlay loosening, or inability to maintain spinal alignment led to substantial decrease in popularity of PMMA fixations [157]. In our opinion, PMMA can still be used as a palliative measure in UCS tumor surgery and/or if used in other indications, it has to be supplemented with additional bone grafting.

Halifax Atlantoaxial Interlaminar Clamps After a very successful application of interlaminar clamps in the treatment of subaxial cervical spine trauma [101, 227], this method was applied in the treatment of atlantoaxial instability as well. Cybulski et al. [39] confirmed the safety of interlaminar clamps with one out of eight clamps loosening prior to fusion. Clamps with claw-type construct avoid the need for sublaminar cable and its potential risks but the risk of neural injury is not completely depleted (Fig.  6.43). The exact fit of the clamps can be of concern under certain anatomical conditions, so can be the loosening of compressing screws. Previously described problems were the main reason for failure and revision of up to 20% of cases, especially in the absence of bony graft

104

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

Fig. 6.43  Halifax clamp with interlaminar grafts

[4, 163, 210]. If iliac crest autograft is interposed and the halo vest applied then fusion rate improves ­dramatically [109]. Historically, other claw type constructs were also used. In summary, methods using posterior C1 and C2 arches to fix the segment do have some advantages but also some important disadvantages: Advantages – Simple to apply – Valuable addict to other more firm fusion methods – Can be a salvage procedure (AV injury during screw techniques) Disadvantages – Posterior elements must be intact – Not possible in low bone quality of arches (osteopenic status) – Cable has to pass bellow the arch of C1 (dangerous in dislocation) – High failure rate – Necessity for hard external support – Autologous graft-related complications

Our Preference Posterior wiring techniques have long been the mainstay of surgical stabilization of atlantoaxial complex. They are relatively easy but of limited stiffness. In particular, they cannot reach the same stability in translation and rotation when compared to other screw constructs. The “parallelogram effect” was well described by Panjabi [176]. The other concern is that they provide less than optimal fusion rates and external rigid immobilization is mandatory. Graft breakage or

wire loosening are relatively common complications of wiring techniques [251]. If there is a need for standalone posterior wiring technique, then the Sonntag modification should be used supplemented by external hard brace (SOMI) or a halo vest. It is generally accepted that solid bony fusion, when desired, is most reliably achieved when the segmental fixation minimizes motion. This is why the previously mentioned techniques were not the last point in C1-2 fixation development and more rigid screw constructs followed.

Transarticular C2-1 Screw Fixation (Magerl) Magerl’s technique of C1-2 fixation (Fig. 6.44), introduced in 1987 [152], gradually achieved a high degree of acceptance and success, mainly because of high fusion rates, instant stability, and relatively low incidence of complications. It has gained popularity over the wiring techniques used earlier for posterior atlantoaxial stabilization especially because of higher proven biomechanical stability avoiding the necessity of postoperative hard external support [82, 164, 188, 237]. High fusion rate (between 87 and 100%) was documented by many authors [36, 44, 69, 86, 91, 116, 154, 212, 214, 219]. As described earlier, a meticulous technique must be used to achieve a safe C2 transisthmic passage and adequate C1 lateral mass anchorage. Thorough preoperative anatomical analysis with 3D modeling in potentially risky cases is invaluable in identifying patients/sides unsuitable for the transisthmic screw [234]. As the risk of VA injury can be as high as 23% per patient in borderline cases, navigational systems can prove to be a useful adjunct [99, 100, 234]. Strong C1 anchorage is important, therefore the screw should be placed at least 5 mm within the bone and should not protrude more than 5 mm. Grob et al. [86] evaluated 161 patients in a multicentre Swiss study and found a 15% rate of less than perfectly positioned screws with 3.4% of them missing the C1 lateral mass. In our cohort of 80 patients collected from 4 centers, 150 screws were placed [219]. The morphology on one side of six patients prevented adequate screw placement and two procedures needed to be converted to posterior wire fusion. VA injury was encountered in four patients (5%) without clinical consequences. We found 28.6% of screws to be suboptimally placed. They were adequately imbedded in the

105

6.4  Monosegmental Fusion Constructs Fig. 6.44  Transarticular C2-1 fixation according to Magerl. (a, b) Posterior and lateral view. (c) Transarticular AA fixation supplemented with posterior graft and wire fusion

C1 lateral mass but in 6% they were too short, in 5.3% too long, and in the other 17.3% deviated out of middle third of the mass (Figs. 6.45–6.47). Four screws (2.7%) were misplaced (i.e., out of C1 lateral mass). Computerized navigation can substantially improve accuracy. Foley et al. [65] were the first who used stereotactic guidance to place atlantoaxial transarticular screws with improved safety. Weidner et  al. [234]

placed 72 screws with image guidance and significantly decreased the frequency of suboptimally placed screws. On postoperative CT, they found lateral deviation only on three sides and medial, on only one. The anchorage within C1 lateral mass was always sufficient. Several authors reported a successful and safe placement of transarticular C1-2 screws in pediatric population as young as 3 years of age [23, 97].

Fig. 6.45  Correct transarticular screw position and length. (a) Transoral radiograph. (b) Lateral view

106

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

Biomechanical superiority over posterior graft fixation techniques is given by more central position of the screws and was, indeed, confirmed in many biomechanical studies [82, 93, 132, 164, 188, 237]. The stability of the construct in flexion-extension is certainly improved by and comparable to Goel-Harms procedure if posterior wire-fixed graft is added [202]. On the

other hand, Naderi et  al. [164] found that unilateral screw fixation is much less stable than bilateral one in all directions, especially in rotation. Sven Olerud suggested a modification of Magerl’s method to decrease the frequency of graft-related problems but mostly to increase the stability of the construct [174]. He connected transarticular screws to an

Fig. 6.46  Postoperative evaluation of transarticular screw position and length from plain X-rays (suboptimal means sufficient anchorage of C1 lateral mass). (a) Suboptimal purchase, too long screws. (b) Too short screws, suboptimal introduction.

(c)  Suboptimal position, too medially located left screw. (d)  Suboptimal purchase, too laterally located left screw. (e) Wrong trajectory, the screw is outside of C1 lateral mass

107

6.4  Monosegmental Fusion Constructs

Advantages of Magerl’s technique – Immediate and strong stability, no external support needed – Laminae could be absent, no wiring necessary – Longer construct incorporation is feasible if polyaxial screw heads are used – High fusion rate – If bone block-wire fusion added, biomechanically superior – Cost (it is significantly cheaper than other techniques) – Low profile of hardware Disadvantages of Magerl’s technique

Fig. 6.46  (continued)

adaptable claw attached to the posterior arch of C1 with a 3.5 mm rod. This innovative technique allows for use of morselized graft only because a structural block of bone was not necessary to create a posterior band. Another theoretical advantage is that the claw can grasp even a partially defective posterior C1 arch in situations where cable fixation would not be possible. The Olerud modification was later successfully used in 26 patients with 91% fusion rate treated for AA instability caused by trauma, RA, and developmental anomalies [35]. In our experience, there are advantages and disadvantages of the described technique:

– VA injury risk in up to 23% of patients – The angle of C2 transpedicular trajectory is not easy to achieve – Reduction feasibility of atlantoaxial dislocation possible, but limited – Cannot be used if target structures are comminuted or destructed – C1-2 joint is damaged – Learning curve necessary – Fluoroscopy mandatory

Our Preference Magerl’s fixation is often described as unsafe but it is not dangerous because of the technique itself. It is the surgeon who can make the procedure unsafe, usually as a result of inadequate preoperative radiological analysis

Fig. 6.47  CT verification of correct transarticular screw positions. (a) Oblique axial reconstruction. (b, c) Parasagittal reconstruction of both sides of pars interarticularis screw position

108

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

and deviating from the recommended trajectory through the C2 isthmus. Incorrect screw angle can often be a result of inappropriate patient positioning, a short neck, spine deformity, or hyperlordotic spine curve. Further, atlas settling, absent C1 anterior tubercle after transoral decompression or VA variability (in approximately 20% of patients) can lead to VA injury. On the other hand, if atlantoaxial transarticular fixation is correctly indicated and performed, it represents a very stable fixation avoiding the need for hard external bracing and providing high fusion rates. Certainly, in some countries, the lower cost of two screws when compared to other methods may be an important factor.

 1 Lateral Mass – C2 Pedicle Screw C and Rod Fixation (Goel, Harms) Goel and Laheri performed the first C1-2 fusion in 1988 [72]. To achieve a wide exposure of the C1-2 joint they amputated the C2 nerve root in all cases. Both the C1 lateral mass and C2 pedicle screw were placed monocortically over a steel plate regularly used for finger bone osteosynthesis (economic reasons – personal communication) (Fig. 6.48). In some of their patients a longer plate was used to allow for occipitocervical fusion. Later, they analyzed a series of 160 patients treated for various C1-2 instabilities by their method [70]. They accepted a bicortical screw purchase as a more stable option and described some of the pitfalls of their method. In few cases, screw placement was impossible due to the encountered morphology. Eighteen patients described specific sensory loss in C2 root area. In four patients, they have seen profound arterial bleeding while drilling the pilot hole for C2 pedicle screw. Although it was not confirmed angiographically as

Fig. 6.48  Screw and plate AA fixation according to Goel

there were no neurological sequelae, they concluded that the bleeding was related to VA injury. They considered all constructs stable after 5 months of follow-up and only one screw was broken after 18 months. Postoperative CT was not routinely performed. The same method, supplemented by joint distraction and placement of a hydroxyapatite or titanium spacer was used in fixed atlantoaxial dislocation in 19 patients [71] to treat atlas settling or dislocation. Most of the patients (18) were suffering from fixed dislocated odontoid pseudoarthrosis and/or os odontoideum. Construct stability was not biomechanically tested and the authors had their patients wear a hard cervical collar for 3 months. Harms and Melcher, who advocated bicortical screw purchase, improved the technology of previously described fixation [94]. They developed polyaxial screws allowing top loaded rod connection (Fig.  6.49). The longer C1 screws with a smooth proximal shaft allowed sparing of the C2 nerve root and indirect manipulation of the atlas (in case of fracture distraction) and consequently in connection to C2 pedicle screws to manipulate the atlantoaxial joint. Their construct can easily be extended either cranially if occipitocervical fusion is required or caudally to fix the UCS to subaxial cervical spine. Bone graft can be packed either directly into the atlantoaxial joint or as a posterior onlay graft. The above-described techniques enable significantly more comfortable angles of screw placement; however, the risk of VA injury is not eliminated. Increasing number of publications documented successful use of Goel’s technique. Stulik et  al. [215] performed atlantoaxial fusion with Harms instrumentation in 46 patients. They used the system either for temporary or permanent fixation. Twenty four patients were evaluated retrospectively with at least a 12-month follow-up. Three C2 screws (5.4%) were considered malpositioned on the postoperative CT scan, with one breaching the canal cortex and two encroaching the FT. They did not have any clinical complications and documented 100% fusion rate. Arayan et al. [14] retrospectively evaluated 102 patients treated with Harms fixation/fusion. In this probably largest available series, most of the patients (48 patients) were treated for instability caused by odontoid pseudoarthrosis. The authors used navigational system in first third of treated patients and neuromonitoring in all of them. They always cut the C2 nerve root with only one case of neuropathic pain in follow-up. They distracted the atlantoaxial joint when

109

6.4  Monosegmental Fusion Constructs

Fig. 6.49  (a) posterior (b) lateral view

needed using force transmitted through screws and in 38% of procedures introduced bilateral allograft spacer intra-articularly. In 23% of patients, at least one side was anatomically not suitable for C2 pedicle screw. Therefore, pars screws were placed (probably, the short ones) instead. The risk of VA injury in this study is similar to that of Magerl. Two VA injuries were encountered but happened during subperiostal exposure of C1 due to atypical VA loop. The strength of the construct was increased by regular use of crosslink. The followup revealed 98% fusion rate. The Goel-Harms technique was also successfully used in a limited number of pediatric patients (6 patients, the youngest being 7 years old). Similar restrictions applied to children as they do for adults [98]. The biomechanical stability of Harms technique was tested in multiple studies and it is similar to transarticular screw fixation if supplemented by posterior wiring [106, 132, 159]. Again, there are advantages as well as disadvantages of the described technique: Advantages of Goel-Harms technique – Immediate stability, no external support needed – The angle of C2 pedicle screw trajectory is easy to achieve – The screw placement is not dependent on relative C1-2 position – Reduction of atlantoaxial dislocation feasible during procedure – C1-2 joint is not damaged, temporary fixation is possible – Isolated C1 fracture compression practicable – Laminae could be absent, no wiring necessary

– Prolonging of construct easy – High fusion rate Disadvantages of Goel-Harms technique – VA injury risk underestimated and possibly as high as 23% of patients – Cannot be used if target structures are comminuted or destructed – Steep learning curve – Cost (Harms fixator is much more expensive than other techniques) – Fluoroscopy mandatory – High profile of the polyaxial construct Our Preference The Goel-Harms technique is a valuable tool in all types of C1-2 instability providing excellent immediate immobilization. The specific advantages in comparison to Magerl’s technique are of two kinds. First, the angle necessary for C2 pedicle screw placement is much easier to achieve and secondly, in cases of dislocated or fractured atlas, the vertebra can be manipulated independently. We would like to emphasize, however, that there is no evidence that this method has a lower risk of VA injury.  1 Lateral Mass – C2 Crosslaminar C Screw and Rod Fixation (Wright) Due to limited stability of C1-2 wiring fixations more rigid methods using C2 isthmic and/or pedicle screws were developed. All those techniques required for a

110

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

screw to pass through an anatomically variable area neighboring the C2 VA groove and thus posing a substantial risk to the artery. In order to avoid this technically demanding procedure, Wright [238] suggested to place two crossing laminar screws and connect them to C1 lateral mass screws with rods (Fig. 6.50). The acute stability of C1 lateral mass – C2 laminar screw fixation was biomechanically tested and compared to Harms construct and combination of laminar/ pedicle construct by Gorek et al. [77]. No statistically significant differences in stiffness were found and, at least in cadaver models, the laminar screw construct is comparable to other atlantoaxial fixation techniques. However, other authors demonstrated less rigidity with laminar screw constructs when compared to pedicle screw anchor, most pronounced during lateral bending and rotation [32, 135]. Similar results were confirmed by Dmitriev who studied construct stiffness after an experimental odontoid process transection [47]. The position of laminar polyaxial screw heads does not allow the use of sublaminar wire in C1-2 fixations. The eventual bone graft has to be fashioned to stay below the rods. The other concern, in comparison to Harms technique, is the possibility to manipulate C1 on C2 with the help of distraction/compression forceps when the screw heads are near to the spinous process of C2. Although never biomechanically proven, the proximity of the laminar screw heads to the medial axis and the curvilinear rod needed to connect them can potentially be less resistant to lateral translation, bending, and rotational forces. This can be of increased importance if unexpected polyaxial screw head would not retain initial stability. Advantages – No risk of VA injury – Less technically demanding

Fig. 6.50  Artist’s drawing of C1 lateral mass – C2 crosslaminar AA fixation according to Wright

– Can be used if C2 pedicle or pars is not large enough to accept a 3.5 mm screw – Can be a salvage procedure in cases of C2 pedicle/ pars erroneous placement – Decreased perioperative radiation exposure – Direct visualization of the target structure (i.e., C2 laminae) – Retained possibility to reduce the C1-C2 dislocation – Independence on C1 position – Easy angle of screw trajectory – Immediate C1-2 construct stability compared to other rigid techniques – Good long-term stability of translaminar screws in short C1-3 constructs Disadvantages – Cannot be used if C2 laminae are absent – If the core diameter of lamina is less than 3.5 mm, cortical breach can occur – Potential risk of spinal cord injury – Fixation of Gallie’s type graft can be difficult – Not enough space to connect C2 to C3 lateral mass screw – Offset connectors or significant rod contouring necessary – Problematic long-term stability in long subaxial constructs

Our Preference In our department, we use Wright’s technique with increasing frequency, mostly as a salvage procedure in situations where a high-riding VA prevents a pedicle or transarticular screw placement. This is usually only on one side in the construct. So far only once, has this situation occurred bilaterally (Fig. 6.51).

6.4  Monosegmental Fusion Constructs

111

Fig.  6.51  Bilateral high riding VA in patient with odontoid pseudoarthrosis not allowing any C2 postero-anterior (transisthmic, transpedicular) screw introduction treated by crosslaminar screw C2 – massa lateralis C1 fixation acc. to Wright. (a) Frontal

plane reconstruction. (b) Postoperative axial scan in the C2 arch plane. (c) Plain postoperative laterogram. (d) CT in 3D posterior crosslaminar C2 screw purchase and C1 lateral mass screw fixation supplemented with autologous graft

Intralaminar Screws C1 – Short Pars C2 (Donnellan)

nor lateral mass fixation or wiring of the atlas. Donnellan et al. [48] suggested to use a combination of intralaminar C1 screws connected to short C2 pars screws (Fig. 6.52). They have documented good results and 100% fusion rate in three patients treated with this technique. The method seems safe, avoiding all the known risks of arterial injury but has not been biomechanically tested so far.

In rare cases, the posterior C1 arch can be defective and the lateral mass eroded by inflammatory or degenerative process together with concomitant C2 high riding VA. This situation does not allow either placement of transarticular or pedicle screw fixation in the axis

112

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

Fig. 6.52  Artistic drawing of intralaminar C1 – short isthmic C2 AA fixation suggested by Donnellan

Our Preference In our opinion, only rarely does a situation arise requiring the use of this construct. Nonetheless, it does represent a salvage option if more robust constructs are impossible. We suppose that a construct connecting C1 and C1 laminar screws is another less stable, salvage option of atlantoaxial fixation.

6.4.2 Anterior Monosegmental Fusion Constructs 6.4.2.1 Anterior Screw Fixation of C2-1 Lesoin et  al. [141] were the first to perform anterior transarticular C2-1 fusion in six patients with acute or chronic posttraumatic C1-2 instability (Fig.  6.14). They used right-sided high anterolateral approach to place 25–35  mm long screws from C2 body perpendicularly through the joint reaching the C1 lateral masses. However, their approach was quite extensive and the head rotated 15° away from the exposed side. Also, the indications were not in concordance with current expert opinions. A similar technique, but performed from less invasive, oblique approach (like for odontoid screw) with intraarticular curette decortications, was described by Sonntag and Dickman [45, 209]. The same group of authors later used a combination of odontoid screw and two anterior transarticular screws as a salvage procedure to stabilize an 85-yearold man with combined atlas-axis fracture [11]. Analyzing dry specimens and embalmed cadavers, anatomical guidelines for anterior atlantoaxial fixation were suggested by Lu et  al. [146]. Using the same entry points as previous authors, they recommended to place 15–25 mm long screws in lateral angle of 5°–25°

and posterior angle of 10°–25°. Vaccaro et  al. [230] performed bilateral high cervical approach, directly decorticated the atlantoaxial joints and packed them with bone graft and placed bilateral transarticular screws. They used K-wires introduced under biplanar fluoroscopy followed by tapping and cannulated screw placement. Although, the usual lateral trajectory angle of 30° is described, they actually used a 0° angle in coronal plane and 25° posterior tilt in sagittal plane. The screws were 26  mm long on both sides in their case. The patient was placed in long-term, postoperative halo vest as the original indication for surgery was a failed posterior Brooks’s fixation for odontoid pseudoarthrosis. In 1999, Knöller et al. presented the first larger series of patients (11) treated for odontoid process pseudoarthrosis with only temporary anterior transarticular fixation [127]. Another case of fixation of combined unstable atlantoaxial fractures (odontoid and C1) treated with anterior triple screw fixation was published by Reindl et al. [184]. The initial concerns about the short-term stability of anterior atlantoaxial transarticular screw fixation techniques were disputed by biomechanical work of Sen et al. [202]. They tested posterior versus anterior transarticular fixation techniques in nine cadavers and did not find any significant differences between these two techniques. However, they discovered that if a cable fixed graft is added to the posterior fixation, stability was significantly higher especially in flexion and extension. They placed the screw in lateral angle of 20° and posteriorly tilted in an angle of 30°. The screws were introduced perpendicularly to the joint fissure in the middle third of articulation. The entry point was located in the groove created by articulation process in the middle of C2 body. Koller et al. suggested different trajectory for anterior transarticular screw fixation, having performed a thorough anatomical analysis of fine CT scans with 3D reconstructions in 42 healthy individuals and

113

6.4  Monosegmental Fusion Constructs

comprehensive literature review [131]. They described a safe zone of approximately 14 mm in the midsagittal base of C2 vertebral body. They proposed a precise algorithm for safe screw placement from the base of C2 through the body to the C1 lateral mass (Fig. 6.15). The transcorporeal route logically enables the use of a longer screw bone passage and therefore also a higher construct stiffness. On the other hand, using the C2 pinafore does not allow triple screw introduction if odontoid screw is needed. Respecting their own anatomical results, the authors successfully fused atlantoaxial joint in seven patients concluding that the indication for anterior can be the same as for posterior fusion and thus anterior procedure can be recommended especially if anatomical situation precludes the posterior one. Advantages – – – – – –

Less muscular damage Smaller risk of VA injury Relatively easy approach (no cavity opened) Trajectory angle is easy to achieve Possible combination with odontoid screw technique Comparable short-time stability to other screw techniques – Theoretical possibility to extend the screw trajectory to fix the occipital condyle

occipital joint. The exact preoperative radiological analysis and perioperative biplanar fluoroscopy are mandatory. In our opinion, the main disadvantage is that only intra-articular surface can be used to potentiate the fusion but no additional graft can be added out of it. Questionable long-term stability makes this method more exceptional and useful in rare specific cases as it is documented by the lack of published larger series of patients.

6.4.2.3 Anterior Plate or Construct C1-2 Anterior atlantoaxial plate to stabilize the C1-2 after transoral odontoidectomy was proposed by Harms et  al. [95]. The plate was fixed by two screws to the anterior lateral C1 masses and by two other to the C2 vertebral body (Fig.  6.53). Fifth screw transfixed the axis body at the base of the odontoid process. Their technique was biomechanically tested by Kandziora et al. [121] who found out that, only if supplemented by posterior Brooks fusion, is this technique comparable in rigidity to Magerl’s procedure alone. Previous biomechanical conclusion was confirmed by clinical series of 15 patients treated for irreducible atlantoaxial

Disadvantages – – – – –

Limited amount of bone graft, only intraarticularly Long-term stability questionable Biplanar fluoroscopy necessary Decompression, if necessary, hardly possible Possible violation of spinal canal (too posterior ­trajectory) – Possible violation of atlanto-occipital joint (too long screw) 6.4.2.2 Our Preference The technique is very elegant in some atlantoaxial combination fractures where the anterior transarticular screw technique can be combined with odontoid screw or anterior C2/3 plate (Fig.  14.3, Chap. 14). Also, in cases of failed posterior fusion or anatomical situation not allowing posterior fixation, this method can serve as a salvage procedure. The attention should be focused not only to the angles of purchase but also to the length of screws, not to unintentionally damage the atlanto-

Fig. 6.53  Anterior plate for transoral AA transoral fixation designed by Harms

114

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

kyphosis [125]. Kerschbaumer et al. saw two cases of screw loosening in their first three patients treated with Harms plate as a standalone fixation and they, therefore, always supplemented the anterior plate with posterior Brooks’s fusion with good long-term results [125]. Kandziora et al. [122] also criticized the design of Harms plate and suggested a new one with locked screws located in more condensed subarticular bone in C2 near to the base of odontoid (subarticular atlantoaxial locking plate (SAALP)). They confirmed higher biomechanical stability for their type of plate; however, clinical data for the new plate are not available. In order to avoid the need for posterior stabilization, another plate with locked screws and option to reduce atlantoaxial kyphosis was designed by Yin et al. [247]. This is done with special forceps distracting the cranial part of plate against temporarily inserted C2 body screw ( transoral atlantoaxial reduction plate (TARP)). They have successfully reduced and fixed 4 patients with odontoid pseudoarthrosis and kyphotic deformity. The demand to reconstruct anterior column with sparing of adjacent segment motion, resulted in design of more complex constructs replacing the C2 vertebral body in tumor surgery. Sar and Eralp [197] used a ­custom-modified Harms cage fixed with screws in C1 and C2 to replace the C2 anterior body in a patient with C2 sarcoma. For the same purpose, C2 body prosthesis was developed and used by Jeszenszky et al. [118].

achieve this goal, anterior atlantoaxial release must be possible. This is not always the case. In RA patients and some developmental anomalies, the joint could be severely deformed and it can be very difficult to circumferentially release it. If incompletely released, it can be dangerous to use inappropriate force to reduce it. The bone is often very weak and does not provide strong support for intra-articular distraction instruments. C1 lateral masses can be so deformed and pronounced that any screw purchase is problematic. In our opinion, if the reduction is not achievable with traction prior to the surgery, its surgical reducibility can only be realized during the procedure thus postponing the decision to use or not to use the plate to this moment. The infection risk is higher than for standard transoral procedure because the time necessary for fixation prolongs the surgery and also the lateral exposure necessary to fix the plate to C1 is much larger implicating important soft retropharyngeal tissue damage. In summary, anterior C1-2 plating can be advantageous if preoperatively irreducible AAD can be released during the surgery. Then, the most sophisticated locking plate should be used. The complex constructs used to support anterior column after tumor resection represent a different topic, which will be described in tumor chapter.

6.4.3 Lateral Monosegmental Fusion

Advantages – Direct anterior decompression – Sophisticated plate can reach enough stability to avoid posterior fixation – Atlantoaxial release possible – Atlantoaxial kyphotization can be reduced – Atlantoaxial joint can be distracted Disadvantages – Risk of infection of intracavitary approach – Wide exposure causing more damage of soft tissue (approx. 4 cm) – Possible screw loosening – Risky revision (infection) – If reduction fails, posterior compression cannot be eliminated 6.4.2.4 Our Preference A sophisticated locking plate with realignment ability can really address the problem of IAAD. However, to

Lateral atlantoaxial fusion was first used by Barbour from Australia to treat the odontoid fractures in 1971 [17]. As he mentioned, starting in 1956, he was probably the first who used stable screw fixation of C1-2. He used a skin incision along the anterior border of sternocleidomastoid and extended it behind the mandibular angle to reach the lateral position of C1 transverse process. He nibbled it partially away and asked anesthesiologist to turn the head to neutral position and introduced the screw from C1 lateral mass medially and downward transarticularly to C2 (Fig. 6.54). The same procedure was then performed from the opposite side. He recommended placing iliac crest onlay grafts on the lateral vertebral surface. Unfortunately, neither number of treated patients nor the follow-up data are presented in his original paper. Encouraged by Barbour’s work but not satisfied with approach difficulties, Du Toit modified the technique and successfully treated a patient with odontoid fracture [49]. He learnt from cadaveric dissections and used an angled skin incision to cut off the proximal attachment of SCM to reach the

115

6.5  CVJ and UCS as a Part of Multisegmental Constructs

Fig.  6.54  Drawing of lateral down-slope transarticular C1-2 fixation performed from bilateral approach according to Barbour and DuToit

transverse process of C1. Anterolateral aspect of atlas and lateral joint fissure was reached strictly subperiostally. Then he denuded the joint and filled it with autologous morselized bone. The drilling was performed caudally (25° below horizontal plane) and posteromedially with custom-made drill guide allowing maximal 10° of backward angle and had 24 mm depth stop. The pilot hole was tapped and the joint transfixed with AO navicular screws. To avoid eventual spinal canal penetration, he established 20° as a maximal posterior tilt of drilling. Later, this group described four other cases (os odontoideum and dens pseudoarthrosis) successfully treated with bilateral transarticular screw fusion [205].

6.4.3.1 Our Preference We see some important drawbacks to have this procedure in standard fusion armamentarium. First of all the bilateral access is needed, the anatomy of the approach is complicated with the accessory nerve, auricular nerve, jugular vein but namely VA in dangerous positions. Second, there is not enough bony surface for eventual bone graft insertion and one has to believe that intra-articular fusion potential would be enough.

6.5 CVJ and UCS as a Part of Multisegmental Constructs CVJ and UCS instability are disorders caused by ­various etiologies including trauma, inflammation, developmental anomaly, tumors, degenerative disease,

and/or iatrogenic decompression. During stabilization procedures, one should avoid undesirable fusion of disease-free segments and attempt to fix only the unstable spinal motion units. This is of utmost importance in CVJ, the most mobile area of the spine. Generally, we can divide long constructs to those involving the occiput (occipitocervical) and those starting subaxially and ending at C2 or C1 (suboccipital constructs). The decision process always involves striking a balance between the loss of range of motion versus the required extent of construct anchorage. Occipitocervical fusion is indicated when the CVJ, namely occipitoatlantal segment, is unstable or it is expected that further progression of pathological process can involve this joint (RA patients, tumors, etc.). Any multisegmental fusion has to be supplemented with bone grafts with or without addition of bone growth accelerators. The exceptions from this rule are those with secondary bone tumors with limited life expectancy.

6.5.1 Occipitocervical Constructs By definition, this technique always involves the occipital bone. The fusion extends at least to C2, often to subaxial vertebrae, and sometimes is performed as skip framework starting at occiput skipping C1 and/or C2 and ending usually in two or three level fixation to subaxial lateral masses or pedicles. Performing any procedure fixating the cervical spine to head the surgeon must adapt the craniospinal angle in sagittal plane and respect the neutral rotational position [155]. This is to allow the patient neutral horizontal view otherwise the dictated position in non physiological flexion, extension, and/or rotation will lead to compensatory deformation of spine balance. In patients with substantial sagittal profile derangement, we also have to calculate the whole spine profile with potential planning of corrective spinal osteotomy in other regions. It is well known that hyperflexed cervical spine can also cause swallowing and breathing difficulties [15]. Foerster was the first who described the occipitocervical fusion with the use of fibular strut graft, in 1927 [64]. Newman and Sweetman, in 1969, reported a series of nine patients treated with occipitocervical onlay autograft. They had only one pseudoarthrosis but their patients were postoperatively treated with 6 weeks of tong traction and then placed in Minerva

116

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

jacket for another 6 weeks [169]. Freely lodged standalone grafts did not secure any stability and the fusion rate can be also very low [62]. Therefore, wiring techniques were introduced to fix these grafts [78, 92, 110, 235]. Also, methylmethacrylate was used to enhance the occipitocervical stabilization [25]. Later, Ransford et al. introduced contoured Lugue rod again fixed with wire to posterior elements as a more stable technique [183]. This wire and frame application was further modified using different metal rods or pins [10, 59, 113]. All the techniques using wires for fixation, however, had some drawbacks. The stability was not very high and the external fixation with halo vest or Minerva jacket had to be used, the posterior elements (occipital bone, laminae) had to be intact, sublaminary introduced hardware can injure the underlying neural structures and their coverings [149, 171, 211, 221] and the wires often have a tendency to abrade through the bone [43]. Despite prolonged external immobilization, in halo vest or Minerva cast, the wire fixation methods have failure rates up to 30% [145, 190]. The previously mentioned techniques most often used doubled holes in the occipital calvarium to pass the wires through. To increase the stability of so called “semi-rigid techniques,” Grob et  al. [81] described Y plate connecting the transarticular C1-2 screws with the occiput, where two screws fixed it to midline. Comparing cohort of patients treated with this technique with another group of patients fixed with older graft wiring, the same authors found pseudoarthrosis rate of 6% vs. 27%, respectively [85]. Grob was also the first who performed the occipital midline screw placement as he knew that there is the thickest bone available. This new plate and screw “rigid” technique was confirmed as good enough to provide up to 100% of fusion without indispensability of hard external supports by Sasso et al. using two AO plates fixed by screws to occiput and to Magerl screws, caudally [199]. Promising results of other authors gained popularity for plate and screw occipitocervical fusion [175, 196, 207]. Screw and plate techniques were also confirmed as much more stable than all the other previous by many biomechanical studies thus allowing to fuse less segments than before [7, 111, 173]. Their main concern is that the plate, although contoured, has defined holes for screws and thus determines their position often to less than optimal location. The other problem is in lining of the construct. This means that most of plates even if bended are reaching lateral occipital regions where

there is limited thickness of the bone. Screw purchase cannot be strong enough and bicortical introduction is potentially dangerous. Pait et al. suggested the insideoutside technique to avoid the screw loosening, subdural injury, and allow lateral occipital purchase [175]. They performed a trephination out of finally planed screw location and then cut a slit with craniotome ending at the desired position. The flat screw head is passed to this position epidurally from trephination hole and the nut is then used to fix it to the plate. Nevertheless, this sophisticated technique could be impossible in patients with very thin bone. Another concern of plate and screw technique is higher frequency of screw breakage and pullout caused by stress transmitted to hardware interfaces. To overcome the problem, modular systems were introduced [1, 170, 178]. Most of them have isolated (non dependent) occipital plate with a variety of available screw positions. This plate is connected with malleable rod to the screws (mostly, polyaxial ones) introduced in C1 lateral masses, C2 anchors, and/or to subaxial spine screws. Currently, the instrumentation is made up from Titanium alloy and the rods usually fortified or thickened in the place of craniovertebral bending. Although the variability of modular systems is providing much more flexibility in surgical decision there are still advocates of the use of the plate and screw systems [170]. Their main argument is higher resistance in lateral bending confirmed by biomechanical works [7] and the ability to correct deformity according the pre-shaped contour of the plate during tightening of the screws. Nevertheless, the potential to correct a deformity is probably higher using the modular systems [1].

6.5.1.1 Our Preference According to Grob [84], the ideal CVJ fixating system should: fix only the target segment, not encroach into the spinal canal, provide immediate reduction and stability, and be effective if laminae are absent. We would like to add that currently there is also a desire to have a system strong enough to provide long-term stability, made up from biocompatible and MRI-friendly materials and modular – means easy to use. The modular facility is very important not only because it makes the surgical work more comfortable but also it allows correct placement of anchoring elements (screws) without any stress created either to hardware or the underlying

117

References

bone. The other demand is its low profile especially in occipital region and connectivity to eventual continuing caudal fusion constructs (RA patients). In summary, there is a very rare indication for semirigid fixation of adult CVJ nowadays. On the other hand, there is still a place for autologous grafts and wires in very small children with surgically corrected instabilities and deformities in our opinion. Their potential for bony healing is very high as well their adaptability to external hard brace wearing. The correctly longitudinally placed autograft (often, the rib) can grow without limitations given by fixed hardware. In some cases, temporary metal fixation can be considered as well. Nowadays, we are not using the plates and screws. The restraints of screw positioning determined by fixed position of holes do not allow introducing the screws correctly and the mechanical stress created after tightening of screws is too large. The consequent potential screw breakage and/or loosening usually require revision and prolong the bone fusion. Also, the required contouring of the plates to reach the acceptable angle in craniovertebral transition (up to 80°) can lead to material microfractures and weakness. The midline occipital screw anchorage, which is in our opinion the strongest and safest, can be hardly achieved even with pre-bend and medially angulated plates. The biomechanical argument supporting the use of plates because of greater stiffness can be overcome with the new developments where the connecting rods are reinforced and the use of strut grafts is a regular part of the constructs. Modern modular systems allow independent placement of screws in the most suitable positions without any stress (Figs. 19.23 and 19.29, Chap. 19). They allow multidirectional manipulation (reduction, distraction, rotation, etc.) with the help of supplemented forceps adapted to be able to connect different screws. The grafts can be placed easily because the constructs are more subtle and not covering the acceptor side. The fusion rates and stability achieved are similar or even superior to previously described modalities. As in other long constructs, the final goal – bony fusion has to be achieved. This is of utmost importance, especially in the dangerous UCS region. Correct decortication of acceptor area and preferably autologous bone should be used in our opinion. Despite all the potential complications accompanying the autologous bone harvesting, its osteogenic, osteoinductive, and osteoconductive potential cannot be substituted by any other material. Although the allograft, bone

substitute, or BMP can be added, autologous bone should always form the majority of the graft content.

6.5.2 Suboccipital Constructs Whenever possible, we prefer to exclude the occiput out of the fusion. Atlanto-occipital joint is responsible for up to 40% of cervical flexion and extension and therefore should be spared. The same can be said about the atlantoaxial connection responsible for 60% of cervical spine rotations. It is clear that if a solid anchorage can be used then it is not necessary to prolong the fixation to a desirable length. For example, a C2 pedicle screw, if acceptable as cranial construct end, is one of the most firm anchorages available in cervical spine. On the other hand, if short isthmic screw is chosen, then one can hesitate about its strength and extend the fixation to C1. The majority of suboccipital multisegmental fusions are performed for complex, combined surgeries treating multilevel stenosis, deformity, tumors, or infection [201].

6.5.3 Anterior Multisegmental Constructs Large decompressions in the deformity cases and mainly in tumors can lead to important loss of structures supporting anterior CVJ and UCS. In such cases, posterior occipitocervical fusion can be considered as insufficient and the anterior column is reconstructed with custom-made titanium mesh cages fixed with or without plates to different anatomical structures, anteriorly [187, 217].

References   1. Abumi, K., Takada, T., Shono, Y., et al.: Posterior occipitocervical reconstruction using cervical pedicle screws and platerod systems. Spine (Phila Pa 1976) 24, 1425–1434 (1999)   2. Aebi, M., Etter, C., Coscia, M.: Fractures of the odontoid process. Treatment with anterior screw fixation. Spine (Phila Pa 1976) 14, 1065–1070 (1989)   3. Ai, F., Yin, Q., Wang, Z., et al.: Applied anatomy of transoral atlantoaxial reduction plate internal fixation. Spine (Phila Pa 1976) 31, 128–132 (2006)

118

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

  4. Aldrich, E.F., Weber, P.B., Crow, W.N.: Halifax interlaminar clamp for posterior cervical fusion: a long-term follow-up review. J Neurosurg 78, 702–708 (1993)   5. Alosh, H., Parker, S.L., McGirt, M.J., et  al.: Preoperative radiographic factors and surgeon experience are associated with cortical breach of C2 pedicle screws. J Spinal Disord Tech 23, 9–14 (2010)   6. Ames, C.P., Crawford, N.R., Chamberlain, R.H., et  al.: Biomechanical evaluation of a bioresorbable odontoid screw. J Neurosurg Spine 2, 182–187 (2005)   7. Anderson, P.A., Oza, A.L., Puschak, T.J., et al.: Biomechanics of occipitocervical fixation. Spine (Phila Pa 1976) 31, 755–761 (2006)   8. Aota, Y., Honda, A., Uesugi, M., et  al.: Vertebral artery injury in C-1 lateral mass screw fixation. Case illustration. J Neurosurg Spine 5, 554 (2006)   9. Apfelbaum, R.I., Lonser, R.R., Veres, R., et al.: Direct anterior screw fixation for recent and remote odontoid fractures. J Neurosurg 93, 227–236 (2000) 10. Apostolides, P.J., Dickman, C.A., Golfinos, J.G., et  al.: Threaded steinmann pin fusion of the craniovertebral junction. Spine (Phila Pa 1976) 21, 1630–1637 (1996) 11. Apostolides, P.J., Theodore, N., Karahalios, D.G., et  al.: Triple anterior screw fixation of an acute combination atlasaxis fracture. Case report. J Neurosurg 87, 96–99 (1997) 12. Arlet, V., Datta, J.C.: Upper cervical spine. In: Aebi, M., Arlet, V., Webb, J.K. (eds.) AO spine manual: principles and techniques, vol I, pp. 265–288. Thieme Verlag, Stuttgart, New York (2007) 13. Arnautovic, K.I., al-Mefty, O., Pait, T.G., et al.: The suboccipital cavernous sinus. J Neurosurg 86, 252–262 (1997) 14. Aryan, H.E., Newman, C.B., Nottmeier, E.W., et  al.: Stabilization of the atlantoaxial complex via C-1 lateral mass and C-2 pedicle screw fixation in a multicenter clinical experience in 102 patients: modification of the Harms and Goel techniques. J Neurosurg Spine 8, 222–229 (2008) 15. Bagley, C.A., Witham, T.F., Pindrik, J.A., et  al.: Assuring optimal physiologic craniocervical alignment and avoidance of swallowing-related complications after occipitocervical fusion by preoperative halo vest placement. J Spinal Disord Tech 22, 170–176 (2009) 16. Bambakidis, N.C., Feiz-Erfan, I., Horn, E.M., et  al.: Biomechanical comparison of occipitoatlantal screw fixation techniques. J Neurosurg Spine 8, 143–152 (2008) 17. Barbour, J.R.: Screw fixation in fracture of the odontoid process. S Aust Clin 5, 20–24 (1971) 18. Blagg, S.E., Don, A.S., Robertson, P.A.: Anatomic determination of optimal entry point and direction for C1 lateral mass screw placement. J Spinal Disord Tech 22, 233–239 (2009) 19. Bloch, O., Holly, L.T., Park, J., et  al.: Effect of frameless stereotaxy on the accuracy of C1-2 transarticular screw placement. J Neurosurg 95, 74–79 (2001) 20. Bohler, J.: Anterior stabilization for acute fractures and nonunions of the dens. J Bone Joint Surg Am 64, 18–27 (1982) 21. Bohm, H., Kayser, R., El Saghir, H., et al.: Direct osteosynthesis of instable Gehweiler Type III atlas fractures. Presentation of a dorsoventral osteosynthesis of instable atlas fractures while maintaining function. Unfallchirurg 109, 754–760 (2006) 22. Borne, G.M., Bedou, G.L., Pinaudeau, M.: Treatment of pedicular fractures of the axis. A clinical study and screw fixation technique. J Neurosurg 60, 88–93 (1984)

23. Brockmeyer, D.L., York, J.E., Apfelbaum, R.I.: Anatomical suitability of C1-2 transarticular screw placement in pediatric patients. J Neurosurg 92, 7–11 (2000) 24. Brooks, A.L., Jenkins, E.B.: Atlanto-axial arthrodesis by the wedge compression method. J Bone Joint Surg Am 60, 279–284 (1978) 25. Bryan, W.J., Inglis, A.E., Sculco, T.P., et  al.: Methyl­ methacrylate stabilization for enhancement of posterior cervical arthrodesis in rheumatoid arthritis. J Bone Joint Surg Am 64, 1045–1050 (1982) 26. Cacciola, F., Phalke, U., Goel, A.: Vertebral artery in relationship to C1-C2 vertebrae: an anatomical study. Neurol India 52, 178–184 (2004) 27. Cassinelli, E.H., Lee, M., Skalak, A., et al.: Anatomic considerations for the placement of C2 laminar screws. Spine (Phila Pa 1976) 31, 2767–2771 (2006) 28. Castillo, C., Vinas, F.C., Gutikhonda, M., et al.: Microsurgical anatomy of the suboccipital segment of the vertebral artery. Neurol Res 20, 201–208 (1998) 29. Chamoun, R.B., Whitehead, W.E., Curry, D.J., et  al.: Computed tomography morphometric analysis for C-1 lateral mass screw placement in children. Clinical article. J Neurosurg Pediatr 3, 20–23 (2009) 30. Chen, J.F., Wu, C.T., Lee, S.C., et al.: Posterior atlantoaxial transpedicular screw and plate fixation. Technical note. J Neurosurg Spine 2, 386–392 (2005) 31. Christensen, D.M., Eastlack, R.K., Lynch, J.J., et  al.: C1 anatomy and dimensions relative to lateral mass screw placement. Spine (Phila Pa 1976) 32, 844–848 (2007) 32. Claybrooks, R., Kayanja, M., Milks, R., et al.: Atlantoaxial fusion: a biomechanical analysis of two C1-C2 fusion techniques. Spine J 7, 682–688 (2007) 33. Cone, W., Turner, W.G.: The treatment of fracture-dislocations of the cervical vertebrae by skeletal traction and fusion. J Bone Joint Surg Am 19, 584–602 (1937) 34. Cooper, D.F.: Bone erosion of the cervical vertebrae secondary to tortuosity of the vertebral artery: case report. J Neurosurg 53, 106–108 (1980) 35. Cornefjord, M., Henriques, T., Alemany, M., et al.: Posterior atlanto-axial fusion with the Olerud Cervical Fixation System for odontoid fractures and C1-C2 instability in rheumatoid arthritis. Eur Spine J 12, 91–96 (2003) 36. Coyne, T.J., Fehlings, M.G., Wallace, M.C., et  al.: C1-C2 posterior cervical fusion: long-term evaluation of results and efficacy. Neurosurgery 37, 688–692 (1995). discussion 692-683 37. Currier, B.L., Maus, T.P., Eck, J.C., et al.: Relationship of the internal carotid artery to the anterior aspect of the C1 vertebra: implications for C1-C2 transarticular and C1 lateral mass fixation. Spine (Phila Pa 1976) 33, 635–639 (2008) 38. Currier, B.L., Todd, L.T., Maus, T.P., et al.: Anatomic relationship of the internal carotid artery to the C1 vertebra: a case report of cervical reconstruction for chordoma and pilot study to assess the risk of screw fixation of the atlas. Spine (Phila Pa 1976) 28, E461–E467 (2003) 39. Cybulski, G.R., Stone, J.L., Crowell, R.M., et  al.: Use of Halifax interlaminar clamps for posterior C1-C2 arthrodesis. Neurosurgery 22, 429–431 (1988) 40. Cyr, S.J., Currier, B.L., Eck, J.C., et al.: Fixation strength of unicortical versus bicortical C1-C2 transarticular screws. Spine J 8, 661–665 (2008)

References 41. De Iure, F., Donthineni, R., Boriani, S.: Outcomes of C1 and C2 posterior screw fixation for upper cervical spine fusion. Eur Spine J 18(Suppl 1), 2–6 (2009) 42. Dickman, C.A., Mariano, F.F.: Principles and techniques of screw fixation in the cervical spine. In: Menezes, A.H., Sonntag, V.K.H. (eds.) Principles of spinal surgery, pp. 123–139. McGraw-Hill, New York (1996) 43. Dickman, C.A., Papadopoulos, S.M., Crawford, N.R., et al.: Comparative mechanical properties of spinal cable and wire fixation systems. Spine (Phila Pa 1976) 22, 596–604 (1997) 44. Dickman, C.A., Sonntag, V.K.: Posterior C1-C2 transarticular screw fixation for atlantoaxial arthrodesis. Neurosurgery 43, 275–280 (1998). discussion 280–271 45. Dickman, C.A., Sonntag, V.K.H., Marcotte, P.J.: Techniques of screw fixation of the cervical spine. BNI Q 8, 9–26 (1992) 46. Dickman, C.A., Sonntag, V.K., Papadopoulos, S.M., et al.: The interspinous method of posterior atlantoaxial arthrodesis. J Neurosurg 74, 190–198 (1991) 47. Dmitriev, A.E., Lehman Jr., R.A., Helgeson, M.D., et  al.: Acute and long-term stability of atlantoaxial fixation methods: a biomechanical comparison of pars, pedicle, and intralaminar fixation in an intact and odontoid fracture model. Spine (Phila Pa 1976) 34, 365–370 (2009) 48. Donnellan, M.B., Sergides, I.G., Sears, W.R.: Atlantoaxial stabilization using multiaxial C-1 posterior arch screws. J Neurosurg Spine 9, 522–527 (2008) 49. Du Toit, G.: Lateral atlanto-axial arthrodesis. A screw fixation technique. S Afr J Surg 14, 9–12 (1976) 50. Duff, T.A., Khan, A., Corbett, J.E.: Surgical stabilization of cervical spinal fractures using methyl methacrylate. Technical considerations and long-term results in 52 patients. J Neurosurg 76, 440–443 (1992) 51. Dull, S.T., Toselli, R.M.: Preoperative oblique axial computed tomographic imaging for C1-C2 transarticular screw fixation: technical note. Neurosurgery 37, 150–151 (1995). discussion 151–152 52. Dvorak, M.F., Fisher, C., Boyd, M., et al.: Anterior occiputto-axis screw fixation: part I: a case report, description of a new technique, and anatomical feasibility analysis. Spine (Phila Pa 1976) 28, E54–E60 (2003) 53. Dvorak, M.F., Sekeramayi, F., Zhu, Q., et  al.: Anterior occiput to axis screw fixation: part II: a biomechanical comparison with posterior fixation techniques. Spine (Phila Pa 1976) 28, 239–245 (2003) 54. Ebraheim, N.A., Fow, J., Xu, R., et al.: The location of the pedicle and pars interarticularis in the axis. Spine (Phila Pa 1976) 26, E34–E37 (2001) 55. Ebraheim, N.A., Lu, J., Biyani, A., et al.: An anatomic study of the thickness of the occipital bone. Implications for occipitocervical instrumentation. Spine (Phila Pa 1976) 21, 1725–1729 (1996). discussion 1729-1730 56. Ebraheim, N.A., Misson, J.R., Xu, R., et  al.: The optimal transarticular C1-2 screw length and the location of the hypoglossal nerve. Surg Neurol 53, 208–210 (2000) 57. Ebraheim, N.A., Xu, R., Ahmad, M., et al.: The quantitative anatomy of the vertebral artery groove of the atlas and its relation to the posterior atlantoaxial approach. Spine (Phila Pa 1976) 23, 320–323 (1998) 58. Eck, J.C., Walker, M.P., Currier, B.L., et al.: Biomechanical comparison of unicortical versus bicortical C1 lateral mass screw fixation. J Spinal Disord Tech 20, 505–508 (2007)

119 59. Fehlings, M.G., Errico, T., Cooper, P., et al.: Occipitocervical fusion with a five-millimeter malleable rod and segmental fixation. Neurosurgery 32, 198–207 (1993). discussion 207–198 60. Feiz-Erfan, I., Gonzalez, L.F., Dickman, C.A.: Atlantooccipital transarticular screw fixation for the treatment of traumatic occipitoatlantal dislocation. Technical note. J Neurosurg Spine 2, 381–385 (2005) 61. Finn, M.A., Apfelbaum, R.I.: Atlantoaxial transarticular screw fixation: update on technique and outcomes in 269 patients. Neurosurgery 66, A184–A192 (2010) 62. Finn, M.A., Bishop, F.S., Dailey, A.T.: Surgical treatment of occipitocervical instability. Neurosurgery 63, 961–968 (2008). discussion 968–969 63. Floyd, T., Grob, D.: Translaminar screws in the atlas. Spine (Phila Pa 1976) 25, 2913–2915 (2000) 64. Foerster, O.: Leitungsbahnen des Schmerzgefuhls und die chirurgische Behandlung der Schmerzzustande. Urban & Schwarzenberg, Berlin (1927) 65. Foley, K.T., Silveri, C.P., Vaccaro, A.R., et al.: Atlantoaxial transarticular screw fixation: Risk assessment and bone morphology using an image guidance system. J Bone Joint Surg Br 80(Suppl), 245 (1998) 66. Fried, L.C.: Atlanto-axial fracture-dislocations. Failure of posterior C.1 to C.2 fusion. J Bone Joint Surg Br 55, 490– 496 (1973) 67. Gallie, W.E.: Fractures and dislocations of the cervical spine. Am J Surg 46, 495–499 (1939) 68. Geisler, F.H., Cheng, C., Poka, A., et  al.: Anterior screw fixation of posteriorly displaced type II odontoid fractures. Neurosurgery 25, 30–37 (1989). discussion 37–38 69. Gluf, W.M., Schmidt, M.H., Apfelbaum, R.I.: Atlantoaxial transarticular screw fixation: a review of surgical indications, fusion rate, complications, and lessons learned in 191 adult patients. J Neurosurg Spine 2, 155–163 (2005) 70. Goel, A., Desai, K.I., Muzumdar, D.P.: Atlantoaxial fixation using plate and screw method: a report of 160 treated patients. Neurosurgery 51, 1351–1356 (2002). discussion 1356–1357 71. Goel, A., Kulkarni, A.G., Sharma, P.: Reduction of fixed atlantoaxial dislocation in 24 cases: technical note. J Neurosurg Spine 2, 505–509 (2005) 72. Goel, A., Laheri, V.: Plate and screw fixation for atlanto-axial subluxation. Acta Neurochir (Wien) 129, 47–53 (1994) 73. Goel, A., Shah, A., Rajan, S.: Vertical mobile and reducible atlantoaxial dislocation. Clinical article. J Neurosurg Spine 11, 9–14 (2009) 74. Goffin, J., Van Brussel, K., Martens, K., et  al.: Threedimensional computed tomography-based, personalized drill guide for posterior cervical stabilization at C1-C2. Spine (Phila Pa 1976) 26, 1343–1347 (2001) 75. Gonzalez, L.F., Crawford, N.R., Chamberlain, R.H., et al.: Craniovertebral junction fixation with transarticular screws: biomechanical analysis of a novel technique. J Neurosurg 98, 202–209 (2003) 76. Gonzalez, L.F., Klopfenstein, J.D., Crawford, N.R., et  al.: Use of dual transarticular screws to fixate simultaneous occipitoatlantal and atlantoaxial dislocations. J Neurosurg Spine 3, 318–323 (2005) 77. Gorek, J., Acaroglu, E., Berven, S., et  al.: Constructs ­incorporating intralaminar C2 screws provide rigid stability for atlantoaxial fixation. Spine (Phila Pa 1976) 30, 1513–1518 (2005)

120

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

78. Grantham, S.A., Dick, H.M., Thompson Jr., R.C., et  al.: Occipitocervical arthrodesis. Indications, technic and results. Clin Orthop Relat Res 65, 118–129 (1969) 79. Graziano, G., Jaggers, C., Lee, M., et  al.: A comparative study of fixation techniques for type II fractures of the odontoid process. Spine (Phila Pa 1976) 18, 2383–2387 (1993) 80. Griswold, D.M., Albright, J.A., Schiffman, E., et al.: Atlantoaxial fusion for instability. J Bone Joint Surg Am 60, 285–292 (1978) 81. Grob, D.: Transarticular screw fixation for atlanto-occipital dislocation. Spine (Phila Pa 1976) 26, 703–707 (2001) 82. Grob, D., 3rd Crisco, J.J., Panjabi, M.M., et al.: Biomechanical evaluation of four different posterior atlantoaxial fixation techniques. Spine (Phila Pa 1976) 17, 480–490 (1992) 83. Grob, D., Bremerich, F.H., Dvorak, J., et al.: Transarticular screw fixation for osteoarthritis of the atlanto axial segment. Eur Spine J 15, 283–291 (2006) 84. Grob, D., Dvorak, J., Panjabi, M., et al.: Posterior occipitocervical fusion. A preliminary report of a new technique. Spine (Phila Pa 1976) 16, S17–S24 (1991) 85. Grob, D., Dvorak, J., Panjabi, M.M., et al.: The role of plate and screw fixation in occipitocervical fusion in rheumatoid arthritis. Spine (Phila Pa 1976) 19, 2545–2551 (1994) 86. Grob, D., Jeanneret, B., Aebi, M., et al.: Atlanto-axial fusion with transarticular screw fixation. J Bone Joint Surg Br 73, 972–976 (1991) 87. Grob, D., Magerl, F.: Operative stabilisierung bei frakturen von C1 und C2. Orthopäde 16, 46–54 (1987) 88. Gunnarsson, T., Massicotte, E.M., Govender, P.V., et al.: The use of C1 lateral mass screws in complex cervical spine surgery: indications, techniques, and outcome in a prospective consecutive series of 25 cases. J Spinal Disord Tech 20, 308–316 (2007) 89. Gupta, S., Goel, A.: Quantitative anatomy of the lateral masses of the atlas and axis vertebrae. Neurol India 48, 120–125 (2000) 90. Haher, T.R., Yeung, A.W., Caruso, S.A., et  al.: Occipital screw pullout strength. A biomechanical investigation of occipital morphology. Spine (Phila Pa 1976) 24, 5–9 (1999) 91. Haid Jr., R.W., Subach, B.R., McLaughlin, M.R., et  al.: C1-C2 transarticular screw fixation for atlantoaxial instability: a 6-year experience. Neurosurgery 49, 65–68 (2001). discussion 69–70 92. Hamblen, D.L.: Occipito-cervical fusion. Indications, technique and results. J Bone Joint Surg Br 49, 33–45 (1967) 93. Hanson, P.B., Montesano, P.X., Sharkey, N.A., et  al.: Anatomic and biomechanical assessment of transarticular screw fixation for atlantoaxial instability. Spine (Phila Pa 1976) 16, 1141–1145 (1991) 94. Harms, J., Melcher, R.P.: Posterior C1-C2 fusion with polyaxial screw and rod fixation. Spine (Phila Pa 1976) 26, 2467–2471 (2001) 95. Harms, J., Schmelze, R., Stolze, D.: Osteosynthesen im occipito-cervikalen Übergang vom transoralen Zugang aus, XVII SICOT World Congress Abstracts. Demeter Verlag, Munich (1987) 96. Hashizume, H., Kawakami, M., Kawai, M., et al.: A clinical case of endoscopically assisted anterior screw fixation for the type II odontoid fracture. Spine (Phila Pa 1976) 28, E102–E105 (2003)

97. Hedequist, D., Proctor, M.: Screw fixation to C2 in children: a case series and technical report. J Pediatr Orthop 29, 21–25 (2009) 98. Heuer, G.G., Hardesty, D.A., Bhowmick, D.A., et  al.: Treatment of pediatric atlantoaxial instability with traditional and modified Goel-Harms fusion constructs. Eur Spine J 18, 884–892 (2009) 99. Holly, L.T., Foley, K.T.: Intraoperative spinal navigation. Spine (Phila Pa 1976) 28, S54–S61 (2003) 100. Holly, L.T., Foley, K.T.: Percutaneous placement of posterior cervical screws using three-dimensional fluoroscopy. Spine (Phila Pa 1976) 31, 536–540 (2006). discussion 541 101. Holness, R.O., Huestis, W.S., Howes, W.J., et al.: Posterior stabilization with an interlaminar clamp in cervical injuries: technical note and review of the long term experience with the method. Neurosurgery 14, 318–322 (1984) 102. Hong, X., Dong, Y., Yunbing, C., et  al.: Posterior screw placement on the lateral mass of atlas: an anatomic study. Spine (Phila Pa 1976) 29, 500–503 (2004) 103. Hong, J.T., Lee, S.W., Son, B.C., et al.: Hypoglossal nerve palsy after posterior screw placement on the C-1 lateral mass. Case report. J Neurosurg Spine 5, 83–85 (2006) 104. Hong, J.T., Lee, S.W., Son, B.C., et  al.: Analysis of anatomical variations of bone and vascular structures around the posterior atlantal arch using three-dimensional computed tomography angiography. J Neurosurg Spine 8, 230–236 (2008) 105. Hong, J.T., Yi, J.S., Kim, J.T., et al.: Clinical and radiologic outcome of laminar screw at C2 and C7 for posterior instrumentation-review of 25 cases and comparison of C2 and C7 intralaminar screw fixation. Surg Neurol 6, 6 (2009) 106. Hott, J.S., Lynch, J.J., Chamberlain, R.H., et  al.: Biomechanical comparison of C1-2 posterior fixation techniques. J Neurosurg Spine 2, 175–181 (2005) 107. Howington, J.U., Kruse, J.J., Awasthi, D.: Surgical anatomy of the C-2 pedicle. J Neurosurg 95, 88–92 (2001) 108. Hu, Y., Ma, W., Xu, R.: Transoral osteosynthesis C1 as a function-preserving option in the treatment of bipartite atlas deformity: a case report. Spine (Phila Pa 1976) 34, E418–E421 (2009) 109. Huang, C.I., Chen, I.H.: Atlantoaxial arthrodesis using Halifax interlaminar clamps reinforced by halo vest immobilization: a long-term follow-up experience. Neurosurgery 38, 1153–1156 (1996). discussion 1156-1157 110. Huhn, S.L., Wolf, A.L., Ecklund, J.: Posterior spinal osteosynthesis for cervical fracture/dislocation using a flexible multistrand cable system: technical note. Neurosurgery 29, 943–946 (1991) 111. Hurlbert, R.J., Crawford, N.R., Choi, W.G., et al.: A biomechanical evaluation of occipitocervical instrumentation: screw compared with wire fixation. J Neurosurg 90, 84–90 (1999) 112. Igarashi, T., Kikuchi, S., Sato, K., et al.: Anatomic study of the axis for surgical planning of transarticular screw fixation. Clin Orthop Relat Res 408, 162–166 (2003) 113. Itoh, T., Tsuji, H., Katoh, Y., et al.: Occipito-cervical fusion reinforced by Luque’s segmental spinal instrumentation for rheumatoid diseases. Spine (Phila Pa 1976) 13, 1234–1238 (1988)

References 114. Jea, A., Sheth, R.N., Vanni, S., et  al.: Modification of Wright’s technique for placement of bilateral crossing C2 translaminar screws: technical note. Spine J 8, 656–660 (2008) 115. Jeanneret, B.: Posterior rod system of the cervical spine: a new implant allowing optimal screw insertion. Eur Spine J 5, 350–356 (1996) 116. Jeanneret, B., Magerl, F.: Primary posterior fusion C1/2 in odontoid fractures: indications, technique, and results of transarticular screw fixation. J Spinal Disord 5, 464–475 (1992) 117. Jenkins, J.D., Coric, D., Branch Jr., C.L.: A clinical comparison of one- and two-screw odontoid fixation. J Neurosurg 89, 366–370 (1998) 118. Jeszenszky, D., Fekete, T.F., Melcher, R., et al.: C2 prosthesis: anterior upper cervical fixation device to reconstruct the second cervical vertebra. Eur Spine J 16, 1695–1700 (2007) 119. Jones, E.L., Heller, J.G., Silcox, D.H., et al.: Cervical pedicle screws versus lateral mass screws. Anatomic feasibility and biomechanical comparison. Spine (Phila Pa 1976) 22, 977–982 (1997) 120. Jun, B.Y.: Anatomic study for ideal and safe posterior C1-C2 transarticular screw fixation. Spine (Phila Pa 1976) 23, 1703–1707 (1998) 121. Kandziora, F., Kerschbaumer, F., Starker, M., et  al.: Biomechanical assessment of transoral plate fixation for atlantoaxial instability. Spine (Phila Pa 1976) 25, 1555– 1561 (2000) 122. Kandziora, F., Pflugmacher, R., Ludwig, K., et  al.: Biomechanical comparison of four anterior atlantoaxial plate systems. J Neurosurg 96, 313–320 (2002) 123. Kandziora, F., Schulze-Stahl, N., Khodadadyan-Klostermann, C., et al.: Screw placement in transoral atlantoaxial plate systems: an anatomical study. J Neurosurg 95, 80–87 (2001) 124. Kelly Jr., D.L., Alexander Jr., E., Davis Jr., C.H., et  al.: Acrylic fixation of atlanto-axial dislocations. Technical note. J Neurosurg 36, 366–371 (1972) 125. Kerschbaumer, F., Kandziora, F., Klein, C., et al.: Transoral decompression, anterior plate fixation, and posterior wire fusion for irreducible atlantoaxial kyphosis in rheumatoid arthritis. Spine (Phila Pa 1976) 25, 2708–2715 (2000) 126. Klekamp, J.W., Ugbo, J.L., Heller, J.G., et  al.: Cervical transfacet versus lateral mass screws: a biomechanical comparison. J Spinal Disord 13, 515–518 (2000) 127. Knoller, S., Jeszenszky, D., Willms, R., et  al.: Transaxial spongiosa-plasty and ventral, temporary atlanto-axial fixation for therapy of dens pseudarthrosis. Z Orthop Ihre Grenzgeb 137, 232–235 (1999) 128. Knopf, W., Vater, H., Meier, U., et al.: Axial osteosynthesis of dens axis fracture as a routine procedure in neurotraumatology. Diagnosis–Surgical-technique–Complications– Follow-up. Zentralbl Neurochir 51, 85–90 (1990) 129. Knöringer, P.: Internal fixation of dens fractures by doublethreaded screws. Orthop Traumatol 4, 231–245 (1992) 130. Kocis, J., Wendsche, P., Visna, P., et al.: Traumatic spondylolisthesis of the axis. Acta Chir Orthop Traumatol Cech 70, 214–218 (2003) 131. Koller, H., Kammermeier, V., Ulbricht, D., et al.: Anterior retropharyngeal fixation C1-2 for stabilization of atlantoaxial instabilities: study of feasibility, technical descrip-

121 tion and preliminary results. Eur Spine J 15, 1326–1338 (2006) 132. Kuroki, H., Rengachary, S.S., Goel, V.K., et al.: Biomechanical comparison of two stabilization techniques of the atlantoaxial joints: transarticular screw fixation versus screw and rod fixation. Neurosurgery 56, 151–159 (2005). discussion 151-159 133. La Marca, F., Zubay, G., Morrison, T., et  al.: Cadaveric study for placement of occipital condyle screws: technique and effects on surrounding anatomic structures. J Neurosurg Spine 9, 347–353 (2008) 134. Lang, J.: The cranio-cervical junction – Anatomy. In: Voth, D., Glees, P. (eds.) Diseases in the cranio-cervical junction. Anatomical and pathological aspects and detailed clinical accounts, pp. 27–61. Gruyter, Berlin, New York (1987) 135. Lapsiwala, S.B., Anderson, P.A., Oza, A., et  al.: Biomechanical comparison of four C1 to C2 rigid fixative techniques: anterior transarticular, posterior transarticular, C1 to C2 pedicle, and C1 to C2 intralaminar screws. Neurosurgery 58, 516–521 (2006). discussion 516–521 136. Latal, J., Lohnert, J.: Treatment of fractures of the upper cervical spine (C1 and C2). Bratisl Lek Listy 94, 483–488 (1993) 137. Leconte, P.: Fracture et luxation des deux premieres vertebres cervicales. In: Judet, R. (ed.) Luxation Congenitale de la Hanche. Fractures du Cou-de-pied Rachis Cervical. Actualites de Chirurgie Orthopedique de l’Hospital Raymond-Poincare, vol. 3, pp. 147–166. Masson et Cie, Paris (1964) 138. Lee, M.J., Cassinelli, E., Riew, K.D.: The feasibility of inserting atlas lateral mass screws via the posterior arch. Spine (Phila Pa 1976) 31, 2798–2801 (2006) 139. Lee, J.H., Jahng, T.A., Chung, C.K.: C1-2 transarticular screw fixation in high-riding vertebral artery: suggestion of new trajectory. J Spinal Disord Tech 20, 499–504 (2007) 140. Lehman Jr., R.A., Sasso, R.C., Helgeson, M.D., et  al.: Accuracy of intraoperative plain radiographs to detect violations of intralaminar screws placed into the C2 vertebrae: a reliability study. Spine (Phila Pa 1976) 32, 3036–3040 (2007) 141. Lesoin, F., Autricque, A., Franz, K., et  al.: Transcervical approach and screw fixation for upper cervical spine pathology. Surg Neurol 27, 459–465 (1987) 142. Levine, A.M., Dacre, A.: Traumatic spondylolisthesis of the axis: hangman’s fracture. In: Clark, C.R., Benzel, E.C., Currier, B.L., et al. (eds.) The cervical spine, pp. 629–650. Lippincott, Philadelphia (2004) 143. Lieberman, I.H., Webb, J.K.: Occipito-cervical fusion using posterior titanium plates. Eur Spine J 7, 308–312 (1998) 144. Lohnert, J., Latal, J.: Fracture of the axis–surgical treatment. II. Axial isthmus. Acta Chir Orthop Traumatol Cech 60, 47–50 (1993) 145. Lowry, D.W., Pollack, I.F., Clyde, B., et al.: Upper cervical spine fusion in the pediatric population. J Neurosurg 87, 671–676 (1997) 146. Lu, J., Ebraheim, N.A., Yang, H., et al.: Anatomic considerations of anterior transarticular screw fixation for atlantoaxial instability. Spine (Phila Pa 1976) 23, 1229–1235 (1998). discussion 1236

122

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

147. Lu, S., Xu, Y.Q., Zhang, Y.Z., et  al.: A novel computerassisted drill guide template for placement of C2 laminar screws. Eur Spine J 18, 1379–1385 (2009) 148. Ma, X.Y., Yin, Q.S., Wu, Z.H., et  al.: C1 pedicle screws versus C1 lateral mass screws: comparisons of pullout strengths and biomechanical stabilities. Spine (Phila Pa 1976) 34, 371–377 (2009) 149. MacKenzie, A.I., Uttley, D., Marsh, H.T., et  al.: Craniocervical stabilization using Luque/Hartshill rectangles. Neurosurgery 26, 32–36 (1990) 150. Madawi, A.A., Casey, A.T., Solanki, G.A., et  al.: Radiological and anatomical evaluation of the atlantoaxial transarticular screw fixation technique. J Neurosurg 86, 961–968 (1997) 151. Madawi, A.A., Solanki, G., Casey, A.T., et al.: Variation of the groove in the axis vertebra for the vertebral artery. Implications for instrumentation. J Bone Joint Surg Br 79, 820–823 (1997) 152. Magerl, F., Seemann, P.S.: Stable posterior fusion of the atlas and axis by transarticular screw fixation. In: Kehr, P., Weidner, A. (eds.) Cervical spine, pp. 322–327. Springer, Wien (1987) 153. Mandel, I.M., Kambach, B.J., Petersilge, C.A., et  al.: Morphologic considerations of C2 isthmus dimensions for the placement of transarticular screws. Spine (Phila Pa 1976) 25, 1542–1547 (2000) 154. Marcotte, P., Dickman, C.A., Sonntag, V.K., et al.: Posterior atlantoaxial facet screw fixation. J Neurosurg 79, 234–237 (1993) 155. Matsunaga, S., Onishi, T., Sakou, T.: Significance of occipitoaxial angle in subaxial lesion after occipitocervical fusion. Spine (Phila Pa 1976) 26, 161–165 (2001) 156. Maughan, P.H., Horn, E.M., Theodore, N., et al.: Avulsion fracture of the foramen magnum treated with occiput-to-c1 fusion: technical case report. Neurosurgery 57, E600 (2005). discussion E600 157. McAfee, P.C., Bohlman, H.H., Ducker, T., et al.: Failure of stabilization of the spine with methylmethacrylate. A retrospective analysis of twenty-four cases. J Bone Joint Surg Am 68, 1145–1157 (1986) 158. McGraw, R.W., Rusch, R.M.: Atlanto-axial arthrodesis. J Bone Joint Surg Br 55, 482–489 (1973) 159. Melcher, R.P., Puttlitz, C.M., Kleinstueck, F.S., et  al.: Biomechanical testing of posterior atlantoaxial fixation techniques. Spine (Phila Pa 1976) 27, 2435–2440 (2002) 160. Menezes, A.H., VanGilder, J.C.: Transoral-transpharyngeal approach to the anterior craniocervical junction. Ten-year experience with 72 patients. J Neurosurg 69, 895–903 (1988) 161. Mixter, S.J., Osgood, R.B.: IV. Traumatic lesions of the atlas and axis. Ann Surg 51, 193–207 (1910) 162. Moftakhar, P., Gonzalez, N.R., Khoo, L.T., et al.: Osseous and vascular anatomical variations within the C1-C2 complex: a radiographical study using computed tomography angiography. Int J Med Robot 4, 158–164 (2008) 163. Moskovich, R., Crockard, H.A.: Atlantoaxial arthrodesis using interlaminar clamps. An improved technique. Spine (Phila Pa 1976) 17, 261–267 (1992) 164. Naderi, S., Crawford, N.R., Song, G.S., et al.: Biomechanical comparison of C1-C2 posterior fixations. Cable, graft, and screw combinations. Spine (Phila Pa 1976) 23, 1946–1955 (1998). discussion 1955–1946

165. Naderi, S., Korman, E., Citak, G., et  al.: Morphometric analysis of human occipital condyle. Clin Neurol Neurosurg 107, 191–199 (2005) 166. Nakanishi, K., Sasaki, T., Tokita, N., et al.: Internal fixation for the odontoid fracture. Orthop Trans 6, 176 (1982) 167. Nassos, J.T., Ghanayem, A.J., Sasso, R.C., et  al.: Biomechanical evaluation of segmental occipitoatlantoaxial stabilization techniques. Spine (Phila Pa 1976) 34, 2740–2744 (2009) 168. Neo, M., Matsushita, M., Iwashita, Y., et al.: Atlantoaxial transarticular screw fixation for a high-riding vertebral artery. Spine (Phila Pa 1976) 28, 666–670 (2003) 169. Newman, P., Sweetnam, R.: Occipito-cervical fusion. An operative technique and its indications. J Bone Joint Surg Br 51, 423–431 (1969) 170. Nockels, R.P., Shaffrey, C.I., Kanter, A.S., et  al.: Occipitocervical fusion with rigid internal fixation: longterm follow-up data in 69 patients. J Neurosurg Spine 7, 117–123 (2007) 171. Nordt, J.C., Stauffer, E.S.: Sequelae of atlantoaxial stabilization in two patients with Down’s syndrome. Spine (Phila Pa 1976) 6, 437–440 (1981) 172. Nottmeier, E.W., Foy, A.B.: Placement of C2 laminar screws using three-dimensional fluoroscopy-based image guidance. Eur Spine J 17, 610–615 (2008) 173. Oda, I., Abumi, K., Sell, L.C., et al.: Biomechanical evaluation of five different occipito-atlanto-axial fixation techniques. Spine (Phila Pa 1976) 24, 2377–2382 (1999) 174. Olerud, S., Olerud, C.: The C1 claw device: a new instrument for C1-C2 fusion. Eur Spine J 10, 345–347 (2001) 175. Pait, T.G., Al-Mefty, O., Boop, F.A., et al.: Inside-outside technique for posterior occipitocervical spine instrumentation and stabilization: preliminary results. J Neurosurg 90, 1–7 (1999) 176. Panjabi, M.M.: The stabilizing system of the spine. Part I. Function, dysfunction, adaptation, and enhancement. J Spinal Disord 5, 383–389 (1992). discussion 397 177. Papagelopoulos, P.J., Currier, B.L., Stone, J., et  al.: Biomechanical evaluation of occipital fixation. J Spinal Disord 13, 336–344 (2000) 178. Paquis, P., Breuil, V., Lonjon, M., et al.: Occipitocervical fixation using hooks and screws for upper cervical instability. Neurosurgery 44, 324–330 (1999). discussion 330–321 179. Paramore, C.G., Dickman, C.A., Sonntag, V.K.: The anatomical suitability of the C1-2 complex for transarticular screw fixation. J Neurosurg 85, 221–224 (1996) 180. Parker, S.L., McGirt, M.J., Garces-Ambrossi, G.L., et al.: Translaminar versus pedicle screw fixation of C2: comparison of surgical morbidity and accuracy of 313 consecutive screws. Neurosurgery 64, 343–348 (2009). discussion 348349 181. Payer, M., Luzi, M., Tessitore, E.: Posterior atlanto-axial fixation with polyaxial C1 lateral mass screws and C2 pars screws. Acta Neurochir (Wien) 151, 223–229 (2009). discussion 229 182. Pointillart, V., Orta, A.L., Freitas, J., et al.: Odontoid fractures. Review of 150 cases and practical application for treatment. Eur Spine J 3, 282–285 (1994) 183. Ransford, A.O., Crockard, H.A., Pozo, J.L., et  al.: Craniocervical instability treated by contoured loop fixation. J Bone Joint Surg Br 68, 173–177 (1986)

References 184. Reindl, R., Sen, M., Aebi, M.: Anterior instrumentation for traumatic C1-C2 instability. Spine (Phila Pa 1976) 28, E329–E333 (2003) 185. Resnick, D.K., Benzel, E.C.: C1-C2 pedicle screw fixation with rigid cantilever beam construct: case report and technical note. Neurosurgery 50, 426–428 (2002) 186. Resnick, D.K., Lapsiwala, S., Trost, G.R.: Anatomic suitability of the C1-C2 complex for pedicle screw fixation. Spine (Phila Pa 1976) 27, 1494–1498 (2002) 187. Rhines, L.D., Fourney, D.R., Siadati, A., et  al.: En bloc resection of multilevel cervical chordoma with C-2 involvement. Case report and description of operative technique. J Neurosurg Spine 2, 199–205 (2005) 188. Richter, M., Schmidt, R., Claes, L., et al.: Posterior atlantoaxial fixation: biomechanical in vitro comparison of six different techniques. Spine (Phila Pa 1976) 27, 1724–1732 (2002) 189. Roberts, D.A., Doherty, B.J., Heggeness, M.H.: Quantitative anatomy of the occiput and the biomechanics of occipital screw fixation. Spine (Phila Pa 1976) 23, 1100–1107 (1998). discussion 1107–1108 190. Robertson, S.C., Menezes, A.H.: Occipital calvarial bone graft in posterior occipitocervical fusion. Spine (Phila Pa 1976) 23, 249–254 (1998). discussion 254–245 191. Rocha, R., Safavi-Abbasi, S., Reis, C., et al.: Working area, safety zones, and angles of approach for posterior C-1 lateral mass screw placement: a quantitative anatomical and morphometric evaluation. J Neurosurg Spine 6, 247–254 (2007) 192. Roy-Camille, R.: Current trends in surgery of the spine. Int Orthop 13, 81–87 (1989) 193. Roy-Camille, R., Saillant, G.: Surgery of the cervical spine. 4. Osteosynthesis of the upper cervical spine. Nouv Presse Med 1, 2847–2849 (1972) 194. Roy-Camille, R., Saillant, G., Laville, C., et al.: Treatment of lower cervical spinal injuries – C3 to C7. Spine (Phila Pa 1976) 17, S442–S446 (1992) 195. Ruf, M., Melcher, R., Harms, J.: Transoral reduction and osteosynthesis C1 as a function-preserving option in the treatment of unstable Jefferson fractures. Spine (Phila Pa 1976) 29, 823–827 (2004) 196. Sandhu, F.A., Pait, T.G., Benzel, E., et al.: Occipitocervical fusion for rheumatoid arthritis using the inside-outside stabilization technique. Spine (Phila Pa 1976) 28, 414–419 (2003) 197. Sar, C., Eralp, L.: Transoral resection and reconstruction for primary osteogenic sarcoma of the second cervical vertebra. Spine (Phila Pa 1976) 26, 1936–1941 (2001) 198. Sasso, R., Doherty, B.J., Crawford, M.J., et  al.: Biomechanics of odontoid fracture fixation. Comparison of the one- and two-screw technique. Spine (Phila Pa 1976) 18, 1950–1953 (1993) 199. Sasso, R.C., Jeanneret, B., Fischer, K., et al.: Occipitocervical fusion with posterior plate and screw instrumentation. A long-term follow-up study. Spine (Phila Pa 1976) 19, 2364–2368 (1994) 200. Sciubba, D.M., Noggle, J.C., Vellimana, A.K., et  al.: Laminar screw fixation of the axis. J Neurosurg Spine 8, 327–334 (2008) 201. Sciubba, D.M., Noggle, J.C., Vellimana, A.K., et  al.: Radiographic and clinical evaluation of free-hand placement of C-2 pedicle screws. Clinical article. J Neurosurg Spine 11, 15–22 (2009)

123 202. Sen, M.K., Steffen, T., Beckman, L., et  al.: Atlantoaxial fusion using anterior transarticular screw fixation of C1-C2: technical innovation and biomechanical study. Eur Spine J 14, 512–518 (2005) 203. Senoglu, M., Ozbag, D., Gumusalan, Y.: C2 intralaminar screw placement: a quantitative anatomical and morphometric evaluation. Turk Neurosurg 19, 245–248 (2009) 204. Shalayev, S.G., Mun, I.K., Mallek, G.S., et al.: Retrospective analysis and modifications of retractor systems for anterior odontoid screw fixation. Neurosurgical Focus 16, 1–4 (2004) 205. Simmons, E.H., du Toit Jr., G.: Lateral atlantoaxial arthrodesis. Orthop Clin North Am 9, 1101–1114 (1978) 206. Six, E., Kelly Jr., D.L.: Technique for C-1, C-2, and C-3 fixation in cases of odontoid fracture. Neurosurgery 8, 374–377 (1981) 207. Smith, M.D., Anderson, P., Grady, M.S.: Occipitocervical arthrodesis using contoured plate fixation. An early report on a versatile fixation technique. Spine (Phila Pa 1976) 18, 1984–1990 (1993) 208. Solanki, G.A., Crockard, H.A.: Peroperative determination of safe superior transarticular screw trajectory through the lateral mass. Spine (Phila Pa 1976) 24, 477–1482 (1999) 209. Sonntag, V.K., Dickman, C.A.: Craniocervical stabilization. Clin Neurosurg 40, 243–272 (1993) 210. Statham, P., O’Sullivan, M., Russell, T.: The Halifax Interlaminar Clamp for posterior cervical fusion: initial experience in the United Kingdom. Neurosurgery 32, 396–398 (1993). discussion 398–399 211. Stevenson, K.L., Wetzel, M., Pollack, I.F.: Delayed intracranial migration of cervical sublaminar and interspinous wires and subsequent cerebellar abscess. Case report. J Neurosurg 97, 113–117 (2002) 212. Stillerman, C.B., Wilson, J.A.: Atlanto-axial stabilization with posterior transarticular screw fixation: technical description and report of 22 cases. Neurosurgery 32, 948–954 (1993) 213. Stokes, J.K., Villavicencio, A.T., Liu, P.C., et al.: Posterior atlantoaxial stabilization: new alternative to C1-2 transarticular screws. Neurosurg Focus 12, E6 (2002) 214. Stulík, J., Krbec, M.: Magerl’s Technique of C1-2 Fixation. Acta Chir Orthop Traumatol Cech 67, 93–99 (2000) 215. Stulik, J., Vyskocil, T., Sebesta, P., et al.: Atlantoaxial fixation using the polyaxial screw-rod system. Eur Spine J 16, 479–484 (2007) 216. Subach, B.R., Morone, M.A., Haid Jr., R.W., et  al.: Management of acute odontoid fractures with single-screw anterior fixation. Neurosurgery 45, 812–819 (1999) 217. Suchomel, P., Buchvald, P., Barsa, P., et  al.: Single-stage total C-2 intralesional spondylectomy for chordoma with three-column reconstruction. Technical note. J Neurosurg Spine 6, 611–618 (2007) 218. Suchomel, P., Hradil, J., Barsa, P., et al.: Surgical treatment of fracture of the ring of axis – “hangman’s fracture”. Acta Chir Orthop Traumatol Cech 73, 321–328 (2006) 219. Suchomel, P., Stulik, J., Klezl, Z., et al.: Transarticular fixation of C1-C2: a multicenter retrospective study. Acta Chir Orthop Traumatol Cech 71, 6–12 (2004) 220. Suchomel, P., Taller, S., Lukas, R., et al.: Surgical treatment of fractures of the odontoid process. Rozhl Chir 79, ­301–308 (2000) 221. Sudo, H., Abumi, K., Ito, M., et al.: Spinal cord compression by multistrand cables after solid posterior atlantoaxial

124

6  Specific Reconstruction Techniques of Upper Cervical Spine and Craniovertebral Junction

fusion. Report of three cases. J Neurosurg 97, 359–361 (2002) 222. Taitz, C., Arensburg, B.: Erosion of the foramen transversarium of the axis. Anatomical observations. Acta Anat (Basel) 134, 12–17 (1989) 223. Taitz, C., Arensburg, B.: Vertebral artery tortuosity with concomitant erosion of the foramen of the transverse process of the axis. Possible clinical implications. Acta Anat (Basel) 141, 104–108 (1991) 224. Taller, S., Suchomel, P., Lukas, R., et al.: CT-guided internal fixation of a hangman’s fracture. Eur Spine J 9, 393–397 (2000) 225. Tan, M., Wang, H., Wang, Y., et al.: Morphometric evaluation of screw fixation in atlas via posterior arch and lateral mass. Spine (Phila Pa 1976) 28, 888–895 (2003) 226. Theodore, N., Partovi, S., Walker, M.P., et al.: Preoperative helical CT angiography for C1-2 transarticular screw placement. A new technique. BNI Q 17, 49–52 (20001) 227. Tucker, H.H.: Technical report: method of fixation of subluxed or dislocated cervical spine below C1-C2. Can J Neurol Sci 2, 381–382 (1975) 228. Uribe, J.S., Ramos, E., Baaj, A., et al.: Occipital cervical stabilization using occipital condyles for cranial fixation: technical case report. Neurosurgery 65, E1216–E1217 (2009) 229. Uribe, J.S., Ramos, E., Vale, F.: Feasibility of occipital condyle screw placement for occipitocervical fixation: a cadaveric study and description of a novel technique. J Spinal Disord Tech 21, 540–546 (2008) 230. Vaccaro, A.R., Lehman, A.P., Ahlgren, B.D., et al.: Anterior C1-C2 screw fixation and bony fusion through an anterior retropharyngeal approach. Orthopedics 22, 1165–1170 (1999) 231. Vlach, O., Leznar, M., Bayer, M.: Diagnosis, classification and treatment of so-called hangman’s fractures. Acta Chir Orthop Traumatol Cech 55, 456–466 (1988) 232. Wang, M.Y.: C2 crossing laminar screws: cadaveric morphometric analysis. Neurosurgery 59, ONS84–ONS88 (2006). discussion ONS84–88 233. Wang, M.Y., Samudrala, S.: Cadaveric morphometric analysis for atlantal lateral mass screw placement. Neurosurgery 54, 1436–1439 (2004). discussion 1439–1440 234. Weidner, A., Wahler, M., Chiu, S.T., et al.: Modification of C1-C2 transarticular screw fixation by image-guided surgery. Spine (Phila Pa 1976) 25, 2668–2673 (2000). discussion 2674 235. Wertheim, S.B., Bohlman, H.H.: Occipitocervical fusion. Indications, technique, and long-term results in thirteen patients. J Bone Joint Surg Am 69, 833–836 (1987) 236. Wickbom, G.I., Williamson, M.R.: Anomalous foramen transversarium of C2 simulating erosion of bone. Neuroradiology 19, 43–45 (1980) 237. Wilke, H.J., Fischer, K., Kugler, A., et al.: In vitro investigations of internal fixation systems of the upper cervical

spine. II. Stability of posterior atlanto-axial fixation techniques. Eur Spine J 1, 191–199 (1992) 238. Wright, N.M.: Posterior C2 fixation using bilateral, crossing C2 laminar screws: case series and technical note. J Spinal Disord Tech 17, 158–162 (2004) 239. Wright, N.M.: Translaminar rigid screw fixation of the axis. Technical note. J Neurosurg Spine 3, 409–414 (2005) 240. Wright, N.M., Lauryssen, C.: Vertebral artery injury in C1-2 transarticular screw fixation: results of a survey of the AANS/CNS section on disorders of the spine and peripheral nerves. American Association of Neurological Surgeons/Congress of Neurological Surgeons. J Neurosurg 88, 634–640 (1998) 241. Xiao, Z.M., Zhan, X.L., de Gong, F., et al.: C2 pedicle screw and plate combined with C1 titanium cable fixation for the treatment of atlantoaxial instability not suitable for placement of C1 screw. J Spinal Disord Tech 21, 514–517 (2008) 242. Xu, R., Burgar, A., Ebraheim, N.A., et al.: The quantitative anatomy of the laminas of the spine. Spine (Phila Pa 1976) 24, 107–113 (1999) 243. Xu, R., Nadaud, M.C., Ebraheim, N.A., et al.: Morphology of the second cervical vertebra and the posterior projection of the C2 pedicle axis. Spine (Phila Pa 1976) 20, 259–263 (1995) 244. Yamazaki, M., Koda, M., Aramomi, M.A., et al.: Anomalous vertebral artery at the extraosseous and intraosseous regions of the craniovertebral junction: analysis by three-dimensional computed tomography angiography. Spine (Phila Pa 1976) 30, 2452–2457 (2005) 245. Yan, W., Zhang, C., Zhou, X., et al.: Safe angle scope for posterior atlanto-occipital transarticular screw fixation. Neurosurgery 65, 499–504 (2009). discussion 504 246. Yan, W.J., Zhou, X., Zhang, Y., et al.: The feasibility study on the posterior occipito-atlanto-axial screw fixation. J Spinal Surg 2, 289–293 (2004) 247. Yin, Q., Ai, F., Zhang, K., et al.: Irreducible anterior atlantoaxial dislocation: one-stage treatment with a transoral atlantoaxial reduction plate fixation and fusion. Report of 5 cases and review of the literature. Spine (Phila Pa 1976) 30, E375–E381 (2005) 248. Yoshida, M., Neo, M., Fujibayashi, S., et al.: Comparison of the anatomical risk for vertebral artery injury associated with the C2-pedicle screw and atlantoaxial transarticular screw. Spine 31, E513–E517 (2006) 249. Young, J.P., Young, P.H., Ackermann, M.J., et al.: The ponticulus posticus: implications for screw insertion into the first cervical lateral mass. J Bone Joint Surg Am 87, 2495–2498 (2005) 250. Zipnick, R.I., Merola, A.A., Gorup, J., et al.: Occipital morphology. An anatomic guide to internal fixation. Spine (Phila Pa 1976) 21, 1719–1724 (1996). discussion 1729–1730 251. Zoma, A., Sturrock, R.D., Fisher, W.D., et al.: Surgical stabilisation of the rheumatoid cervical spine. A review of indications and results. J Bone Joint Surg Br 69, 8–12 (1987)

7

Virtual and Real Time Navigational Techniques P. Suchomel and O. Choutka

All new technology should enable physicians to perform previously undoable tasks or solve unsolved problems, rather than dominating over reason as once elegantly stated by Scott Boden [60]. This certainly applies to spinal surgery and navigational techniques that have developed over the past decade. Such development opened the field of spinal surgery also to less experienced surgeons by improving accuracy and safety. Image-guided spinal surgery (IGS) is certainly a good addition to the armamentarium of a spine surgeon but, by any means, does not replace experience. Due to fairly unique anatomy of the region, experience and sophisticated reconstructive techniques are usually necessary for cases involving the upper cervical spine (UCS) and craniovertebral junction (CVJ). The literature is full of descriptions of various surgical techniques for treating pathologies of the CVJ and UCS [14, 29, 35, 46–51, 59]. Image-guided techniques, therefore, do not necessarily further enhance those solutions. However, there are certain types of operations where the extent of pathology is not directly visible, the resection volume of pathological tissue cannot be predicted, or the implant trajectory is out of sight of direct vision. Here, although useful, simple fluoroscopy can be rather limited and therefore, image-guided techniques have the potential to enhance the safety of such procedures.

Spatial orientation is an important asset for reaching deep-seated lesions in the brain. Similarly, in CVJ, it is essential to understand not only the target but also vital structures en route to the target as they are at risk of injury. Frame stereotaxy [11] and frameless navigational systems [9, 37] proved to be rather useful in this task. Cranial navigation is simplified due to the relatively stable relationship of the brain to the skull (Fig. 7.1). Similar techniques in the spine did succeed in the past [11]; however, the mobility of this structure called for a different approach to stereotaxy. The improvement of modern imaging techniques (CT, MRI, and isocentric fluoroscopy) together with improved computer software, allowed for significantly better, three-dimensional thinking in spinal surgery [5, 15, 45, 54]. The selection of an actual type of image guidance tool is dependent on the goal of the procedure. Different tasks have different needs. For example, tumor resection requires good localization and border demarcation to continuously monitor the extent of resection. On the other hand, a spinal fracture with mobile fragments and deformity can only be registered preoperatively but is inaccurate after any reduction, unless re-registered. It is fairly obvious that technical errors during instrumentation of UCS and CVJ can result in neurological, vascular, or mechanical complications.

P. Suchomel Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic

7.1 Technique Description

O. Choutka Department of Neurosurgery, University of Cincinnati College of Medicine, 231 Albert Sabin Way, Cincinnati, OH 45267-0515, USA

Surgeons new to navigation techniques must realize that there are two main, relatively different, image guidance options. There are techniques utilizing onetime image acquisition (pre- or intraoperative) with subsequent navigation based on the initial registration

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_7, © Springer-Verlag Berlin Heidelberg 2011

125

126

Fig.  7.1  Large clival chordoma approached transmaxillar and removed with the help of cranial frameless navigation. (a) Preoperative MRI showing the extent of tumor. (b) Surface

7  Virtual and Real Time Navigational Techniques

registration for virtual navigation. (c) Preoperative picture depicting the transmaxillar approach. (d) Postoperative MRI showing no residual tumor

127

7.1  Technique Description

(virtual image guidance) and those that utilize intraoperative real-time image techniques that do not require constant re-registration after any data shift (e.g., positioning, tumor removal, reduction of malalignment).

7.1.1 Virtual Image-Guided Surgery (vIGS) Foley and Smith introduced navigation to spine surgery in 1994 after a positive cranial experience [9, 20]. Lack of reliable external surface markers was an initial concern; however, this issue was solved with the registration of dorsal spine bony anatomic landmarks in association with dynamic reference array (DRA) directly attached to the target vertebra [9, 18, 19, 37]. Initially, vIGS was mostly used for

a

lumbar pedicle screw placement [8, 9, 31, 36, 37]. Later, thoracic spine was added to the lumbar and it quickly became obvious that the technique can result in higher accuracy than traditional methods [12, 31, 40]. Now, navigational techniques have been applied to slightly more delicate structures of the UCS [1, 5, 15, 57, 58]. Basically, anatomical data are collected by thinsliced (1–1.5  mm) CT, MRI, or isofluoroscopy and transmitted to a computer workstation. DRA is connected to a solid part of the target area so that it does not obstruct the operative field. Navigated vertebra and instruments being tracked are then registered by the electro-optical camera connected to the computer (Fig. 7.2). Most systems use either passive arrays (reflective spheres) or active arrays (light-emitting diodes) for spatial orientation of the tracked instruments. The virtual image is then displayed on the computer screen and

c

b

Fig. 7.2  Set up of surgical theater prepared for virtual C1-2 posterior transarticular screw image guided surgery (preoperative data acquisition). (a) Navigational computer screen, tracking

optic camera, screen of fluoroscope. (b) Active DRA fixed to spinous process of C2, passive DRA probe C2 surface registration. (c) Image guided drill connected to active DRA

128

7  Virtual and Real TimeNavigational Techniques

positioning of the patient or even after spine exposure but prior to placement of any hardware. One has to be aware of possible movement or intersegmental relationship change after data acquisition and appropriately re-register or use automatic updates. This is more of a potential problem with systems that rely on preoperative rather than intraoperative scan. A comprehensive data acquisition and computer analysis allows essentially for any intervention to be virtually planned (Fig. 7.4). In terms of UCS and CVJ, this usually applies to determination whether a certain screw is anatomically feasible, the location and course of VA, the best tumor approach, and clival anatomy. Virtual planning can determine whether a complex procedure is feasible and can even be utilized in procedures where further intraoperative navigation is not possible or necessary [26, 33, 34].

7.1.1.1 Preoperative Imaging Based vIGS

Fig. 7.3  Comparison of virtual plan with real drill position visible on navigational computer screen

the procedure follows a preoperative virtual plan and can be checked directly on the screen (Fig. 7.3). Images for navigation systems are currently obtained prior to surgery or, more frequently, after

CT-based preoperative vIGS is the most accurate method of spinal bone navigation, especially when it comes to deformity and anatomically difficult regions. With accurate registration (1.5 mm accuracy), this guidance can be utilized for any screw placement in the UCS and CVJ. Its most useful application is the placement of transarticular C1/2 screw [1, 5, 15, 17, 56] due to the lack of direct visualization of anatomical structures and its accuracy is certainly better than with traditional methods [12, 27, 32, 53]. The downside of preoperative CT-based navigation is the need for a specific scanning

Fig. 7.4  Virtual plan for transarticular screw introduction. (a) Left side. (b) Right side (notice high riding VA)

129

7.1  Technique Description

protocol, which obviously increases the cost in cases where a CT has already been obtained but is incompatible with the navigation system. Registration process is lengthy, has a steep learning curve and each individual vertebra requires separate registration. Usually, only C2 vertebra is registered/guided and C1-reduced position is confirmed by lateral fluoroscopy.

7.1.1.2 Intraoperative Imaging Based vIGS Intraoperative use of CT scanner has been well described. Navigation system registers images obtained after patient positioning and spinal exposure and thus minimizing intersegmental movement that can occur with preoperatively obtained data [20]. It also allows for immediate repeat intraoperative scan to check hardware position. However, the cost and common small gentry’s window of a mobile scanner are the major downsides of this technique. Other drawbacks include the need for a special table, draping technique and the fact that, for example, mobile fracture fragments cannot be seen in a changed position without new registration. Virtual two-dimensional fluoroscopy is perhaps better known to most spine surgeons as it combines the use of familiar fluoroscopy and image guidance techniques [4, 20, 37]. Although its accuracy and virtual guidance is only improved in one plane at a time, application to UCS and CVJ surgery has not been described. Nevertheless, when compared to classic fluoroscopy, the radiation exposure is decreased. This type of navigation depends on regular fluoroscope image acquisition and is, therefore, problematic in obese or osteopenic patients and in poorly visualized regions of the spine. UCS anatomy is not always visible, in particular, in patients with deformity and certainly lacks the necessary detail. Therefore, surgeons have adopted three-dimensional fluoroscopy with the use of isocentric C-arm that automatically rotates around the patient, obtaining fluoroscopic images in the surgical position with the spine centered [21, 22]. The iso-C then generates axial, sagittal, and coronal images of the anatomy that are close to CT quality. The iso-C arm can be connected to a navigation station and images are acquired after the patient is positioned on the table prior to or after the exposure. Spinal exposure prior to registration is not necessary thus obviating the surgeon-driven registration process completely. This fact is important for minimally invasive or even percutaneous spinal

procedures [12, 21]. The accuracy in UCS and CVJ has been established in the initial series using iso-C navigational techniques and is frequently good enough [22, 23, 41, 42]. Axial images can be reconstructed three-dimensionally and sent to the navigation system. A real-time position of spinal elements and implants can be transmitted to the computer and any subsequent navigation is actualized and more accurate. Also, the final position of hardware can be monitored at the end of the procedure [22]. The radiation dose is reduced to 57–77% [22] in comparison with standard CT protocols. Images obtained with iso-C arm are obviously of lesser quality than CT and the volume of the scan is limited to 12 cm3, which means visualization of only four cervical or three lumbar vertebrae. Intraoperative spine image acquisition with MRI became possible with open design of the scanner [44]. This has found application mostly in intracranial surgery of intra-axial lesions to define extent of resection. Similarly, resection of intramedullary spine tumors can be controlled with MRI-based navigation techniques; however, bony structures are poorly defined by this modality [28]. The most limiting nature of an intraoperative MRI is its cost, size, and need for nonmagnetic equipment.

7.1.2 “Real Time” Image Guided Surgery (rIGS) Ideally, “real time” image-guided surgery would allow for a continuous check of the extent of an intervention. Although a real-time feedback in its true sense of a word does not currently exist, the real-time techniques available allow for an immediate check of each step of the procedure. It is safer than any virtual navigation. Surgical strategy can be adapted based on changes during the procedure or target movement. Any dynamic process in surgery (e.g., resection of tumor, extent of decompression, deformity reduction, and/or fracture fragment reposition) will be automatically reflected on subsequent intraoperative scan. The drill/screw angle or length can be modified and visualized. Subsequent steps of surgery minimize the risk to vascular or neural structures as long as frequent updated scans take place. The longer the time interval between “real time” updates is, the higher the risk of inadvertent event. Minimally invasive or even percutaneous spine surgery is clearly safer due to real-time imaging guidance.

130

However, the price for improved safety and accuracy is paid in time necessary for repeated scans, need for a good radiographer or radiologist, higher radiation exposure, need for special table hardware connection for CT or MRI, and surgeons’ discomfort caused by ergonomic problems. The most ergonomic real-time navigational tool is probably three-dimensional isofluoroscopy [22, 23]. The machine is smaller and thus offers some flexibility in the management of operating room and the procedure. It fares favorably against CT or MRI in terms of image acquisition times [24, 40]. Bony structures are well visualized on iso-C arm and it is, therefore, a good tool for delicate procedures of the UCS and CVJ (e.g., transpedicular or transarticular C2 screw placement, odontoid-compressive osteosynthesis or C1–0 screw) [22, 23]. Real-time navigation can also be combined with a virtual one whenever necessary. A good example is the placement of an odontoid-compressive screw that can be guided Iso-C arm and also monitored by lateral fluoroscopy instead of using biplanar fluoroscopy [22]. Three-dimensional isofluoroscopy does account for motion of bone fragments or deformity corrections in real time, which is one of its main advantages. Image quality in obese patients as well as increased radiation exposure remains an issue. The use of direct spinal CT guidance in stationary scanners in radiology suites is well known from percutaneous interventions [3, 43]. Alternatively, it was limited to procedures performed in stationary scanners in radiology departments and suites adapted for use of general anesthesia and surgery [10, 54]. Mobile CT scanners with radiolucent surgical tables then allowed for real-time navigational surgery to move to real operating rooms for larger, open procedures [10, 25]. The superior image quality of CT when compared to isofluoroscopy resulted in broader range of intraoperative applications of this modality and also enabled percutaneous procedures at the UCS and CVJ. As with any intraoperative scanner, the major problems are of ergonomic nature due to the gantry size and need for a special operating table. Likewise, intraoperative MRI (iMRI) surgical systems have been extensively described in the literature [28, 60]. Soft tissue pathologies (extra/intradural tumors, intramedullary tumors, or CVJ anomalies compressing the neural tissue) can be imaged very well with iMRI. Again, the advantages are clear and drawbacks were already mentioned above. The widespread

7  Virtual and Real TimeNavigational Techniques

use of iMRI in spine surgery is prohibited by the low yield of high start-up costs, poor bone visualization, significant metal artifact, and low MRI image quality.

7.2 Our Preference We advocate the use of computer guidance in the surgery of UCS and CVJ, in particular in situations when safety and accuracy can be significantly enhanced. This is especially important when navigating invisible areas. Neural compression can be caused by developmental or acquired deformity, trauma, inflammation or tumor. The use of virtual or real-time CT or MRI to localize or monitor the extent of decompression is well described [55, 57, 58]. Whenever pure bony compression exists, the Iso-C arm may be sufficient. Reconstructive techniques during transoral procedures may require the use of navigational techniques for planning and guiding anchorage of construct to the clivus (Fig.  6.6, Chap. 6; Fig.  19.22, Chap. 19). The position of hypoglossal canal can also be monitored in cases of atlanto-occipital instrumentation when screws pass through the condyle [23]. Although, the anatomy of atlas is relatively simple and rarely requires navigational techniques, this technology facilitated percutaneous fixation of an unusual C1 fracture in our department. The fracture comprised a unilateral sagittal split of the lateral mass with an intact transverse ligament (Figs. 7.5–7.8). The reason for surgical treatment in our situation was the intraarticular nature of the fracture and we believe such fractures result in subsequent pain syndrome, deformity, and poor healing, despite the advocated use of conservative treatment only. Indeed, Bransford et al. [7] later reported on three out of six patients with similar injuries that were initially treated with external orthosis developing a late cockrobin deformity with significant pain and rotatory restriction. All three patients were eventually successfully treated with an occipitocervical fusion that might have been avoided with an initial aggressive treatment. Image guidance can be applied to just about any type of a C2 screw. Anterior approaches are complicated by the lack of a stable DRA attachment to the navigated vertebra but IGS is not impossible as demonstrated in placement of odontoid screws with iso-C navigation [23, 52]. Deformity correction and fracture dislocations can make posterior vIGS difficult. Any mobile part of spine

7.2  Our Preference

131

Fig. 7.5  Unstable two part C1 ring fracture with intact TAL. (a) Axial CT image showing intraarticular extend of fracture. (b) CT reconstruction in coronal plane. (c) MRI confirming the intact TAL

Fig. 7.6  First step of “real time” percutaneous CT guided osteosynthesis of distracted C1 lateral mass fracture (the same patient from Fig. 7.5). (a) Planning of introductory angle. (b–d) Consecutive “step by step” introduction of K-wire through the fracture. Each step controlled by repeated CT image

or fracture fragment cannot be registered preoperatively; perioperative data update is time-consuming and it only depicts one actual position of target in space. This becomes obvious when reducing hangman’s fracture posteriorly. This issue was overcome by real-time CT-based guidance [54]. Judet’s transpedicular screw

compressive osteosynthesis [2, 29, 41, 42] can effectively be performed in patients with some Effendi type II fractures without disk bulge or, more appropriately, in Effendi type I injuries, where the fracture gap is larger than 3 mm on the CT scan (see Chap. 12). The UCS is approached through a standard midline approach and

132 Fig. 7.7  Percutaneous drilling and cannulated screw purchase along the K-wire. (a) Cannulated drill introduced along the K-wire. (b) Cannulated screw driver in position. (c, d) Consecutive cannulated lag screw passage

7  Virtual and Real TimeNavigational Techniques

a

b

c

d

Fig. 7.8  Final tightening of the lag screw compressing the fracture (result after 1.5 year on Fig. 10.13, Chap. 10). (a) Before tightening. (b) Final fracture compression

entry points are planned according to the navigation computer optimal trajectory. The gantry of a scanner is accordingly to maximize screw visualization as it passes through the axis. Repeated CT scans monitor the stepby-step gradual screw introduction. The fracture is finally compressed by tightening of the lag screw; the length of which is chosen based on navigation images

(Fig. 7.9; Fig. 12.16, Chap. 12). We believe that, soon many of those procedures will be achievable percutaneously thanks to the development of mobile scanners, Iso-C arms, and better computer software. Currently, we use two main surgical techniques for C1–2 fixation: atlantoaxial screw fixation as described by Magerl in 1987 [35] and the Harms [16] modification

133

7.2  Our Preference Fig. 7.9  Open surgical compression of hangman type I fracture with the use of “real time” CT navigation. (a) Before tightening of the lag screw. (b) After final tightening

of Goel’s [14] method of screw and rod construct between C1 lateral mass and the C2 pedicles/isthmi (see Chap. 6). Because both techniques involve a passage of the screws through the axis and may potentially result in inadvertent injuries of neural or vascular structures (i.e., spinal cord and vertebral artery), it would appear very reasonable to use surgical navigation in such cases. We use a CT-based vIGS where images are obtained by a CT scanner according to a specific protocol and transferred to the workstation. We focus the scanner on the C2 vertebra mainly in order to create its large

a

Fig. 7.10  Navigational plan (a) and postoperative axial CT scan (b) confirming correct position of crosslaminary (with help of vIGS) introduced screws in thin C2 laminas

and precise 3D model. On the preoperative virtual plan, we determine whether the isthmus is large enough to accommodate a 3.5 mm screw. Once the feasibility is confirmed virtually, we proceed to define clearly visible anatomical feducial points for registration purposes. The patient is then positioned prone, standard posterior exposure carried out, and a DRA frame is firmly attached to the C2 spinous process as not to hinder the procedure (Fig.  7.2b). Registration then takes place and instruments are tracked. The virtual picture should always be checked against the visible anatomy in order

134

7  Virtual and Real TimeNavigational Techniques

b

real-time data acquisition and thus enable the growth of minimally invasive or even percutaneous [6, 12, 21, 50] and robotic [38] surgeries of the UCS and CVJ.

References

Fig. 7.10  (continued)

to avoid any registration mistakes as a result of vertebral movement (e.g., after drilling of the first screw hole). If atlas dislocation requires open reduction, we prefer to do this by C1–2 wire fixation or C2 traction. Performing this step first, obviously, not only reduces the dislocation but also limits any movement of atlas during preparation of the transarticular screw hole (e.g., tapping). Any atlantal movement could disrupt the continuity of a screw path and make screw passage through the joint difficult. Navigation is then used to mark the appropriate entry points and predict the best possible screw trajectory. Tracked instruments are then used to complete the instrumentation safely (Fig. 7.2c). We usually confirm the accuracy of a navigated procedure with lateral fluoroscopy intraoperatively and then again with a CT scan on the first postoperative day (if CT not used during the navigation) (Fig. 7.10). Although, many anatomical studies raise a concern of VA injury during instrumentation of UCS [30, 34, 39], navigational systems were not used in the largest published series of Harms/Goel method of C1–2 fixation [13, 16]. Virtual or real-time image guidance is indeed helpful during atlantoaxial fixation, especially when dealing with thin pedicles identified during preoperative evaluation. We believe that three-dimensional measurements and optimal pedicle screw trajectory planning should be done whenever this surgery is being considered. No navigational technology can substitute surgeon’s planning; however, safety of the instrumentation of the UCS and CVJ can be significantly enhanced by the new technology. In the near future, image-guided application process will become faster with improved

  1. Acosta Jr., F.L., Quinones-Hinojosa, A., Gadkary, C.A., et  al.: Frameless stereotactic image-guided C1-C2 transarticular screw fixation for atlantoaxial instability: review of 20 patients. J Spinal Disord Tech 18, 385–391 (2005)   2. Arand, M., Hartwig, E., Kinzl, L., et al.: Spinal navigation in cervical fractures – a preliminary clinical study on Judetosteosynthesis of the axis. Comput Aided Surg 6, 170–175 (2001)   3. Barsa, P., Suchomel, P., Lukas, R., et  al.: Percutaneous CT-guided radiofrequency ablation in spinal osteoid osteoma treatment. Acta Chir Orthop Traumatol Cech 74, 401–405 (2007)   4. Battaglia, T.C., Tannoury, T., Crowl, A.C.: A cadaveric study comparing standard fluoroscopy with fluoroscopy-based computer navigation for screw fixation of the odontoid. J Surg Orthop Adv 14, 175–180 (2005)   5. Bolger, C.: Preliminary experience with computer assisted surgery for C1/C2 transarticular screw placement. Computer Assisted Orthopedic Surgery, 4th international symposium, Davos, 17–19 March 1999, p. S25.   6. Borm, W., Konig, R.W., Albrecht, A., et  al.: Percutaneous transarticular atlantoaxial screw fixation using a cannulated screw system and image guidance. Minim Invasive Neurosurg 47, 111–114 (2004)   7. Bransford, R., Falicov, A., Nguyen, Q., et al.: Unilateral C-1 lateral mass sagittal split fracture: an unstable Jefferson fracture variant. J Neurosurg Spine 10, 466–473 (2009)   8. Braun, V., Rath, S.A., Antoniadis, G., et al.: In vivo experiences with frameless stereotactically guided screw placement in the spine – results from 75 consecutive cases. Neurosurg Rev 24, 74–79 (2001)   9. Foley, K.T., Smith, M.M.: Image-guided spine surgery. Neurosurg Clin N Am 7, 171–186 (1996) 10. Fritz, H.G., Kuehn, D., Haberland, N., et  al.: Anesthesia management for spine surgery using spinal navigation in combination with computed tomography. Anesth Analg 97, 863–866 (2003) 11. Gabriel, E.M., Nashold Jr., B.S.: History of spinal cord stereotaxy. J Neurosurg 85, 725–731 (1996) 12. Gebhard, F., Weidner, A., Liener, U.C.: Navigation at the spine. Injury 35(Suppl 1), S-A35–S-A45 (2004) 13. Goel, A., Desai, K.I., Muzumdar, D.P., et  al.: Atlantoaxial fixation using plate and screw method: a report of 160 treated patients. Neurosurgery 51, 1351–1356 (2002). discussion 1356–1357 14. Goel, A., Laheri, V.: Plate and screw fixation for atlanto-axial subluxation. Acta Neurochir (Wien) 129, 47–53 (1994) 15. Goffin, J., Van Brussel, K., Martens, K.: Three-dimensional computed tomography-based, personalized drill guide for posterior cervical stabilization at C1-C2. Spine (Phila Pa 1976) 26, 1343–1347 (2001)

References 16. Harms, J., Melcher, R.P.: Posterior C1-C2 fusion with polyaxial screw and rod fixation. Spine (Phila Pa 1976) 26, 2467–2471 (2001) 17. Herz, T., Franz, A., Giacomuzzi, S.M., et al.: Accuracy of spinal navigation for magerl screws. Clin Orthop Relat Res 409, 124–130 (2003) 18. Holly, L.T.: Image-guided spinal surgery. Int J Med Robot 2, 7–15 (2006) 19. Holly, L.T., Bloch, O., Johnson, J.P.: Evaluation of registration techniques for spinal image guidance. J Neurosurg Spine 4, 323–328 (2006) 20. Holly, L.T., Foley, K.T.: Intraoperative spinal navigation. Spine (Phila Pa 1976) 28, S54–S61 (2003) 21. Holly, L.T., Foley, K.T.: Percutaneous placement of posterior cervical screws using three-dimensional fluoroscopy. Spine (Phila Pa 1976) 31, 536–540 (2006). discussion 541 22. Hott, J.S., Deshmukh, V.R., Klopfenstein, J.D.: Intraoperative Iso-C C-arm navigation in craniospinal surgery: the first 60 cases. Neurosurgery 54, 1131–1136 (2004). discussion 1136-1137 23. Hott, J.S., Papadopoulos, S.M., Theodore, N.: Intraoperative Iso-C C-arm navigation in cervical spinal surgery: review of the first 52 cases, 29th edn, pp. 2856–2860. Spine, Phila Pa 1976 (2004) 24. Hufner, T., Gebhard, F., Grutzner, P.A.: Which navigation when? Injury 35(Suppl 1), S-A30–S-A34 (2004) 25. Hum, B., Feigenbaum, F., Cleary, K.: Intraoperative computed tomography for complex craniocervical operations and spinal tumor resections. Neurosurgery 47, 374–380 (2000). discussion 380-371 26. Igarashi, T., Kikuchi, S., Sato, K., et al.: Anatomic study of the axis for surgical planning of transarticular screw fixation. Clin Orthop Relat Res 408, 162–166 (2003) 27. Ito, H., Neo, M., Yoshida, M., et al.: Efficacy of computerassisted pedicle screw insertion for cervical instability in RA patients. Rheumatol Int 27, 567–574 (2007) 28. Jolesz, F.A.: Future perspectives for intraoperative MRI. Neurosurg Clin N Am 16, 201–213 (2005) 29. Judet, R., Roy-Camille, R., Saillant, G.: Fractures du raches cervical. Actualités de chirurgie orthopédique de l’hospital Raymond-Poincaré 8, 174–175 (1970) 30. Kazan, S., Yildirim, F., Sindel, M., et al.: Anatomical evaluation of the groove for the vertebral artery in the axis vertebrae for atlanto-axial transarticular screw fixation technique. Clin Anat 13, 237–243 (2000) 31. Kosmopoulos, V., Schizas, C.: Pedicle screw placement accuracy: a meta-analysis. Spine (Phila Pa 1976) 32, E111–E120 (2007) 32. Kotani, Y., Abumi, K., Ito, M., et  al.: Improved accuracy of  computer-assisted cervical pedicle screw insertion. J Neurosurg 99, 257–263 (2003) 33. Lee, J.H., Jahng, T.A., Chung, C.K.: C1-2 transarticular screw fixation in high-riding vertebral artery: suggestion of new trajectory. J Spinal Disord Tech 20, 499–504 (2007) 34. Madawi, A.A., Solanki, G., Casey, A.T., et al.: Variation of the groove in the axis vertebra for the vertebral artery. Implications for instrumentation. J Bone Joint Surg Br 79, 820–823 (1997) 35. Magerl, F., Seemann, P.S.: Stable posterior fusion of the atlas and axis by transarticular screw fixation. In: Kehr, P., Weidner, A. (eds.) Cervical spine, pp. 322–327. Springer, Wien (1987)

135 36. Merloz, P., Tonetti, J., Pittet, L., et  al.: Computer-assisted spine surgery. Comput Aided Surg 3, 297–305 (1998) 37. Nolte, L.P., Slomczykowski, M.A., Berlemann, U., et  al.: A new approach to computer-aided spine surgery: fluoroscopy-based surgical navigation. Eur Spine J 9(Suppl 1), S78–S88 (2000) 38. Ortmaier, T., Weiss, H., Dobele, S., et  al.: Experiments on robot-assisted navigated drilling and milling of bones for pedicle screw placement. Int J Med Robot 2, 350–363 (2006) 39. Paramore, C.G., Dickman, C.A., Sonntag, V.K.: The anatomical suitability of the C1-2 complex for transarticular screw fixation. J Neurosurg 85, 221–224 (1996) 40. Rajasekaran, S., Vidyadhara, S., Ramesh, P., et al.: Randomized clinical study to compare the accuracy of navigated and nonnavigated thoracic pedicle screws in deformity correction surgeries. Spine (Phila Pa 1976) 32, E56–E64 (2007) 41. Rajasekaran, S., Vidyadhara, S., Shetty, A.P.: Iso-C3D fluoroscopy-based navigation in direct pedicle screw fixation of Hangman fracture: a case report. J Spinal Disord Tech 20, 616–619 (2007) 42. Rajasekaran, S., Vidyadhara, S., Shetty, A.P.: Intra-operative Iso-C3D navigation for pedicle screw instrumentation of hangman’s fracture: a case report. J Orthop Surg (Hong Kong) 15, 73–77 (2007) 43. Rosenthal, D.I., Springfield, D.S., Gebhardt, M.C., et  al.: Osteoid osteoma: percutaneous radio-frequency ablation. Radiology 197, 451–454 (1995) 44. Schenck, J.F., Jolesz, F.A., Roemer, P.B., et  al.: Superconducting open-configuration MR imaging system for image-guided therapy. Radiology 195, 805–814 (1995) 45. Schlenzka, D., Laine, T., Lund, T.: Computer-assisted spine surgery. Eur Spine J 9(Suppl 1), S57–S64 (2000) 46. Skaf, G.S., Sabbagh, A.S., Hadi, U.: The advantages of submandibular gland resection in anterior retropharyngeal approach to the upper cervical spine. Eur Spine J 16, 469– 477 (2007) 47. Stulik, J., Suchomel, P., Lukas, R., et al.: Primary osteosynthesis of the odontoid process: a multicenter study. Acta Chir Orthop Traumatol Cech 69, 141–148 (2002) 48. Suchomel, P., Buchvald, P., Barsa, P., et al.: Pyogenic osteomyelitis of the odontoid process: single stage decompression and fusion. Spine (Phila Pa 1976) 28, E239–E244 (2003) 49. Suchomel, P., Buchvald, P., Barsa, P., et  al.: Single-stage total C-2 intralesional spondylectomy for chordoma with three-column reconstruction. Technical note. J Neurosurg Spine 6, 611–618 (2007) 50. Suchomel, P., Buchvald, P., Barsa, P., et  al.: Instability of craniovertebral junction and upper cervical spine. Abstract, 8th ESBS Congress Prague 2007, Skull Base 17 (2007) 51. Suchomel, P., Hradil, J., Barsa, P., et al.: Surgical treatment of fracture of the ring of axis – “hangman’s fracture”. Acta Chir Orthop Traumatol Cech 73, 321–328 (2006) 52. Summers, L.E., Kouri, J.G., Yang, M., et al.: Odontoid screw placement using Isocentric 3-dimensional C-arm fluoroscopy. J Spinal Disord Tech 21, 45–48 (2008) 53. Takahashi, J., Shono, Y., Nakamura, I., et  al.: Computerassisted screw insertion for cervical disorders in rheumatoid arthritis. Eur Spine J 16, 485–494 (2007) 54. Taller, S., Suchomel, P., Lukas, R., et al.: CT-guided internal fixation of a hangman’s fracture. Eur Spine J 9, 393–397 (2000)

136 55. Ugur, H.C., Kahilogullari, G., Attar, A., et al.: Neuronavigationassisted transoral-transpharyngeal approach for basilar invagination – two case reports. Neurol Med Chir (Tokyo) 46, 306–308 (2006) 56. Van Cleynenbreugel, J., Schutyser, F., Goffin, J.: Imagebased planning and validation of C1-C2 transarticular screw fixation using personalized drill guides. Comput Aided Surg 7, 41–48 (2002) 57. Veres, R., Bago, A., Fedorcsak, I.: Early experiences with image-guided transoral surgery for the pathologies of the upper cervical spine. Spine (Phila Pa 1976) 26, 1385–1388 (2001)

7  Virtual and Real TimeNavigational Techniques 58. Vougioukas, V.I., Hubbe, U., Schipper, J.: Navigated transoral approach to the cranial base and the craniocervical junction: technical note. Neurosurgery 52, 247–250 (2003). discussion 251 59. Wang, M.Y.: C2 crossing laminar screws: cadaveric morphometric analysis. Neurosurgery 59, ONS84–ONS88 (2006). discussion ONS84-88 60. Woodard, E.J., Leon, S.P., Moriarty, T.M.: Initial experience with intraoperative magnetic resonance imaging in spine surgery. Spine (Phila Pa 1976) 26, 410–417 (2001)

Section Indications for Surgery and Examples of Reconstruction

III

8

Traumatic Atlantooccipital Dislocation (AOD) P. Suchomel and V. Beneš

Atlanto-occipital dislocation (AOD) is a rare, highly unstable injury of the craniovertebral junction (CVJ) and, as such, is associated with high mortality and neurological morbidity. Its first description dates back to 1908, when Blackwood described a patient surviving his injury for nearly 35 h [3]. AOD may be defined as acute traumatic osteoligamentous instability between the occiput and the atlas [2]. According to the relationship of atlas and occiput, Traynelis, in 1986, classified AOD into three types [28]. Type I involves anterior displacement of the occiput with respect to the atlas. Type II is a longitudinal distraction, whereas Type III results when the occiput is displaced posteriorly relative to the atlas.

8.1 Etiology Although AOD is relatively rare, it was found in up to 90% of fatal cervical spine injuries [1, 4] and represented 1% of all (dead or surviving) cervical spine injuries [23]. High energy trauma, such as motor vehicle accident (MVA), is the usual cause, with an incidence of 8–31% among traffic victims [4, 29]. AOD is more common among children and young adults due to increased laxity of ligaments and the disparity between occipital condyles and articular surfaces of the atlas [5, 19, 26]. The disproportionate size of the head in relation to spine can also play a significant role [14].

8.2 Clinical Symptoms AOD victims can have surprisingly few or no neurological symptoms [2, 14]. In one series, 27% of patients did not present with neurological deficits [14]. More commonly, patients present with spinal cord injury of a varying degree, lower cranial nerve deficits, or bulbar-cervical dissociation [2, 10, 14, 24, 25, 28]. AOD frequently results into sudden death and is diagnosed at postmortem examination [1, 20]. Other associated injuries are common and include, traumatic brain injury, skull fractures, concomitant traumatic instability of lower cervical spine, and other injuries to multiple organ systems in context of a polytrauma victim [1, 2, 9, 14, 17, 18, 24]. Due to improved rescue services and emergency protocols, more AOD victims nowadays reach hospital alive. High index of suspicion for AOD must be maintained, since delay in diagnosis can have devastating consequences [2, 14]. Impaired consciousness and other associated injuries in a polytraumatized victim can lead to an under-appreciation of the situation and delay in diagnosis. In Bellabarba’s series, a delay in diagnosis occurred in 75% of patients and 38% of those were diagnosed with AOD only after a neurological deterioration [2]. Therefore, a careful and thorough radiological evaluation is of utmost importance and AOD must be ruled out in every polytrauma patient, especially with impaired level of consciousness.

8.3 Radiology P. Suchomel and V. Beneš Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic

CT especially in 3D or midsagittal reconstruction (Fig.  8.1) and lateral cervical spine radiography (Fig.  8.2) can show increased both basion – dens

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_8, © Springer-Verlag Berlin Heidelberg 2011

139

140

8  Traumatic Atlantooccipital Dislocation (AOD)

Fig. 8.1  Atlantooccipital dislocation (AOD) with distraction type II according to Traynelis. (a) Lateral view of 3D reconstruction, note increased BDI. (b) Reconstruction in coronal plane

Fig. 8.2  Plain X-rays of AOD with anterior dislocation, type I according to Traynelis in two different patients. (a) Note contrast media in vessels. Picture obtained after angiography confirming the brain death because of legal reasons

interval (BDI) as well as basion – axial interval (BAI); the cut-off value is considered 12 mm for both parameters [2, 12, 13]. The Powers ratio is greater than 1.0 [23]. Associated injuries to the C0-C2 complex can also be appreciated this way and are rather common in AOD patients [2, 14]. Cranial CT may reveal evidence of intracranial traumatic injuries. If a CT scan is without abnormalities and a high suspicion for upper cervical spine injury still exists, an MRI can confirm AOD diagnosis. The findings then include abnormal signal in the C0-C1 joint capsule or ligamentous structures, such as the posterior atlantooccipital membrane, alar, apical, and cruciate ligaments [6,

14, 16]. Based on ligamentous injury, Bellabarba et al. classified AOD into three stages. Stage 1 was defined as stable minimally or non-displaced injury with sufficiently preserved ligamentous integrity. Such injury included unilateral alar ligament avulsion or partial ligamentous injury or sprain. Stage 2 injury was partially or completely spontaneously reduced bilateral AOD with minimal displacement, where traction test confirms loss of ligamentous integrity. Both BDI and BAI are no more than 2 mm beyond normal values. Stage 3 injury denotes a gross craniocervical displacement with BDI and BAI more than 2 mm beyond upper limit [2]. Horn [14] pointed out a drawback of this classification

141

8.5  Our Preference

system: its reliance on plain and dynamic radiography to determine instability, and proposed his own classification. It relies on described CT and MRI features. Grade I injury denotes normal findings on CT (Powers ratio, BDI, BAI) with moderately abnormal findings on MRI (high signal in posterior ligaments or occipitoatlantal joints). Grade II injury shows more than one abnormal finding on CT or grossly abnormal MRI findings in occipitoatlantal joints, tectorial membrane, or alar or cruciate ligaments [14]. Previously mentioned two classification scales [2, 14] have therapeutic implications.

8.4 Treatment Strategy It is important to emphasize again that AOD represents a highly unstable pure ligamentous injury, which can be seen in up to 31% of traffic accident victims. Therefore, it is extremely important to consider this entity during first aid and transportation of a trauma victims. Any undesirable head movement can cause fatal upper spinal cord injury. This is of particular importance during extrication of MVA victims trapped in their vehicles. AOD is the most dangerous spine injury in terms of inappropriate handling of patients and its potential occurrence was one of the main reasons for changing rescue guidelines in the past. It is now a standard that emergency service personnel always immobilize the victim’s cervical spine with a hard collar at the scene prior to any manipulation. AOD treatment is either conservative and relies on external immobilization, or surgical, in which case the mainstay is posterior occipitocervical fusion. External orthosis, such as halo vest or Philadelphia collar are possible therapeutical solutions in selected cases (Bellabarba Stage 1 and Horn Grade I injuries). However, this therapy can also be associated with neurological deterioration [7, 8, 15] or late segmental instability [8, 15, 22, 27]. These therapeutical measures can be recommended as a provisional fixation until a definite treatment is undertaken or in relatively stable injuries in children, where potential of ligamentous healing is expected. Since AOD is primarily a ligamentous injury, posterior occipitocervical fixation and fusion with a bone graft offer a definitive treatment. As emphasized earlier, AOD is associated with high rates of morbidity and mortality. The high fatality rate suggests that a chance of surviving this injury is remote [1]. However, upon reaching a hospital, the best predictor of outcome is the

severity of neurological injuries at presentation. In a series of 17 treated survivors, functional status improved after surgery in 85% of those presenting with spinal cord injury and no change was noted in two patients initially neurologically intact [2]. Another institutional series of 33 patients reported 28 survivors; remarkably, 14 of these were neurologically intact [14]. Surgical complications with neurological consequences from these two only series reported to date were seen in one [2] and two [14] patients, respectively. When appropriate, high dose steroid regimen is implemented upon presentation and any surgical intervention is undertaken under electrophysiological monitoring. The patient is usually intubated fiberoptically and maintained under total intravenous anesthesia (TIVA) to allow for motor-evoked potential monitoring. Halo ring or halo vest are applied without traction to stabilize the CVJ, baseline electrophysiological data are obtained, and then the patient is carefully positioned prone and secured to the operating table. The occiput and at least the first two cervical vertebrae are exposed using a standard posterior midline approach. The occipitocervical fixation is then carried out with the C0-C1 joints anatomically aligned. The exact type of fixation and fusion is dependent on the severity of injury. In mild forms, especially in children and young adults, only a short construct can be used. Grob et al. published a technique of direct atlanto-occipital screw fixation in such cases [11]. Similarly, Maughan et al. used a short occipitoatlantal fixation using an occipital plate connected with rods to polyaxial C1 lateral mass screws for the treatment of circular occipital bone fracture [21]. However, in more severe grades of AOD, a strong construct between occiput and at least two cervical vertebrae is preferred. External immobilization (e.g., halo, SOMI, or hard collar) is commonly continued for 3 months postoperatively even following a solid fixation and fusion.

8.5 Our Preference Unfortunately, large series of AOD survivors are mostly published in the United States. This is probably due to a superior organization and standard protocols adopted by the emergency medical services when it comes to extrication and transport of polytrauma victims. In Europe, survivors are discussed in case reports and the majority of our patients with AOD are

142

admitted to the hospital in pentaplegic status. Thus, a proper education of public and emergency services is essential to avoid a secondary injury associated with improper manipulation of the head of an injured victim. This includes traction. A prompt placement of rigid external orthosis and full spinal precautions at the scene of an accident are a standard ATLS protocol. In our opinion, the morphological classification is slightly misleading. If the head is structurally disconnected from the spine, it can freely move in any direction. Therefore, its position at the time of CT image acquisition may be quite different from that at the time of impact. In terms of surgical indication, the abovementioned Horn grading system is more practical. Upon arrival to the hospital, most trauma patients are stabilized in the trauma bay and then transported from the emergency department to a CT scanner. A spiral CT scanner may reveal a possible AOD and thus dictate further manipulation precautions (Fig.  8.3). In such cases of suspected AOD, it is time-consuming and inefficient to be obtaining plain radiographs. If the patient is hemodynamically stable, we prefer a direct MRI evaluation of the CVJ (Fig. 8.4). In indicated cases, the patient should undergo surgical fixation as soon as possible, provided no other emergent surgical treatments are necessary. Immediate surgical fixation rids the patient of any subsequent manipulation hazard. Intubation (preferably fiberoptic) and prone positioning on the surgical table need to be carried out with extreme caution. Electrophysiological monitoring should be instituted prior to those maneuvers. Depending on the severity of the injury, the occiput

Fig. 8.3  AOD type I according to Traynelis. (a) Whole spine spiral CT. (b) Detail of 3D reconstruction in sagittal plane

8  Traumatic Atlantooccipital Dislocation (AOD)

and UCS and/or more caudal cervical vertebrae are exposed via a standard posterior midline approach in all cases of AOD. Being careful around the vertebral artery behind the C1 lateral mass, the degree of dislocation of the atlanto-occipital joint should be visualized directly. Reduction, if necessary, is achieved directly by manipulation of the head relative to C1 under direct vision and fluoroscopic guidance. Our preferred construct involves an occipital plate firmly anchored in the midline and directly connected to C1 lateral mass screws and C2 pedicle screws by two appropriately shaped rods (Fig.  8.5). In situations where such construct is not possible due to the morphology of the injury, we extend the fixation to lower levels. Iliac crest bone graft is always added to allow for fusion between the occiput and UCS (both laminae and spinous process). A short monosegmental construct in mild forms of AOD may be sufficient as described by some authors but we do not share such experience. We obtain a postoperative CT to check adequacy of the construct and upright films in external orthosis as soon as possible to assess the construct under appropriate loading. The length of period of external immobilization is judged on an individual basis depending on the injury, construct used, compliance of the patient, and follow-up imaging. In conclusion, AOD is a highly unstable injury and must be ruled out in any major trauma victim, because delayed diagnosis frequently results in neurological deterioration. Computer tomography is the study of choice as a fast, first-line evaluation of trauma severity to the CVJ. However, time and situation permitting, an MRI evaluation allows for a better assessment of the

143

References

References

Fig. 8.4  MRI in a patient surviving the AOD type I. Edema of upper spinal cord and lower brain stem

Fig. 8.5  Occipitocervical fusion in patient with AOD

ligaments of the CVJ. Despite the high mortality rate, survival is well documented. External orthosis can be used in carefully selected cases or as a temporary measure prior to a definitive treatment. Occipitocervical fusion is the treatment of choice that offers an immediate, solid stabilization of the CVJ.

  1. Alker Jr., G.J., Oh, Y.S., Leslie, E.V.: High cervical spine and craniocervical junction injuries in fatal traffic accidents: a radiological study. Orthop Clin North Am 9, 1003–1010 (1978)   2. Bellabarba, C., Mirza, S.K., West, G.A., et  al.: Diagnosis and treatment of craniocervical dislocation in a series of 17 consecutive survivors during an 8-year period. J Neurosurg Spine 4, 429–440 (2006)   3. Blackwood, N.J.: III. Atlo-occipital dislocation: a case of fracture of the atlas and axis, and forward dislocation of the occiput on the spinal column, life being maintained for thirty-four hours and forty minutes by artificial respiration, during which a laminectomy was performed upon the third cervical vertebra. Ann Surg 47, 654–658 (1908)   4. Bucholz, R.W., Burkhead, W.Z.: The pathological anatomy of fatal atlanto-occipital dislocations. J Bone Joint Surg Am 61, 248–250 (1979)   5. Bulas, D.I., Fitz, C.R., Johnson, D.L.: Traumatic atlantooccipital dislocation in children. Radiology 188, 155–158 (1993)   6. Bundschuh, C.V., Alley, J.B., Ross, M., et al.: Magnetic resonance imaging of suspected atlanto-occipital dislocation. Two case reports. Spine (Phila Pa 1976) 17, 245–248 (1992)   7. DiBenedetto, T., Lee, C.K.: Traumatic atlanto-occipital instability. A case report with follow-up and a new diagnostic technique. Spine (Phila Pa 1976) 15, 595–597 (1990)   8. Donahue, D.J., Muhlbauer, M.S., Kaufman, R.A., et  al.: Childhood survival of atlantooccipital dislocation: underdiagnosis, recognition, treatment, and review of the literature. Pediatr Neurosurg 21, 105–111 (1994)   9. Dublin, A.B., Marks, W.M., Weinstock, D., et al.: Traumatic dislocation of the atlanto-occipital articulation (AOA) with short-term survival. With a radiographic method of measuring the AOA. J Neurosurg 52, 541–546 (1980) 10. Eismont, F.J., Bohlman, H.H.: Posterior atlanto-occipital dislocation with fractures of the atlas and odontoid process. J Bone Joint Surg Am 60, 397–399 (1978) 11. Grob, D.: Transarticular screw fixation for atlanto-occipital dislocation. Spine (Phila Pa 1976) 26, 703–707 (2001) 12. Harris Jr., J.H., Carson, G.C., Wagner, L.K.: Radiologic diagnosis of traumatic occipitovertebral dissociation: 1 Normal occipitovertebral relationships on lateral ­radiographs of supine subjects. AJR Am J Roentgenol 162, 881–886 (1994) 13. Harris Jr., J.H., Carson, G.C., Wagner, L.K., et al.: Radiologic diagnosis of traumatic occipitovertebral dissociation: 2. Comparison of three methods of detecting occipitovertebral relationships on lateral radiographs of supine subjects. AJR Am J Roentgenol 162, 887–892 (1994) 14. Horn, E.M., Feiz-Erfan, I., Lekovic, G.P., et al.: Survivors of occipitoatlantal dislocation injuries: imaging and clinical correlates. J Neurosurg Spine 6, 113–120 (2007) 15. Hosono, N., Yonenobu, K., Kawagoe, K., et al.: Traumatic anterior atlanto-occipital dislocation. A case report with survival. Spine (Phila Pa 1976) 18, 786–790 (1993) 16. Chaljub, G., Singh, H., Gunito Jr., F.C., et  al.: Traumatic atlanto-occipital dislocation: MRI and CT. Neuroradiology 43, 41–44 (2001)

144 17. Chattar-Cora, D., Valenziano, C.P.: Atlanto-occipital dislocation: a report of three patients and a review. J Orthop Trauma 14, 370–375 (2000) 18. Junge, A., Krueger, A., Petermann, J., et  al.: Posterior atlanto-occipital dislocation and concomitant discoligamentous C3-C4 instability with survival. Spine (Phila Pa 1976) 26, 1722–1725 (2001) 19. Kaufman, R.A., Dunbar, J.S., Botsford, J.A., et al.: Traumatic longitudinal atlanto-occipital distraction injuries in children. AJNR Am J Neuroradiol 3, 415–419 (1982) 20. Lesoin, F., Blondel, M., Dhellemmes, P., et  al.: Posttraumatic atlanto-occipital dislocation revealed by sudden cardiopulmonary arrest. Lancet 2, 447–448 (1982) 21. Maughan, P.H., Horn, E.M., Theodore, N., et al.: Avulsion fracture of the foramen magnum treated with occiput-to-c1 fusion: technical case report. Neurosurgery 57, E600 (2005). discussion E600 22. Palmer, M.T., Turney, S.Z.: Tracheal rupture and atlantooccipital dislocation: case report. J Trauma 37, 314–317 (1994)

8  Traumatic Atlantooccipital Dislocation (AOD) 23. Powers, B., Miller, M.D., Kramer, R.S., et  al.: Traumatic anterior atlanto-occipital dislocation. Neurosurgery 4, 12–17 (1979) 24. Rao, G., Arthur, A.S., Apfelbaum, R.I.: Circumferential fracture of the skull base causing craniocervical dislocation. Case report. J Neurosurg 97, 118–122 (2002) 25. Saeheng, S., Phuenpathom, N.: Traumatic occipitoatlantal dislocation. Surg Neurol 55, 35–40 (2001). discussion 40 26. Shamoun, J.M., Riddick, L., Powell, R.W.: Atlanto-occipital subluxation/dislocation: a “survivable” injury in children. Am Surg 65, 317–320 (1999) 27. Sponseller, P.D., Cass, J.R.: Atlanto-occipital fusion for dislocation in children with neurologic preservation. A case report. Spine (Phila Pa 1976) 22, 344–347 (1997) 28. Traynelis, V.C., Marano, G.D., Dunker, R.O., et  al.: Traumatic atlanto-occipital dislocation. Case report. J Neurosurg 65, 863–870 (1986) 29. Zivot, U., Di Maio, V.J.: Motor vehicle-pedestrian accidents in adults. Relationship between impact speed, injuries, and distance thrown. Am J Forensic Med Pathol 14, 185–186 (1993)

9

Occipital Condyle Fractures P. Suchomel and L. Jurák

Occipital condyle fracture (OCF) was first described by Charles Bell after an autopsy [4]. OCF was considered to be a very rare injury accompanying head trauma with a high rate of mortality in the past [5, 6, 14]. Nowadays, it is encountered more frequently due to developed rescue services allowing higher survival rate, especially for traffic accident victims. The polytrauma patients are usually transported directly to emergency department where the first radiological investigation is a “whole body” spiral CT scan. This means a much earlier diagnosis of UCS injury including OCF [3, 11, 15]. Nevertheless, it is still quite a rare injury, which can easily be overlooked [3, 11, 15]. Despite a lack of uniformly accepted clinical evaluation of the patient’s status, there are two widely accepted classifications of OCFs. The first descriptive classification was suggested by Anderson and Montesano (1988) derived from an analysis of their six patients [1] dividing OCFs into three types. Type I:

occipital condyle comminution with no or minimal displacement

Type II:

dislocated occipital condyle fractures

Type III:

those with stable C0-1-2 complex

The authors proposed that condyle comminution is a result of extreme axial load; the avulsion occurs if the axial force is combined with bending and/or rotation with concomitant participation of alar ligament traction and that the type II is a part of skull base injury usually caused by high energy blunt trauma. P. Suchomel and L. Jurák Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova st. 10, 46063 Liberec, Czech Republic

Tuli et al. opposed that the above classification does not provide any risk stratification or treatment guidance and instead suggested dividing OCFs into two basic types and two other subtypes [22]. Their classification was based on a review of literature and analysis of only three of their own cases. Type 1:

fractures of condyle without dislocation

Type 2:

dislocated occipital condyle fractures

Type 2a:

those with stable C0-1-2 complex

Type 2b:

fractures with evidence of occipitoatlanto-axial instability (OAA)

In this schema, only the very rare type 2b fractures are considered for surgery. Although this classification seems to be more logical than the first one, it is not easy to determine stability based on the authors’ required parameters that can only be measured on thin sliced CT images with some of them obtained in dynamic positions. OCF displacement is usually defined as greater than 2 mm of fragment separation [3, 11]. Currently, two basic questions have to be answered when deciding whether to operate or not [15]: Firstly, is there any reason to decompress neural structures possibly compressed by the condyle fragments? Secondly, is the OCF unstable enough to require surgical fixation?

9.1 Etiology and Epidemiology OCFs are present in 0.1–0.4% of patients admitted to hospitals with traumatic injuries. Often, OCF is caused by high impact blunt trauma [3, 11]. Patients are most frequently involved in MVA (55%) but also falls from height (34%) and assaults (9%) can result in condylar

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_9, © Springer-Verlag Berlin Heidelberg 2011

145

146

9  Occipital Condyle Fractures

injury [15]. TBI is concomitantly present in up to 56% patients and associated injuries of the rest of the cervical spine are found in 20–31% patients [3, 11, 15]. There is a significant male predominance (70%) and the types II and III, according to Anderson and Montesano, are the most frequent types. The vast majority of diagnosed cases are unilateral and do not have any condyle fragment displacement (Tuli Type I) [3, 11, 15].

complex clinical findings are seen in those with impaired consciousness and neurologic deficit caused by TBI [1–3, 7, 8, 13, 20, 22]. Sometimes, clinical symptoms of cranial nerve involvement do not appear initially but present in a delayed fashion. It may be due to osseous and fibrous tissue proliferation as a reparative process or due to inadequate stabilization of bony fragments [7, 9, 16, 19, 22].

9.2 Clinical Symptoms

9.3 Radiology

Neurologic deficit is rarely seen in isolated, unilateral OCFs [15]. The patient usually suffers from a nonspecific pain exacerbated by head movement; however, unilateral hypoglossal paralysis due to OCF has been described [10, 12, 21, 23]. Clinical symptoms related to OCFs are masked by the overwhelming effects of TBI or other traumatic injuries in more than a half of the patients [15]. The most

The diagnosis of OCF is rarely made on plain films. Computerized tomography is the modality of choice. Fracture lines visible on standard axial CT are usually inadequate in defining the exact morphology and extent of the fracture (Fig. 9.1). 3D CT reconstructions are usually necessary to delineate the true fracture pattern and the degree of displacement (Fig.  9.2). Fracture mor­ phology is not the only parameter determining the

Fig. 9.1  CT of broad condyle avulsion. (a) Axial image. (b) Reconstruction in frontal plane

Fig. 9.2  3D CT of broad condyle avulsion, the same patient from Fig. 9.1. (a) Lateral view. (b) Internal view

147

9.5  Our Preference

Fig.  9.3  Normal CCI distance (less than 2  mm) depicted on parasagittal CT reconstruction in a case of condyle avulsion

management of the injury. Stability of the AO joint must also be established. Apart from different techniques used in the evaluation of AOD, the occipital condyle – C1 interval (CCI) is considered to be the most relevant index. Pang et al. [17, 18] described it to be symmetrical and not exceed 2 mm (Fig. 9.3). MRI is indicated to evaluate the ligamental integrity and possible neural compression in patients with suspicion for instability or those with neurological deficit [2, 3, 11, 15, 22]. Certainly, more extensive workup is needed in OCFs combined with other upper cervical spine (UCS) injuries.

9.4 Treatment Strategy Treatment of OCFs is conservative in the vast majority of cases. Surgical decompression is rarely indicated only if neural structures are directly compressed by

Fig. 9.4  Nondisplaced comminuted fracture of occipital condyle (Type I Anderson and Montesano) successfully treated without bracing. (a) Axial CT scan. (b) Coronal reconstruction

displaced fracture fragments. Stabilization procedures are seldom necessary and required only in situations with AO instability, malalignment or complex instabilities of UCS in the combined injuries. Depending on fracture type, conservative management varies from activity restriction only without immobilization to a rigid cervical collar for 6  weeks followed by dynamic plain films or CT. Halo-vest or SOMI brace are rarely used except for combined UCS injuries [15]. Patients with bilateral OCFs or occipitoatlantal or atlantoaxial instability may require halo traction followed by hard external bracing or surgical occipitocervical fusion [11]. In the largest published retrospective analysis of 100 patients with 106 OCFs reported by Masserati et  al. [15], only two patients with AO malalignment required surgical stabilization (one primarily and one delayed) and in the other four halo-vests were used (in three because of combined UCS injury). Conversely, 19.3% of their patients were discharged without any external cervical support.

9.5 Our Preference Single, isolated nondislocated fractures of all Anderson and Montesano types can be treated conservatively, in our opinion. Those patients with type I fractures where the condyle bearing capacity looks damaged by less than 50% can be treated by simple activity restriction without external support (Fig. 9.4). Similarly, external immobilization seems to be unnecessary when it comes to wedge locked, broad based, type III fractures, and single line fractures of skull base extending into occipital condyle (Fig. 9.5). However, type III fractures where the alar ligament insertion tubercle is detached, hard collar should

148

Fig.  9.5  Broad condyle avulsion – more skull base fracture passing through the jugular foramen (Type II Anderson and Montesano). (a) Axial scan showing the fracture passing through

9  Occipital Condyle Fractures

jugular foramen. (b) coronal view showing some degree of ­displacement. (c) 3D CT internal view

be recommended (Fig. 9.6). Type I comminution of the condyle and/or shallow based type III avulsion (Fig. 9.7) can be prone to settling and displacement and therefore, hard collar treatment should be worn for at least 6 weeks with appropriate radiographic follow-up. Transoral images can show progressive asymmetry caused by dislocation of the condyle which has to be confirmed by CT during the followup. The final status of alignment and fracture healing should be evaluated by CT approximately 3–6 months after the injury.

Fig. 9.7  Typical occipital condyle avulsion with mild dislocation (Type III Anderson and Montesano). (a) Reconstruction in frontal plane. (b) Anterior 3D view

Fig. 9.6  Mild avulsion of the occipital condyle at the place of alar ligament attachment (Type III Anderson and Montesano), note the simultaneous fracture of C1 lateral mass

We do not have our own experience with acute or delayed instability and/or deformity in OCF requiring surgical intervention. We suppose that, in some bilateral condyle fractures (Fig. 9.8), circular foramen magnum breach, atlanto-occipital joint disruption, and asymmetric condylar collapse, surgical intervention limited to the damaged segments has to be considered.

References

Fig.  9.8  Bilateral condyle avulsion depicted on coronal CT reconstruction

References   1. Anderson, P.A., Montesano, P.X.: Morphology and treatment of occipital condyle fractures. Spine (Phila Pa 1976) 13, 731–736 (1988)   2. Anonymous: Occipital condyle fractures. Neurosurgery 50, S114–S119 (2002)   3. Aulino, J.M., Tutt, L.K., Kaye, J.J., et al.: Occipital condyle fractures: clinical presentation and imaging findings in 76 patients. Emerg Radiol 11, 342–347 (2005)   4. Bell, C.: Surgical observations. Middlesex Hosp J 4, 469– 470 (1817)   5. Blacksin, M.F., Lee, H.J.: Frequency and significance of fractures of the upper cervical spine detected by CT in patients with severe neck trauma. AJR Am J Roentgenol 165, 1201–1204 (1995)   6. Bloom, A.I., Neeman, Z., Slasky, B.S., et al.: Fracture of the occipital condyles and associated craniocervical ligament injury: incidence, CT imaging and implications. Clin Radiol 52, 198–202 (1997)   7. Bolender, N., Cromwell, L.D., Wendling, L.: Fracture of the occipital condyle. AJR Am J Roentgenol 131, 729–731 (1978)   8. Clayman, D.A., Sykes, C.H., Vines, F.S.: Occipital condyle fractures: clinical presentation and radiologic detection. AJNR Am J Neuroradiol 15, 1309–1315 (1994)

149   9. Deeb, Z.L., Rothfus, W.E., Goldberg, A.L., et  al.: Occult occipital condyle fractures presenting as tumors. J Comput Tomogr 12, 261–263 (1988) 10. Demisch, S., Lindner, A., Beck, R., et al.: The forgotten condyle: delayed hypoglossal nerve palsy caused by fracture of the occipital condyle. Clin Neurol Neurosurg 100, 44–45 (1998) 11. Hanson, J.A., Deliganis, A.V., Baxter, A.B., et al.: Radiologic and clinical spectrum of occipital condyle fractures: retrospective review of 107 consecutive fractures in 95 patients. AJR Am J Roentgenol 178, 1261–1268 (2002) 12. Legros, B., Fournier, P., Chiaroni, P., et al.: Basal fracture of the skull and lower (IX, X, XI, XII) cranial nerves palsy: four case reports including two fractures of the occipital condyle – a literature review. J Trauma 48, 342–348 (2000) 13. Leone, A., Cerase, A., Colosimo, C., et al.: Occipital condylar fractures: a review. Radiology 216, 635–644 (2000) 14. Link, T.M., Schuierer, G., Hufendiek, A., et al.: Substantial head trauma: value of routine CT examination of the cervicocranium. Radiology 196, 741–745 (1995) 15. Maserati, M.B., Stephens, B., Zohny, Z., et  al.: Occipital condyle fractures: clinical decision rule and surgical management. J Neurosurg Spine 11, 388–395 (2009) 16. Orbay, T., Aykol, S., Seckin, Z., et  al.: Late hypoglossal nerve palsy following fracture of the occipital condyle. Surg Neurol 31, 402–404 (1989) 17. Pang, D., Nemzek, W.R., Zovickian, J.: Atlanto-occipital dislocation – part 2: The clinical use of (occipital) condyleC1 interval, comparison with other diagnostic methods, and the manifestation, management, and outcome of atlantooccipital dislocation in children. Neurosurgery 61, 995–1015 (2007). discussion 1015 18. Pang, D., Nemzek, W.R., Zovickian, J.: Atlanto-occipital dislocation: part 1 – normal occipital condyle-C1 interval in  89 children. Neurosurgery 61, 514–521 (2007). discussion 521 19. Savolaine, E.R., Ebraheim, N.A., Jackson, W.T., et  al.: Three-dimensional computed tomography in evaluation of occipital condyle fracture. J Orthop Trauma 3, 71–75 (1989) 20. Spencer, J.A., Yeakley, J.W., Kaufman, H.H.: Fracture of the occipital condyle. Neurosurgery 15, 101–103 (1984) 21. Su, T.M., Lui, C.C., Cheng, M.H., et al.: Occipital condyle fracture with hypoglossal nerve palsy: case report. J Trauma 49, 1144–1146 (2000) 22. Tuli, S., Tator, C.H., Fehlings, M.G., et al.: Occipital condyle fractures. Neurosurgery 41, 368–376 (1997). discussion 376–367 23. Urculo, E., Arrazola, M., Arrazola Jr., M., et  al.: Delayed glossopharyngeal and vagus nerve paralysis following occipital condyle fracture. Case report. J Neurosurg 84, 522–525 (1996)

10

Atlas Fractures P. Suchomel and R. Brabec

Fractures of the atlas comprise approximately 2–13% of cervical spine injuries and about 1–3% of the fractures of the entire spinal column [13, 14, 29, 39]. They are either isolated or occur in 23–57% as combined with other UCS but also subaxial cervical spine injuries [3, 10, 14, 24, 26, 27]. The first description of atlas fracture was an autopsy report by Cooper, in 1823 [4]. Sir Geoffrey Jefferson was the first who comprehensively described the atlas “burst” fracture that bears his name in 1920 [22]. He stated that: “If the atlas was morphologically similar to the other vertebrae, death would be the common result of fracture.” Nevertheless, in his own series of four cases (two patients and two museum specimens) analyzed together with literature survey of 42 others, he found only in 50% of patients a consequent neurological symptomatology and thus as the first also validated the relative clinical benignity of C1 isolated fractures. From those times many large series of patients with atlas fractures were published; however, a conclusive statement giving us the therapeutic guidance on higher level of evidence is not available up to now [10, 13, 14, 26, 29, 39].

10.1 Classification There is no uniformly accepted classification system of atlas fractures to date. Many attempts to classify C1 fractures were published. Levine and Edwards [28, 29] divided their series of 34 patients into three groups.

P. Suchomel and R. Brabec Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova st. 10, 46063 Liberec, Czech Republic

Most frequently, the posterior arch fracture was seen, then the lateral mass area fractures often causing asymmetric single mass displacement were described as second category, and the last group was three and/or four fragmental Jefferson type fractures. More precise is the classification proposed by Landels and Van Peteghem from Vancouver dividing the fractures into three types [26]. Type I are the isolated fractures of single arch not crossing the equator of atlas. Type II are the fractures of arches crossing the equator having two and more fragments including the four-fragmental Jefferson fracture. Unstable were evaluated as those type II fractures with summarized lateral mass overhang more than 6.9 mm, according to Spence [41]. Type III are those fractures involving lateral mass and maximally one arch. In their analyzed series of 35 patients having atlas fracture, they found the concomitant other level cervical spine injury in 57%. No one patient had neurological deficit related to C1 fracture. Another and even more complex classification was proposed by Aebi and Nazarian [1]. Also, case studies describing unusual horizontal fracture of anterior arch probably caused by hyperextension together with counter action of anterior tubercle attachment of longus colli muscle were reported [23, 32, 42]. Dickman and Green suggested the most descriptive classification system respecting the therapeutic con­ sequences [5]. They divided the fractures of C1 into 6 categories: Type A:

fracture of the anterior arch

Type B:

fracture of the posterior arch

Type C:

simple lateral mass fracture

Type D:

comminuted lateral mass fracture

Type E:

four-part ring fracture (Jefferson type)

Type F:

two-part ring fracture

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_10, © Springer-Verlag Berlin Heidelberg 2011

151

152

10  Atlas Fractures

The fractures were recommended for treatment successively from A to F depending also on the status of TAL with increasing stiffness of external support or even surgery. Perhaps, currently, the most widely accepted is the classification of Gehweiler et al. interestingly proposed as one of the first but accepted as the last because of its printed presentation in radiological literature [12]. This classification describes five types of atlas ­fractures (Fig. 10.1): Type I:

fracture of the anterior arch

Type II:

fracture of the posterior arch

Type III:

combined fracture of both anterior and posterior arch (Jefferson incl.)

Type IV:

isolated simple or burst fracture of massa lateralis

Type V:

fracture of transverse process

For all the previously mentioned authors, the crucial point how to define the stability of atlas fractures was the functional integrity of TAL. Initially, most of the surgeons accepted the “rule of Spence” saying that summarized overhang of dislocated lateral masses over C2 superior facets should not exceed 6.9  mm on the transoral radiogram (Fig.  10.2). This idea came from the work of Spence et al. [41] who tested the distractive force causing TAL rupture on ten cadaveric isolated

specimens cut like in Jefferson four-part fractures. They found that the TAL can tear under distractive force of average 580 N (380–1,040 N) having the mean distance 6.9 mm (4.8–7.6 mm) above the normal value. However, their results were criticized by Dickman and Sonntag because of important flaws related to laboratory conditions of testing [8]. In their words, the testing force did not reflect the clinical mechanism and also, the elastic recoil normally oriented against the distraction was lost due to devoid muscular and soft tissue in the experiment. The other drawback of the “rule of Spence” was defined by Heller et al. [18]. Simply, the radiographic

Fig.  10.2  Overhang of C1 over C2 in summary larger than 8 mm suspicious of TAL disruption (“rule of Spence”)

Fig. 10.1  Types of atlas fractures according to Gehweiler. (a) Posterior arch fracture (exceptional single). (b) Anterior arch fracture. (c) Three piece fracture of both arches. (d) Comminuted fracture of lateral mass. (e) fracture extending to transverse process

153

10.2  Etiology

Fig.  10.3  Negative Spence measurement despite TAL abruption. (a) Coronal CT reconstruction analog to transoral radiogram showing no C1-overhang. (b) MRI clearly showing TAL attachment abruption in the same patient

magnification of +18% has to be calculated during evaluation of transoral pictures. They suggested to increase the critical summarized distance of both sides overhang measured on transoral images from 6.9 to 8.1 mm. Dickman et al. in their series of direct MRI assessment of TAL integrity has demonstrated that the “rule of Spence” would have missed 60% of transverse ligament disruptions (Fig. 10.3). They classified those disruptions into two subtypes: type I with pure ligament disruption; type II with an avulsion fracture of the attaching tubercle or the lateral mass comminution. They advocated an early surgical fixation of type I fractures due to subsequent C1-2 instability and poor potential of TAL matrix healing [6]. Recently, Bransford et al. added the sagittal split fracture of lateral mass to classification systems as a specific entity prone to nonunion with painful deformity sequelae in the long term [3]. This intra-articular fracture, in fact, completely disconnects the lateral part of C1 mass from the C1 ring but it does not influence the attachment and function of TAL (Fig. 10.4). The authors found 6 (11%) of previously described fractures in their series of 54 admitted to the hospital. Three of them died due to unrelated causes; however, in three surviving individuals they have demonstrated unsuccessful conservative treatment in rigid external supports (2 × rigid collar, 1 × halo-vest) resulting in painful deformity accompanied by cranial settling, craniolateral odontoid migration, and finally necessitating traction-reduction followed by surgical occipitocervical fusion.

Fig.  10.4  Sagittal split fracture of lateral mass described by Bransford. Note the avulsion fracture of condyle

10.2 Etiology Like in other spine injuries, the incidence of atlas fractures peaks in the second and third decades of life with almost twice male predominance. Vehicle accidents, falls, and miscellaneous other reasons often caused by either heavy object falling on the individual’s head or indirect axial head compression are the common causes of C1 fracture. Most frequently seen is an isolated fracture of the posterior C1 arch (approx 60% of all),which is nearly always bilateral and usually caused by hyperextension and axial load when the arch is compressed between occiput and spinous process of C2. The fracture almost always occurs through its thinnest part in the VA groove. It is obviously visible on lateral radiograph.

154

On lateral plain film also easily visible isolated fracture of anterior arch is usually caused by the direct impact of odontoid process or by hyperextension with counteraction of longus colli muscle attached to anterior tubercle. Jefferson was the first who hypothesized that the axial load transmitted via occipital condyles to the wedge-shaped lateral masses of atlas can cause their lateral dislocation followed by the ring fracture [22]. This “bursting” mechanism is still accepted. The classic Jefferson fracture as four-point fracture of the atlas is rare; however, two- or three-point variants are much more common [17]. These fractures result from axial loading and usually, are not associated with neurological injury. Different positions of head during the axial load (flexion, rotation, lateral bending, etc.) can be responsible for variations of fracture pattern. For example, asymmetric lateral mass burst fracture with or without concomitant asymmetric disconnection of both arches can be a result of axial load in lateral bending. The functional integrity of TAL is considered as the most critical factor determining the stability of C1 fractures [10, 13, 14, 26, 29, 39]. Transverse process fractures are caused almost exclusively by direct blunt impact and can be accompanied by VA tear or thrombosis [21]. Also, open injury caused by sharp object penetration, knife violence, and/or gunshot can involve the atlas [38, 39]. In such trauma, cervical CTA might be necessary to exclude an AV fistula or thrombosis.

10.3 Clinical Symptoms Published mortality rate related to atlas fracture differs depending on the source of analysis. If all trauma patients admitted to a hospital with the C1 fracture are analyzed retrospectively the mortality is as high as 30% [3], but if only those with isolated atlas fractures referred for a specific treatment are included, the mortality is almost zero, and neurological deficit is rarely detected [10, 14]. These observations ­confirming relative benignity of isolated C1 fractures are certainly influenced by eventual inclusion of polytrauma patients and/or those with combined UCS injuries in the survey. The clinical picture can be also modified by concomitant craniocerebral trauma in approximately 20% of patients [14]. Nevertheless, in the patients with isolated C1 injury, the clinical symptoms are usually nonspecific and any

10  Atlas Fractures

neurological deficit clearly related to C1 trauma is observed very rarely. Patients can complain of cervical spine tenderness or pain radiating to the occiput. Also, decreased sensation in occipital nerve region can be registered sometimes in combination with paravertebral reactive muscular spasms limiting the cervical spine motion. If the anterior arch is dislocated or prevertebral hematoma present, the patient can also have swallowing difficulties. Dislocated lateral masses and/or direct injury of vertebral foramen can cause symptoms of vertebrobasilar insufficiency. There are numerous case reports describing VA injuries associated with C1 fractures [3, 31, 40, 48];however, only one case where the C1 fracture caused bilateral VA obliteration [47]. This patient suffered from posterior fossa stroke but survived.

10.4 Diagnosis As mentioned previously, most atlas fractures are detected by spiral CT during admittance of acutely traumatized patient; however, if the patient presents on outpatient basis with nonspecific symptoms, then the lateral, AP, and transoral radiographs are performed as a first choice. Unfortunately, up to 25% of C1 fractures might be missed on plain radiographs [7, 14] and therefore, if any suspicion of UCS injury arise from plain films, CT always follows and thereafter, the MRI is eventually performed to evaluate the status of TAL and exclude neural structure compromise. The stability of atlantal fractures, sometimes hypothetic, although for further treatment decision crucial, can be confirmed by physician-guided flexion–extension skiascopy in cooperating patients. Despite that, it is not widely accepted since a dynamic CT or MRI can also be done for the same purpose with more accuracy today. Often, the first suspicion of instability comes from transoral images if the atlantoaxial joint lateral summarized overhang is more than 8  mm [18, 41]. This always leads to thin sliced CT imaging not only to exactly depict the fracture and bone dislocation extend but especially to see if the TAL tubercle is not detached (Fig. 10.5) as an indirect sign of TAL deficiency [6]. Also, some comminuted lateral mass fractures are not able to hold the ligamental tubercle avulsion (LTA) strength to fix the dens in correct position (Fig. 10.6). The most specific method to evaluate the LTA status is MRI (Fig. 10.6).

10.5  Treatment Strategy

Fig. 10.5  Unstable two part C1 ring fracture with detachment of TAL treated surgically with temporary lateral mass compression by custom made internal fixator fixed behind the C2 spinous process by the interconnecting rod. (a) Axial CT showing the fracture and TAL tubercle avulsion. (b) Perioperative

155

picture, note the rod behind the C2 spinous process. (c) Transoral postoperative view showing the achieved partial lateral mass compression. (d) Laterogram with internal fixator of C1 fixed to spinous process of C2 by the rod

Fig. 10.6  (a) axial MRI showing intact TAL, (b) axial scan at the different level demonstrating burst TAL attachment

10.5 Treatment Strategy Irrespective of the type of treatment, the goal of therapy in fractures of the atlas is to achieve bony healing, maintain atlantoaxial stability, and prevent any neurological or painful sequelae of nonunion or malunion

with good functional outcome. In the past, the treatment began with external bracing almost always [15, 39, 49]. The dislocations were reduced by traction, sometimes lasting more than 6  weeks, ­followed by hard external support with Minerva plaster or halocast and later, halo-vest [26, 29]. Most of the authors

156

Fig.  10.7  Pure ligamentous tear of TAL near to its insertion depicted on axial MRI

described successful outcome in nearly all cases with fusion rate round 95–100% and without delayed AA instability regardless of the TAL integrity status [10, 14, 24, 39]. Surgery was indicated only if the conservative approach failed. Due to contemporary technical development, the surgical fixation was often not stable enough and thus, uncomfortable halo-vest wearing continued for another 12 weeks after the operation [3, 26]. Dickman et al. started the era of more active surgical approach promoting to operate on all acute C1 fractures with MRI-proven traumatic intrasubstance disintegration of TAL (their type I) to prevent subsequent AA instability. They have also reported that LTA (their type II TAL injury) can heal conservatively in only 74% [6]. Segal et  al. [37] found positive relationship between the degree of fracture displacement and nonunion and noticed that nonunions only occurred in comminuted fractures involving the lateral mass with osteoperiosteal avulsion of the transverse ligament. It was also reported that those patients did not return to full level of activity and were classed as poor clinical outcome. As well others, including Jefferson’s initial review [22], concluded that poor functional outcomes may occur in 56–80% of conservatively treated patients when the articular surface of lateral mass is fractured and displaced [26, 29]. Usual argument is that, incongruity of articular surface is responsible for late pain and limited mobility in lateral mass fractures. Recently, Dvorak et al. retrospectively studied a group of 34 patients treated (91% conservatively)

10  Atlas Fractures

for isolated Jefferson type fractures [9]. They psychometrically compared the follow-up status with the normative and found out that the functional ability did not achieve the preinjury state of health approximated to the normal population. The patient’s functional status was much worse in those with lateral mass residual displacement more than 7 mm. Potential risks of all surgical treatments are clear; however, external immobilization techniques are not without their risks. Halo-vest may be associated with extra- or intracranial infections, while rigid collars and braces (Minerva, SOMI) may result in cutaneous ­ulceration or inadequate immobilization [11, 19, 30, 34, 45]. Although immobilization techniques can be adequate, 12 weeks in a halo-vest may not be acceptable to every patient especially if alternatives exist [43]. Moreover, the halo seems not to be superior in terms of rigidity to Philadelphia collar fixation in the UCS region [25, 36]. Traditionally, unstable atlas fractures have been treated surgically through a variety of different fusion techniques ranging from posterior onlay occipito­ cervical fusions through wiring or screw methods [6, 27, 35]. Currently, the modular screw systems for AA fixation with the ability to reduce atlantal fracture and/ or ­dislocation are the most effective surgical alternative [16, 44, 46]. Nevertheless, any AA fixation ­substantially reduces the UCS mobility and therefore, it is reasonable to look for better alternatives. Motionpreserving surgical treatment of isolated atlas fractures is not a novel idea. Böhm et al. recently presented eight patients with unstable Gehweiler Type III atlas fractures treated with an open direct osteosynthesis [2]. They reconstructed the C1 ring and avoided not only fusion to C2 or occiput but also postoperative immobilization. Another method of preserving the AA motion is, to fix the lateral masses together with construct introduced transorally [20, 33]. When associated with other fractures of the cervical spine, it is usually the concomitant injury (most frequently, odontoid fracture) that determines the type of treatment [13, 24].

10.6 Our Preference All our patients with suspected UCS injury have thin sliced CT and MRI to obtain maximal information about the “hard” and “soft” morphology of the injury. More sophisticated investigations like dynamic films, dynamic

157

10.7  Our Treatment Algorithm

MRI, MRA, and/or CTA are added if necessary for finetuning of the diagnosis or planning of surgery. We see isolated atlas fractures less frequently than C1-2 combination injuries. This can be caused by preferable admitting of patients referred for surgical intervention from other hospitals. Apart from frequent single arch fractures the four-part Jefferson fracture is extremely rarely seen; however, three- or two-part ring disconnection is more often admitted. In a similar frequency, we are encountering intraarticular fractures and lateral mass comminutions. We believe that intra-articular fracture extent, the joint incongruence caused by dislocation and the C1 ring functional disintegrity are the most important factors influencing the patient’s outcome in isolated C1 injuries. We have accepted the “Spence rule” as an orientation warning about TAL status. Nevertheless, the negative results seen on TO image should be confirmed by more precise methods (MRI, CT to see the tubercle, dynamic films) of TAL integrity investigation in all C1 ring disconnecting fractures. Generally, the stability of the fracture is dependent on the integrity of the C1 ring itself, in our opinion. The other fact supporting our conviction of C1 ring integrity importance is that the anterior transoral C1 laminectomy can lead to odontoid vertical migration in the long term because the disintegrated ring is not able to hold the weight of head and the wedge-shaped lateral masses are prone to displace laterally. Any shift or rotation of disconnected ring fragments can cause the AA joint incongruence. Moreover, the sagittal split of lateral mass described by Bransford et  al. [3] is, in fact, an unstable intra-articular fracture. In summary, we suppose that if the fracture disconnecting the C1 ring heals under influence of vertical force on wedge-shaped lateral masses, the displacement with joint incongruence can be the result. In such situations, the later onset of posttraumatic arthritis with consecutive pain and movement limitation can be anticipated. Jefferson fracture is not a “burst fracture” in the usual sense of the term but this is an atlas-bursting injury. There are three most important points in decision ­process choosing between surgery and conservative treatment of atlas fractures, in our opinion. First, to establish if the fracture is isolated or a part of combined USC injury. Most often, we can see the combination with odontoid process fracture and the C1 breach itself is not the key point to decide; however, if

it is unstable, more complex surgery can be involved (e.g., triple anterior screw). As second point, the stability of atlas fracture has to be evaluated, which means that the functional integrity of TAL has to be judged. We believe that also comminution of lateral mass cannot hold the TAL strength in normal values. As a third important point in our decision process, is the expected long-term result and patient’s satisfaction. Dislocated fractures not reducible with traction, intra-articular fractures (especially, comminuted) – in summary, all those who cannot heal conservatively with acceptable congruence of AA joint are, at least, candidates for surgical consideration. The busiest joint of the whole spine having healed in incorrect position can cause a whole life’s worth of problems for our patient.

10.7 Our Treatment Algorithm Depending on pain generated by soft tissue injury, the soft or hard collar should be stiff enough for successful treatment of isolated posterior arch fractures. Philadelphia collar for 12  weeks is sufficient to immobilize the anterior arch fractures but more ­frequent, radiological follow-up including dynamic films is necessary to monitor eventual posterior C1 displacement. In nondislocated fractures of both sides of the ring with intact TAL confirmed by MRI also, the hard collar is used as initial treatment. Nonetheless, the fracture can heal slowly (sometimes, more than 6 months) and this process has to be verified by CT and the dynamic films should confirm the AA stability. The same therapeutic regime can be used if the “Jefferson-like” arch disconnection has more than two fragments. Dislocated fractures without intra-articular extend with intact TAL can be reduced by traction. The weight of traction should be increased gradually starting on 2 kg and the reduction effect controlled by TO X-rays, or better by CT (with all the difficulties accompanying the traction during patient’s transport and positioning). If the fracture does not redislocate after gradual ­traction release, the Philadelphia collar with 14 days follow-up transoral radiographic control can be applied, in our opinion. Fractures of C1 lateral masses can be treated similarly to the previous group with hard collar if

158

10  Atlas Fractures

nondislocated (Fig. 10.8). However, if the fracture is dislocated or burst, the traction attempt can fail to effectively reduce the joint congruence. This can happen especially in sagittally oriented mass splits described by Bransford et  al. [3], where the craniocaudal force transmitted by the condyle does not allow the fracture reduction. We call this “axe effect.” Such fractures can be treated conservatively but only with hardly achievable long-term continuous distraction in SOMI brace or halo-vest. A similar problem is with lateral mass burst fracture. It can also be treated conservatively; however, the functional results are poor (Fig. 10.9). All conservatively treated patients are radiographically checked on a regular basis at 6 weeks, 3 months, 6  months, and 1  year. The first dynamic films are

performed at 6 weeks in presumably stable injuries but at 3 months in the others. Only CT can finally confirm the bony fusion. If the fracture is considered as suitable for surgical treatment the patient and/or his family are fully informed about the advantages and possible risks of operation and the hard external support is offered as an alternative. We always emphasize that, nearly, all isolated C1 fractures can be treated conserva­­tively (with exception of clear AA instability and ­documented TAL tear) according to literature and that there is no evidence supporting any decision available. Surprisingly, more patients choose the more aggressive approach believing that this allows them faster mobilization and more active life in the future. There is no doubt that those fractures causing direct pressure to neural structures are indicated for

Fig.  10.8  Non sagittal fracture of C1 lateral mass healed in Philadelphia collar, images (c, d) obtained 3 years after the initial treatment with acceptable clinical results (occasional headache). (a) Axial CT scan showing sagittal like pattern. (b)

Coronal reconstruction depicting that the fracture is not sagittally oriented. (c). axial CT scan 3 years after conservative treatment in hard collar. (d) coronal reconstruction showing healed fracture in “acceptable” AA joint congruence

10.7  Our Treatment Algorithm

159

Fig. 10.9  Comminuted lateral mass of atlas treated 12 weeks in halo-vest, images (c, d) obtained 4 years after the initial treatment, poor clinical result (pain in rotation, headache). (a) Axial CT scan showing the comminution of lateral mass. (b) Initial

coronal reformatted image. (c) Axial CT scan obtained 4 years after the treatment showing “healed” fracture. (d) Coronal reconstruction showing lateral mass deformity and important joint incongruence

decompression: however, such injuries are very rare. They can be seen as a result of direct localized blunt violence or as open injuries related to gun shots or sharp instrument penetration. This way the VA can easily be involved resulting in bleeding and/or thrombosis (Fig. 10.10). Also, the fractures with obvious AA instability and documented TAL tear should be operated by solid method of AA posterior fixation (Fig. 10.11). We ­prefer the transarticular screw AA fixation; however, in cases where the C1 fracture-dislocation manipulation can lead to its reduction, we prefer to use of Harms fixator (Fig.  10.12). Mostly, we supplement the previous fixation with posterior grafting.

In fractures with TAL tubercle avulsion temporary fixation, either with custom-made compression allowing device (Fig. 10.5) or with the Harms technique can be performed. Sagittal split fracture of lateral mass can be effectively treated with CT navigated percutaneous direct compressive osteosynthesis (Fig.  10.13) but open surgical approach allowing the fracture re­­ duction and fixation is also recommendable. In complex injuries, we favor the most important instability as it is described in Chap. 14. Always, we have to bear in mind that as few segments as possible should be fused in CVJ region, especially when it comes to the occipital bone extent of the construct that is often unreasonable.

160

Fig.  10.10  Gun shot with lateral mass destroyed but without VA injury treated with occipitocervical fusion. (a) Axial CT scan showing the antero-posterior pathway of the bullet through C1 lateral mass. (b) Frontal plane reconstruction showing

Fig. 10.11  Patient from Fig. 10.3 with coincidental atlas fracture with TAL incompetence and ­subaxial luxation fracture. (a) Laterogram showing the Harms fixator stabilizing C1– and 360 fixation of subaxial fracture-luxation. (b) lateral X-ray in flexion documenting the stability of the constructs 3 months after surgery

10  Atlas Fractures

l­ateral  mass destruction. (c) CTA confirming the VA patency. (d) Occipitocervical fusion, C2 transpedicular screw on the side of injury, short transarticular on the other side

References

161

Fig. 10.12  Three part fracture of the C1 ring with AA dislocation manipulated by Harms fixator to correct joint position (because of bilateral high riding VA the Wright’s modification with crosslaminar screw purchase was used to fix C2). (a) 3D preoperative image showing the right AA joint posterior displacement. (b) sagittal reconstruction documenting surgically achieved joint reduction

Fig. 10.13  Sagittal split fracture of C1 lateral mass with intact TAL on MRI treated with percutaneous CT guided compressive osteosynthesis (also in Chap. 7). (a) Initial axial CT scan. (b)  Preoperative coronal reconstruction. (c) Parasagittal scan

obtained 1.5 years after the surgery showing the renewed ­congruence of AA joint. (d) The screw position in C1 on plain laterogram

References

  3. Bransford, R., Falicov, A., Nguyen, Q., et al.: Unilateral C– lateral mass sagittal split fracture: an unstable Jefferson fracture variant. J Neurosurg Spine 10, 466–473 (2009)   4. Cooper, A.: A treatise on dislocations and fractures of the joints, pp. 570–576. Longman, Hurst, Rees, Orme, Brown and Cox, London (1823)   5. Dickman, C.A., Green, K.A.: Treatment of atlas fractures. In: Menezes, A.H., Sonntag, V.K.H. (eds.) Principles of spinal surgery, pp. 855–869. McGraw-Hill, New York (1996)

  1. Aebi, M., Nazarian, S.: Classification of injuries of the cervical spine. Orthopade 16, 27–36 (1987)   2. Bohm, H., Kayser, R., El Saghir, H., et al.: Direct osteosynthesis of instable Gehweiler Type III atlas fractures. Presentation of a dorsoventral osteosynthesis of instable atlas fractures while maintaining function. Unfallchirurg 109, 754–760 (2006)

162   6. Dickman, C.A., Greene, K.A., Sonntag, V.K.: Injuries involving the transverse atlantal ligament: classification and treatment guidelines based upon experience with 39 injuries. Neurosurgery 38, 44–50 (1996)   7. Dickman, C.A., Hadley, M.N., Browner, C., et  al.: Neurosurgical management of acute atlas-axis combination fractures. A review of 25 cases. J Neurosurg 70, 45–49 (1989)   8. Dickman, C.A., Sonntag, V.K.: Injuries involving the transverse atlantal ligament: classification and treatment guidelines based upon experience with 39 injuries. Neurosurgery 40, 886–887 (1997)   9. Dvorak, M.F., Johnson, M.G., Boyd, M., et al.: Long-term health-related quality of life outcomes following Jeffersontype burst fractures of the atlas. J Neurosurg Spine 2, 411– 417 (2005) 10. Fowler, J.L., Sandhu, A., Fraser, R.D.: A review of fractures of the atlas vertebra. J Spinal Disord 3, 19–24 (1990) 11. Frangen, T.M., Zilkens, C., Muhr, G., et al.: Odontoid fractures in the elderly: dorsal C1/C2 fusion is superior to halovest immobilization. J Trauma 63, 83–89 (2007) 12. Gehweiler, J.A., Osborne, R.L., Becker, R.F.: The radiology of the vertebral trauma. Saunders, Philadelphia (1980) 13. Greene, K.A., Dickman, C.A., Marciano, F.F., et al.: Acute axis fractures. Analysis of management and outcome in 340 consecutive cases. Spine (Phila Pa 1976) 22, 1843–1852 (1997) 14. Hadley, M.N., Dickman, C.A., Browner, C.M., et al.: Acute traumatic atlas fractures: management and long term outcome. Neurosurgery 23, 31–35 (1988) 15. Han, S.Y., Witten, D.M., Mussleman, J.P.: Jefferson fracture of the atlas. Report of six cases. J Neurosurg 44, 368–371 (1976) 16. Harms, J., Melcher, R.P.: Posterior C1-C2 fusion with polyaxial screw and rod fixation. Spine (Phila Pa 1976) 26, 2467–2471 (2001) 17. Hays, M.B., Alker Jr., G.J.: Fractures of the atlas vertebra. The two-part burst fracture of Jefferson. Spine (Phila Pa 1976) 13, 601–603 (1988) 18. Heller, J.G., Viroslav, S., Hudson, T.: Jefferson fractures: the role of magnification artifact in assessing transverse ligament integrity. J Spinal Disord 6, 392–396 (1993) 19. Horn, E.M., Feiz-Erfan, I., Lekovic, G.P., et al.: Survivors of occipitoatlantal dislocation injuries: imaging and clinical correlates. J Neurosurg Spine 6, 113–120 (2007) 20. Hu, Y., Ma, W., Xu, R.: Transoral osteosynthesis C1 as a function-preserving option in the treatment of bipartite atlas deformity: a case report. Spine (Phila Pa 1976) 34, E418–E421 (2009) 21. Jeanneret, B.: Combined fracture of anterior and posterior arch of atlas due to extreme lateral bending. Case report. In: Kehr, P., Weidner, A. (eds.) Cervical spine, pp. 246–253. Springer, Wien (1987) 22. Jefferson, G.: Fractures of the atlas vertebra: report of four cases and a review of those previously reported. Br J Surg 7, 407–422 (1920) 23. Jevtich, V.: Horizontal fracture of the anterior arch of the atlas. Case report. J Bone Joint Surg Am 68, 1094–1095 (1986) 24. Kesterson, L., Benzel, E., Orrison, W., et al.: Evaluation and treatment of atlas burst fractures (Jefferson fractures). J Neurosurg 75, 213–220 (1991)

10  Atlas Fractures 25. Koller, H., Zenner, J., Hitzl, W., et  al.: In vivo analysis of atlantoaxial motion in individuals immobilized with the halo thoracic vest or Philadelphia collar. Spine (Phila Pa 1976) 34, 670–679 (2009) 26. Landells, C.D., Van Peteghem, P.K.: Fractures of the atlas: classification, treatment and morbidity. Spine (Phila Pa 1976 13, 450–452 (1988) 27. Lee, T.T., Green, B.A., Petrin, D.R.: Treatment of stable burst fracture of the atlas (Jefferson fracture) with rigid cervical collar. Spine (Phila Pa 1976 23, 1963–1967 (1998) 28. Levine, A.M., Edwards, C.C.: Treatment of injuries in the C1-C2 complex. Orthop Clin North Am 17, 31–34 (1986) 29. Levine, A.M., Edwards, C.C.: Fractures of the atlas. J Bone Joint Surg Am 73, 680–691 (1991) 30. Majercik, S., Tashjian, R.Z., Biffl, W.L., et  al.: Halo vest immobilization in the elderly: a death sentence? J Trauma 59, 350–356 (2005). discussion 356–358 31. Muratsu, H., Doita, M., Yanagi, T., et al.: Cerebellar infarction resulting from vertebral artery occlusion associated with a Jefferson fracture. J Spinal Disord Tech 18, 293–296 (2005) 32. Proubasta, I.R., Sancho, R.N., Alonso, J.R., et al.: Horizontal fracture of the anterior arch of the atlas. Report of two cases and review of the literature. Spine (Phila Pa 1976) 12, 615– 618 (1987) 33. Ruf, M., Melcher, R., Harms, J.: Transoral reduction and osteosynthesis C1 as a function-preserving option in the treatment of unstable Jefferson fractures. Spine (Phila Pa 1976) 29, 823–827 (2004) 34. Saeed, M.U., Dacuycuy, M.A., Kennedy, D.J.: Halo pin insertion-associated brain abscess: case report and review of literature. Spine (Phila Pa 1976) 32, E271–E274 (2007) 35. Scharen, S., Jeanneret, B.: Atlas fractures. Orthopade 28, 385–393 (1999) 36. Schneider, A.M., Hipp, J.A., Nguyen, L., et al.: Reduction in head and intervertebral motion provided by 7 contemporary cervical orthoses in 45 individuals. Spine (Phila Pa 1976) 32, E1–E6 (2007) 37. Segal, L.S., Grimm, J.O., Stauffer, E.S.: Non-union of fractures of the atlas. J Bone Joint Surg Am 69, 1423–1434 (1987) 38. Sherk, H.H., Giri, N., Nicholson, J.T.: Gunshot wound with fracture of the atlas and arteriovenous fistula of the vertebral artery. Case report. J Bone Joint Surg Am 56, 1738–1740 (1974) 39. Sherk, H.H., Nicholson, J.T.: Fractures of the atlas. J Bone Joint Surg Am 52, 1017–1024 (1970) 40. Siegel, M., Alberts, R.: Unusual sign of a Jefferson fracture. A case report. Spine (Phila Pa 1976) 17, 605–607 (1992) 41. Spence Jr., K.F., Decker, S., Sell, K.W.: Bursting atlantal fracture associated with rupture of the transverse ligament. J Bone Joint Surg Am 52, 543–549 (1970) 42. Stewart Jr., G.C., Gehweiler Jr., J.A., Laib, R.H., et  al.: Horizontal fracture of the anterior arch of the atlas. Radiology 122, 349–352 (1977) 43. Strohm, P.C., Muller, ChA, Kostler, W., et al.: Halo-fixator vest – indications and complications. Zentralbl Chir 132, 54–59 (2007) 44. Stulik, J., Krbec, M.: Injuries of the atlas. Acta Chir Orthop Traumatol Cech 70, 274–278 (2003) 45. Tashjian, R.Z., Majercik, S., Biffl, W.L., et  al.: Halo-vest immobilization increases early morbidity and mortality in elderly odontoid fractures. J Trauma 60, 199–203 (2006)

References 46. Tessitore, E., Momjian, A., Payer, M.: Posterior reduction and fixation of an unstable Jefferson fracture with C1 lateral mass screws, C2 isthmus screws, and crosslink fixation: technical case report. Neurosurgery 63, ONSE100–ONSE101 (2008). discussion ONSE101 47. Walsh, G.S., Cusimano, M.D.: Vertebral artery injury associated with a Jefferson fracture. Can J Neurol Sci 22, 308– 311 (1995)

163 48. Weller, S.J., Rossitch Jr., E., Malek, A.M.: Detection of vertebral artery injury after cervical spine trauma using magnetic resonance angiography. J Trauma 46, 660–666 (1999) 49. Zimmerman, E., Grant, J., Vise, W.M., et al.: Treatment of Jefferson fracture with a halo apparatus. Report of two cases. J Neurosurg 44, 372–375 (1976)

Odontoid Process Fractures

11

P. Suchomel and L. Jurák

The presence of odontoid fracture (OF) was first described by Lambotte [56]; however, the first patient undergoing treatment of the fracture by delayed posterior surgical atlantoaxial fixation was reported by Mixter and Osgood [61]. Interestingly, despite being the most common UCS injury, OF has a less colorful background than the most frequently, historically mentioned, hangman’s fracture. Throughout the literature, the attention has always been drawn to the fact that significant amount of odontoid fractures are detected late after the injury and they are notoriously prone to nonunion. C1-2 instability caused by loss of restriction of translational AA movement is ­considered the most dangerous consequence of this frequent injury that can potentially result in fatal spinal cord damage. Historically the treatment ranged from conservative external immobilization [13, 33, 70, 74, 84] to surgical posterior AA fusion [17, 43, 95] usually done after failure of external immobilization, i.e., pseudoarthrosis phase. The development of anterior screw fixation by Nakanishi [65] and independently by Magerl [39] added another and, in fact, the most physiological ­surgical treatment option. Currently, odontoid process fractures are diagnosed immediately after the injury and modern imaging techniques certainly facilitate the decision as to which treatment option is the most appropriate for our patients.

11.1 Classification The earliest attempts to classify OF distinguished only two types of fractures: those at the base and those at

P. Suchomel and L. Jurák Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova st. 10, 46063 Liberec, Czech Republic

the neck of the process [19]. Schatzker et al. proposed to classify OFs into two categories depending on the location of fracture either above or below the accessory ligaments [80]. Another, more accepted classification, focused on the fracture site stability was suggested by Roy-Camille [72]. Fractures were divided into subtypes based on the direction of the fracture line on plain lateral and dynamic films. OF was considered to be unstable if displaced at presentation or if dislocation was identified on dynamic films. The classification comprised three fracture line patterns (shortenings derived from French terminology): OBAV – fracture line slopes forward anteriorly with or without anterior displacement; OBAR – fracture line slopes obliquely backward with or without posterior dislocation; and HTAL – horizontal fracture line with or without displacement in any direction. Althoff et al. [3] proposed a scheme where Type A fracture passes through the neck of the odontoid; Type B fracture through the rostral part of the C2 body and Type C through the body of C2 but also the medial aspect of one of the C2 superior articular facets. Type D injury then involved both C2 upper articular processes. The classification most frequently used today is that of Anderson and D’Alonzo [6]. It is based on the location of the fracture line (Fig. 11.1). Type I stands for an oblique fracture of the tip of the dens above the transverse atlantal ligament (TAL). It is a very rare injury commonly considered as stable [15, 36, 83]. Type II is a fracture of the base of the odontoid process. This most frequent subtype of OF is highly unstable and very much prone to nonunion. Type III is represented by a fracture of the dens base extending more or less into the C2 body. Using this classification, other authors found various subtypes of the odontoid neck injury. Hadley et  al. [40] described a Type IIA comminuted fracture of the odontoid base

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_11, © Springer-Verlag Berlin Heidelberg 2011

165

166

11  Odontoid Process Fractures

Fig. 11.1  A schematic drawing of Anderson D’Alonso classification of odontoid fracture

with associated free fracture fragments. Its incidence was estimated as 5% of all Type II fractures. Geisler et al. [30] suggested classifying posteriorly displaced fractures as II-P. Analyzing our series of patients, we proposed to separate transverse odontoid process ­fractures that were above the base but below the transverse ligament as Type IIT, where T means transverse or transitional [91]. Type IIT fracture is unstable in all directions, particularly in rotation and therefore, requires a two-screw anterior osteosynthesis rather than a single-screw technique popularized recently. Gauer et al. proposed a treatment based sub-classification of Anderson Type II fractures [34]. Nondisplaced transverse fracture with no comminution suitable for conservative treatment was classified as Type IIA. Type IIB was assigned to displaced transverse or ­posteriorly oblique fracture that was amenable to anterior screw fixation following fracture reduction. Anteriorly oblique fracture line or a fracture with ­significant comminution was the classified as Type IIC. This type, according to author’s preference, is predetermined for surgical posterior atlantoaxial stabilization. Vertical OF is also described in the literature [51]. It may be considered as stable if the transverse ligament is not involved. Similarly to others, we have adopted the Anderson D’Alonso classification system and therefore, their denomination of fracture types will be used in the remainder of the text.

11.2 Etiology and Epidemiology The fracture of C2 odontoid process represents 50–60% of all fractures of the axis [16, 28, 36] and 8–15% of all cervical acute spine fractures [2, 6, 16, 36, 41]. OF is the most common cervical spine fracture in adults over the age of 70 years [64, 73] and it is the most frequent fracture of all spine injuries in population aged over 80 years [73]. Type II is the most frequent form present in 37–83% of all odontoid fractures with even higher incidence of 95% in the elderly [16, 36, 64, 93]. Odontoid fractures are associated with other spine injuries in 34% of patients, of which, 85% are cervical and 20% are associated with C1 injury [36]. Concurrent TAL malfunction due to abruption of its attachment has also been described [20, 36]. Similar AA instability can also result from concomitant C1 ring disintegration in combined C1-2 fractures. The association of head injury and all C2 fracture subtypes was seen in 20.3% of cases [36]. OF can be caused by hyperflexion with possible anterior dislocation and AA subluxation with or without transverse ligament damage or, more commonly, by hyperextension with concomitant C1 anterior arch fracture and/or posterior displacement of C1. As the majority of OFs are caused by motor vehicle accidents or simple falls [2, 6, 42, 62, 67, 91], the mechanism of the dens fracture is usually not caused by pure sagittal force transmission but is often modified by lateral bending and rotational forces.

11.4  Radiology

11.3 Clinical Symptoms Between 25 and 40% of patients with UCS injury die at the scene of the accident; however, approximately 90% of surviving patients have no major neurological deficit [11, 16, 36, 42, 96]. Fractures of the dens were frequently missed in the past. Difficulty in obtaining adequate lateral and transoral plain films in the acute setting was the main reason for the delay, as hospitaladmitted trauma survivors were often uncooperative and/or unconscious due to the associated head trauma, multiple injuries or intoxication. In the late 1980s, we published that 60% of acutely admitted patients with head injury are under the influence of alcohol [90]. The diagnostic difficulty is currently eliminated by an early and mandatory CT evaluation of all uncooperative and unconscious patients with history of trauma. The cooperating patient usually complains only of poorly localized pain in the posterior part of the neck and has paravertebral muscle spasm, tenderness, and limited movement of the neck. Neurological symptoms and signs vary from a rather rare pentaplegia to the more frequently seen simple occipital neuralgia with limited neck motion.

167

hospitals without 24  h CT service have a standard algorithm of radiological workup based on clinical situation. Therefore, plain films in simple lateral and AP projection are obtained first and if any UCS abnormality is suspected, transoral views are added. Dislocated fractures are, usually, easily seen and identified (Fig. 11.2). However, hairline nondislocated fractures can be overlooked (Fig.  11.3). Classical tomography was historically also a good tool to confirm the presence of OF (Fig. 11.4). Only CT in bone windows with sagittal and coronal, or even better, 3D reconstructions, exactly delineates fracture location, its direction and extent, as well as bone morphology for potential surgical fixation (Fig. 11.5). Spiral CT

11.4 Radiology As mentioned above, most major trauma victims today are usually screened with an early, fast spiral CT. However, there are self-presenting ambulatory patients with minimal symptoms that harbor an ­odontoid fracture. Such patients are not commonly screened by a CT at the first instance. Also, smaller

Fig. 11.3  Non dislocated Type II odontoid fracture. (a) Transoral film of a hairline Type II fracture which was initially overlooked. (b) Odontoid pseudoarthrosis in the same patient 6 month later (no treatment until then)

Fig. 11.2  Lateral radiograph of posteriorly displaced odontoid Type II fracture

168

11  Odontoid Process Fractures

Fig. 11.4  Odontoid Type II fracture seen on sagittal classical tomogram

Fig. 11.6  MRI depicting the lateral tear of the transverse atlantal ligament (TAL)

Fig. 11.5  3D CT showing the odontoid Type II fracture

scanning also confirms or excludes any other associated spinal fractures. MRI evaluation is nonurgent in neurologically intact patients but is helpful in evaluation of the soft tissues, and especially the integrity of TAL (Fig. 11.6) [20].

11.5 Treatment Strategy Holdsworth [46] stated: “any classification has a sense if it can influence our therapeutic decision.” From this point of view, the most widely accepted classification

of Anderson and D’Alonso has some limitations. Generally, it is necessary to define which fractures are unstable and would thus eventually require surgical intervention. Most of the authors accept that Type I and III fractures are stable enough to allow conservative treatment with external bracing [2, 6, 7, 42, 62]. However, the “gray zone” exists between Type II and III fractures. The so called “high, rostral or shallow” Type III fractures should only include fractures that extend into the superior C2 articular surface. However, those that involve rostral C2 body but do not extend into the articular surfaces should be classified as Type II, as elegantly addressed by Grauer et al. [34]. Another classification issue is the caudal extent of the fracture, i.e., what distinguishes whether a fracture is a Type III OF or a fracture of the body (Fig. 11.7). Furthermore, there is also a wide range of different morphologies of Type II fracture patterns, not involved in the original classification, which can substantially influence the treatment decision. Fracture site, comminution, obliquity, and/or dislocation can all be of importance. Historically, most of the pioneers treated delayed fracture presentation. First procedure done for odontoid fracture is credited to Mixter who performed posterior AA wire and graft fusion, in fact, for odontoid pseudoarthrosis [61]. Generally, all types of odontoid fractures were primarily treated conservatively in the past. Following traction reduction of dislocated fractures, hard external orthoses (Minerva jackets, SOMI braces, and hard

11.5  Treatment Strategy

Fig. 11.7  Coronal plane CT showing a horizontal fracture of C2 body which should not be considered as odontoid fracture

collars) were used to stabilize the UCS. The halo-vest was popularized later due to its supposedly higher rigidity [16, 23, 36]. Predominantly, posterior surgical intervention was mostly reserved for fractures that had failed conservative therapy. Many modifications of the Gallie technique have been developed to achieve stable posterior AA fusion [12, 17, 22, 29, 37]. The techniques are described in detail in Chap. 6. All posterior AA fusion techniques were initially supplemented by halo-vest fixation thus adding further stress to the patient and it was not clear if healing was achieved due to the fusion, external immobilization or both. However, with the introduction of transarticular technique of Magerl, immediate AA stability could be achieved by posterior rigid screw techniques and thus lead to much improved fusion rates and clinical results [17, 21, 31, 32, 38, 44, 49, 59, 60]. Nonetheless, any atlantoaxial fusion substantially limits the cervical spine rotation (50%) and therefore, does not represent an ideal solution to the problem. The first attempt to approach the fracture directly was reported by Estridge and Smith [24]. They followed the idea of Fang and Ong [25] who fixed the odontoid pseudoarthrosis by intra-articular C1-2 grafting performed transorally. They directly refreshed the fracture site and implanted vertically oriented autologous iliac crest bone graft. Their patient fused in a Minerva jacket but, unfortunately, successfully committed suicide a year later. Interestingly, a firm bony fusion was confirmed at autopsy in this case. The development of direct screw compressive osteosynthesis by Nakanishi and Magerl was the real

169

breakthrough in the philosophy of the odontoid fracture treatment. This technique is suitable for Type II and shallow Type III fractures and represents the most physiological approach directly targeting the pathology and theoretically not influencing adjacent segment mobility [2, 10, 52, 88, 91]. It has been proven that without any treatment, odontoid fracture have close to 100% nonunion rate [16]. However, given the number of surgical options for odontoid fractures, there is no evidence clearly favoring one particular treatment strategy over the other, including both conservative and surgical methods [7]. In displaced fractures, traction reduction should always precede the final immobilization, whatever it may be [52, 94]. As in the past, patients with the rare Type I OF with no other associated UCS injury (e.g.: AOD) are treated conservatively with collar or more rigid external braces including halo-vest [62, 70, 83]. The majority of the Type III fractures are successfully treated in external braces with expected fusion rates of approximately 87–100% [15, 16, 28, 52, 94]. For this purpose, more surgeons recommend halo-vest immobilization [7, 36] although reports of successful hard collar treatment of Type III fracture were published [55, 63, 70]. Halo-vest fixation was recently criticized because it offers no advantage of a more solid immobilization than a rigid cervical collar and increases the rate of complications, especially in the older population [55, 82]. Frequent complications (26–66%) related to wearing a halo-vest are not benign and can include pressure sores, pin infection, pin loosening, fracture correction loss and, in the worst case scenario, breathing problems and pneumonia [27, 47, 76]. Fatal cardiopulmonary complications resulting in cardiac arrest have been reported in the elderly [92]. In some series of patients treated for odontoid fractures, the halo-vest related mortality was much higher than that for Philadelphia collar and/or surgical treatment [27, 87, 92]. Strohm et  al. [87] also noted that 58% of his patients judged the halo to be intolerable. The most commonly discussed topic in odontoid fractures nowadays is the treatment of Type II and “shallow” Type III fractures. These injuries are highly unstable and external immobilization fails to create bony union, on an average, in 30–50% of patients [36, 54, 57, 84]. Even further, if patients older than 60 years are included in the series, the failure rate dramatically increases up to 77–86% [36, 75].

170

Nonetheless, halo-vest or even hard collar immobilization is still accepted in certain nondisplaced and stable Type II fracture scenarios [36, 63, 69, 70], although the majority of modern surgeons [1, 2, 8, 18, 66, 88, 91, 95] prefer early surgical stabilization in acute forms. The reasons surrounding the high frequency of nonunion following conservative treatment of Type II fractures are also the subject of considerable debate. Some authors propose that there is a weak vascular watershed zone with poor blood supply in the odontoid neck whereas others believe that the odontoid neck is a place of enormous load transmission in the location of structurally poor bone [4, 5, 81]. The following factors have been shown to negatively influence healing of Type II fractures: fracture dislocation greater than 6 mm, angulation greater than 10°, fracture site comminution, osteoporosis, age over 60 years, delayed treatment, and loss of fracture alignment during the follow-up [9, 42, 58, 71, 80]. Despite its obvious logical advantage, the direct anterior screw osteosynthesis does have its opponents who prefer posterior AA fusion, especially in the elderly [14, 64]. The reported fusion rate of Type II fracture treated with anterior screw is 80–96% and for shallow Type III even higher being near 100% in most cases [7, 8, 52, 88, 89, 91]. Different techniques were developed to achieve anterior compressive osteosynthesis, as described in Chap. 6. Initially, fully threaded 3.5 mm steel screws were used [10] with the necessity of proximal canal overdrilling to achieve fracture compression. Later, partially threaded titanium alloy cannulated screws were introduced along the guiding Kirschner wire [2]. Apfelbaum advocated the use of 4 mm noncannulated screws as a stronger option [8]. Knöringer proposed a double-threaded screw (similar to Herbert screw) with self-compressive property [53]. Different types of approach instruments were subsequently developed to minimize morbidity of the surgical approach [45, 85]. Special plates introduced via a high anterolateral approach to fix comminuted, oblique, and delayed fractures were suggested with the aim to spare atlantoaxial movement [68, 86]. Currently, there is still ongoing discussion if one or two anterior screws should be used for fracture stabilization. Originally, it was proposed to introduce the first screw as compressive and the second one to stabilize against the rotational forces [2, 10]. However, bio­ mechanical studies confirmed similar strength for one

11  Odontoid Process Fractures

or two screw construct [35, 78]. This finding was supported by excellent clinical results documenting up to 95% fusion rate using only one screw for fixation [26, 48, 50, 79, 89]. Although far less physiological, the posterior atlantoaxial fixation does indeed have its place in the treatment of odontoid fractures, especially when anterior osteosynthesis is not possible or when pseudoarthrosis already exists [8]. Short neck, barrel chest, hyperkyphotic cervical spine, comminuted fracture site, nonreducible dislocation, certain combined C1-2 injuries, and TAL deficiency can represent relative contraindications to the anterior procedure.

11.6 Our Preference Similar to other CVJ traumatic injuries, all our patients with suspected odontoid process injury undergo mandatory spiral CT with 3D reconstructions. We are also convinced that MRI should be done in the first 24  h even in patients without neurological deficit. In those with neurological compromise, MRI evaluation should be undertaken on an emergent basis. MRI is valuable in evaluation of the spinal cord status and the integrity of transverse ligament. It can also exclude other soft tissue injuries – for example, a disk prolapse. (Fig. 11.8). In selected and cooperating patients, flexion – extension films performed under physician guidance can reveal potential instability in Type III fractures and/or other level of concomitant injury, not immediately detectable on static images (Fig. 11.9). Despite 15 years of experience with UCS injuries and surgical treatment of more than 90 patients with odontoid fracture, we have never encountered a Type I odontoid fracture. Nonetheless, even Type I injury could be unstable in the presence of bilateral apical ligament disruption or combined injury with occipital condyle fractures. In those situations, when initial attempt of conservative treatment with standard hard cervical collar is undertaken, it is essential to confirm CVJ stability radiographically at the end of follow-up as one can discover a potentially dangerous, initially reduced AOD. Transverse ligament damage is, in our experience, rarer than presumed in the literature [20]. In the era of lower quality MRI, most of our patients with Type II fractures were treated with anterior screws without the

11.6  Our Preference Fig. 11.8  Odontoid Type II fracture accompanied by a disk prolapse at the C3/4 level. (a) Sagittal MRI in T2 sequence. (b) Postoperative lateral radiograph showing single screw odontoid fixation performed simultaneously with graft and plate C3/4 fusion

Fig. 11.9  (a) Patient with instability revealed during physician-guided flexion and extension films a. lateral film obtained at admission. (b) Flexion revealed odontoid and C2 ring fracture. (c) Sagittal CT reconstruction showing combination of hangman’s type and odontoid Type III fracture. (d) Double screw fixation of odontoid with concomitant graft and plate C2/3 fusion

171

172

exact knowledge of TAL status. No single case of AA instability was identified on dynamic films performed routinely at 6 weeks after procedure (Fig. 11.10). We do not typically operate on nondisplaced Type III OFs and even more so, we believe that if conservative treatment is selected, a hard cervical collar (e.g., Philadelphia) is sufficient to allow successful healing of this type of fracture if isolated. We use a halo-vest fixation only exceptionally in certain combined UCS injuries and in situations where surgical fixation does not necessarily guarantee sufficient stability. Due to

Fig. 11.10  Dynamic films 6 weeks after double-screw fixation of Type II odontoid fracture confirming AA stability. (a) Lateral film in flexion and (b) extension

Fig. 11.11  Halo-vest fixation was unable to maintain odontoid Type II fracture alignment. (a) Reduced fracture before patient’s discharge. (b) Posterior dislocation registered at 1 month check up

11  Odontoid Process Fractures

our repeated experience of highly morbid halo vests, we never select this option as a first choice treatment in Type II or shallow Type III fractures (Fig. 11.11). We select a direct anterior compressive osteosynthesis for almost all Type II and “shallow” Type III fractures (especially if the AA joint is affected (Fig. 11.12). Also, Type III injuries that are dislocated and/or unstable on dynamic films can be treated with anterior screws if there is enough bone available at the base of the C2 body. In our opinion, as long as the previously described condition of sufficient bone “stock”

11.6  Our Preference

173

Fig. 11.12  CT reconstruction of “shallow” Type III fracture with left AA joint involvement. (a) Reconstruction in coronal plane. (b) Sagittal reconstruction showing the anteriorly oblique pattern. (c) Anterior view of 3D reconstruction

Fig. 11.13  The same patient as Fig. 11.12, treated with a double anterior screw fixation with K-wire safeguarding against anterior redislocation. (a) First, the K-wire introduced through reduced

odontoid fracture on fluoroscopical view. (b) Lateral fluoroscopical view documenting the K-wire protective function during first odontoid screw tightening. (c) Final result shown on transoral film

is respected, even anteriorly oblique fractures do not represent a contraindication to the anterior screw fixation. The described unintentional anterior dislocation can be simply prevented with parallelly-introduced K-wires (Fig.  11.13). However, if the anterior procedure is not possible (barrel chest, fracture site comminution etc.) or failed, we perform a salvage posterior fixation. The posterior approach of choice for us is the solid screw fixation according to Magerl or Goel-Harms technique of atlantoaxial fixation. Prior to scheduled anterior procedure, all dislocated fractures are reduced with halo-ring traction with its continued application during the operation if necessary (Fig.  11.14). We prefer the use of four-point fixation ring instead of Crutschfield, Barton or Gardner-Wells tongs because it is easy to attach, safe, and forces can be applied in any direction. Different weights can be used to achieve reduction, starting with 2  kg,

but frequently reaching much higher values during reduction maneuvers in experienced hands (20  kg on one occasion). If a fracture dislocation cannot be reduced, either posterior fixation with an attempted reduction or fixation and fusion in-situ are the options. If cord compression exists on MRI, a transoral decompression may then be necessary. Theoretically, also, a vice versa approach is possible. In general, we prefer to introduce two screws. The limiting factors for that maneuver are a thin odontoid process unable to accommodate two 4  mm shortthreaded screws or a technical error made during the drilling of the first hole, not allowing sufficient space for a second screw placement. We only consider a singlescrew construct as a bailout option if two are impossible and only when treating broad-based Type II and III fractures. We have repeatedly seen odontoid peg rotation

174

11  Odontoid Process Fractures

Fig. 11.14  Type II odontoid posterior dislocation reduced by traction. (a) Lateral film depicting the initial dislocation. (b) Lateral view of fracture reduced by traction. (c) Double screw fixation

during final tightening of a single screw when treating transverse fractures involving the narrowest part of the dens. Such situations can be corrected by using a double-screw construct or again with a parallel K-wire blocking the rotation during tightening. This is a

Fig. 11.15  Odontoid Type IIT fracture treated with a double-screw anterior fixation. (a) Preoperative sagittal CT reconstruction. (b) Preoperative coronal CT reconstruction. (c) Postoperative AP film. (d) Postoperative lateral film depicting correct length of screws

“conditio sine qua non” in Type IIT fractures where simple apical cortex drill penetration can be difficult due to its rotational instability (Fig. 11.15). The basic prerequisite for a successful endosteal bone formation is close contact and the best is compression of

11.6  Our Preference

175

the fracture surface accompanied by adequate immobilization [81]. To achieve this goal, the lag screw has to penetrate the apical odontoid cortex or must be short enough to adequately fulfill the lag principle. Postoperatively, if the fracture is deemed sufficiently stable, hard cervical orthosis (e.g., Philadelphia collar) is used for the same purpose as stated by Sandler et al. [77]: “not to move the head too far and too quickly,” for 6 weeks. Certainly, any part of our treatment paradigm or all of it can be influenced by coexistence of other and more complex UCS traumatic injuries. In combination fractures, one has to primarily address the major instability (usually AA instability) first. Atlas fractures are the most frequent accompanying injuries. If posterior C1 arch is broken, odontoid fracture stabilization has the priority. However, if the anterior arch of C1 is broken or the entire C1 ring is disconnected, then one has to consider the degree of AA instability in the treatment paradigm. Most frequently, direct anterior odontoid osteosynthesis followed by longer term use of a Philadelphia collar is sufficient. However, in cases of marked AA dislocation and/or MRI-proven transverse ligament damage, anterior triple screw fixation and fusion (odontoid and two anterior transarticular atlantoaxial screws) can be chosen. A solid posterior atlantoaxial fixation needs to be always considered as an option for these injuries.

Not infrequently, one has to treat concomitant s­ ubaxial cervical spine injuries. If this is the case, a combined surgical approach may be required and is best performed during a single session (Fig. 11.16). Given the high complication rate of halo-vest external fixation and the fact that it does not surpass Philadelphia collar in UCS immobilization in experimental studies [55, 82], we logically conclude that halo-vest fixation should be avoided and either surgical intervention or hard cervical collar should be used in the treatment of odontoid fractures. This is especially true in old, unconscious and/or chronically ventilated individuals. In this group of patients, we always prefer surgical stabilization, co-morbidities permitting. Early mobilization and ease of care can save their lives. All patients are followed up regularly at 6 weeks; and then at 3, 6, and 12  months. Lateral dynamic films are performed regularly until fracture healing is confirmed. A healed fracture can nowadays only be proven by CT documenting bridging bone across the fracture site. However, in the elderly, a stable fibrous union can represent a functional and acceptable result. We believe, however, that in young and active patients who do not demonstrate fracture union by one year, a posterior AA fixation and fusion should be performed.

Fig.  11.16  Odontoid fracture with C3 coronal split. (a) Preoperative sagittal CT reconstruction. (b) Postoperative lateral film ­showing simultaneous odontoid fixation with graft and plate

C2-4 fusion. (c)  Sagittal CT reconstruction depicting correct length of screws

176

References   1. Aebi, M.: Surgical treatment of upper, middle and lower cervical injuries and non-unions by anterior procedures. Eur Spine J 19(Suppl 1), S33–S39 (2009)   2. Aebi, M., Etter, C., Coscia, M.: Fractures of the odontoid process. Treatment with anterior screw fixation. Spine (Phila Pa 1976) 14, 1065–1070 (1989)   3. Althoff, B.: Fracture of the odontoid process. An experimental and clinical study. Acta Orthop Scand Suppl 177, 1–95 (1979)   4. Althoff, B., Goldie, I.F.: The arterial supply of the odontoid process of the axis. Acta Orthop Scand 48, 622–629 (1977)   5. Amling, M., Posl, M., Wening, V.J., et al.: Structural heterogeneity within the axis: the main cause in the etiology of dens fractures. A histomorphometric analysis of 37 normal and osteoporotic autopsy cases. J Neurosurg 83, 330–335 (1995)   6. Anderson, L.D., D’Alonzo, R.T.: Fractures of the odontoid process of the axis. J Bone Joint Surg Am 56, 1663–1674 (1974)   7. Anonymous: Isolated fractures of the axis in adults. Neurosurgery 50, S125–S139 (2002)   8. Apfelbaum, R.I., Lonser, R.R., Veres, R., et al.: Direct anterior screw fixation for recent and remote odontoid fractures. Neurosurg Focus 8, 1–10 (2000)   9. Apuzzo, M.L., Heiden, J.S., Weiss, M.H., et al.: Acute fractures of the odontoid process. An analysis of 45 cases. J Neurosurg 48, 85–91 (1978) 10. Bohler, J.: Anterior stabilization for acute fractures and nonunions of the dens. J Bone Joint Surg Am 64, 18–27 (1982) 11. Bohlman, H.H.: Acute fractures and dislocations of the cervical spine. An analysis of three hundred hospitalized patients and review of the literature. J Bone Joint Surg Am 61, 1119–1142 (1979) 12. Brooks, A.L., Jenkins, E.B.: Atlanto-axial arthrodesis by the wedge compression method. J Bone Joint Surg Am 60, 279– 284 (1978) 13. Bucholz, R.W.: Unstable hangman’s fractures. Clin Orthop Relat Res 154, 119–124 (1981) 14. Campanelli, M., Kattner, K.A., Stroink, A., et al.: Posterior C1-C2 transarticular screw fixation in the treatment of ­displaced type II odontoid fractures in the geriatric population – review of seven cases. Surg Neurol 51, 596–600 (1999). discussion 600–591 15. Chiba, K., Fujimura, Y., Toyama, Y., et al.: Treatment protocol for fractures of the odontoid process. J Spinal Disord 9, 267–276 (1996) 16. Clark, C.R., White 3rd, A.A.: Fractures of the dens. A multicenter study. J Bone Joint Surg Am 67, 1340–1348 (1985) 17. Coyne, T.J., Fehlings, M.G., Wallace, M.C., et  al.: C1-C2 posterior cervical fusion: long-term evaluation of results and efficacy. Neurosurgery 37, 688–692 (1995). discussion 692–683 18. Dailey, A.T., Hart, D., Finn, M.A., et al.: Anterior fixation of odontoid fractures in an elderly population. J Neurosurg Spine 12, 1–8 (2010) 19. de Mourgues, G., Fischer, L., Comtet, J.J., et al.: Fractures of the odontoid process of the axis: a series of 80 fractures. Acta Orthop Belg 38, 137–146 (1972)

11  Odontoid Process Fractures 20. Dickman, C.A., Mamourian, A., Sonntag, V.K., et  al.: Magnetic resonance imaging of the transverse atlantal ligament for the evaluation of atlantoaxial instability. J Neurosurg 75, 221–227 (1991) 21. Dickman, C.A., Sonntag, V.K.: Posterior C1-C2 transarticular screw fixation for atlantoaxial arthrodesis. Neurosurgery 43, 275–280 (1998). discussion 280–271 22. Dickman, C.A., Sonntag, V.K., Papadopoulos, S.M., et al.: The interspinous method of posterior atlantoaxial arthrodesis. J Neurosurg 74, 190–198 (1991) 23. Dunn, M.E., Seljeskog, E.L.: Experience in the management of odontoid process injuries: an analysis of 128 cases. Neurosurgery 18, 306–310 (1986) 24. Estridge, M.N., Smith, R.A.: Transoral fusion of odontoid fracture. Case report. J Neurosurg 27, 462–465 (1967) 25. Fang, H.S.Y., Ong, G.B.: Direct anterior approach to the upper cervical spine. J Bone Joint Surg Am 44, 1588–1604 (1962) 26. Fountas, K.N., Kapsalaki, E.Z., Karampelas, I., et al.: Results of long-term follow-up in patients undergoing anterior screw fixation for type II and rostral type III odontoid fractures. Spine (Phila Pa 1976) 30, 661–669 (2005) 27. Frangen, T.M., Zilkens, C., Muhr, G., et al.: Odontoid fractures in the elderly: dorsal C1/C2 fusion is superior to halovest immobilization. J Trauma 63, 83–89 (2007) 28. Fujii, E., Kobayashi, K., Hirabayashi, K.: Treatment in fractures of the odontoid process. Spine (Phila Pa 1976) 13, 604–609 (1988) 29. Gallie, W.E.: Fractures and dislocations of the cervical spine. Am J Surg 46, 495–499 (1939) 30. Geisler, F.H., Cheng, C., Poka, A., et  al.: Anterior screw fixation of posteriorly displaced type II odontoid fractures. Neurosurgery 25, 30–37 (1989). discussion 37–38 31. Gluf, W.M., Schmidt, M.H., Apfelbaum, R.I.: Atlantoaxial transarticular screw fixation: a review of surgical indications, fusion rate, complications, and lessons learned in 191 adult patients. J Neurosurg Spine 2, 155–163 (2005) 32. Goel, A., Laheri, V.: Plate and screw fixation for atlantoaxial subluxation. Acta Neurochir (Wien) 129, 47–53 (1994) 33. Govender, S., Grootboom, M.: Fractures of the dens – the results of non-rigid immobilization. Injury 19, 165–167 (1988) 34. Grauer, J.N., Shafi, B., Hilibrand, A.S., et al.: Proposal of a modified, treatment-oriented classification of odontoid fractures. Spine J 5, 123–129 (2005) 35. Graziano, G., Jaggers, C., Lee, M., et  al.: A comparative study of fixation techniques for type II fractures of the odontoid process. Spine (Phila Pa 1976) 18, 2383–2387 (1993) 36. Greene, K.A., Dickman, C.A., Marciano, F.F., et al.: Acute axis fractures. Analysis of management and outcome in 340 consecutive cases. Spine (Phila Pa 1976) 22, 1843–1852 (1997) 37. Griswold, D.M., Albright, J.A., Schiffman, E., et al.: Atlantoaxial fusion for instability. J Bone Joint Surg Am 60, 285– 292 (1978) 38. Grob, D., Jeanneret, B., Aebi, M., et al.: Atlanto-axial fusion with transarticular screw fixation. J Bone Joint Surg Br 73, 972–976 (1991) 39. Grob, D., Magerl, F.: Operative stabilisierung bei frakturen von C1 und C2. Orthopäde 16, 46–54 (1987)

References 40. Hadley, M.N., Browner, C.M., Liu, S.S., et al.: New subtype of acute odontoid fractures (type IIA). Neurosurgery 22, 67–71 (1988) 41. Hadley, M.N., Browner, C., Sonntag, V.K.: Axis fractures: a comprehensive review of management and treatment in 107 cases. Neurosurgery 17, 281–290 (1985) 42. Hadley, M.N., Dickman, C.A., Browner, C.M., et al.: Acute axis fractures: a review of 229 cases. J Neurosurg 71, 642– 647 (1989) 43. Hanigan, W.C., Powell, F.C., Elwood, P.W., et al.: Odontoid fractures in elderly patients. J Neurosurg 78, 32–35 (1993) 44. Harms, J., Melcher, R.P.: Posterior C1-C2 fusion with polyaxial screw and rod fixation. Spine (Phila Pa 1976) 26, 2467–2471 (2001) 45. Hashizume, H., Kawakami, M., Kawai, M., et al.: A clinical case of endoscopically assisted anterior screw fixation for the type II odontoid fracture. Spine (Phila Pa 1976) 28, E102–E105 (2003) 46. Holdsworth, F.: Fractures, dislocations, and fracture-dislocations of the spine. J Bone Joint Surg Am 52, 1534–1551 (1970) 47. Horn, E.M., Theodore, N., Feiz-Erfan, I., et al.: Complications of halo fixation in the elderly. J Neurosurg Spine 5, 46–49 (2006) 48. Hrabalek, L., Burval, S., Vaverka, M.: Anterior osteosynthesis of odontoid fractures. Acta Chir Orthop Traumatol Cech 75, 332–338 (2008) 49. Jeanneret, B., Magerl, F.: Primary posterior fusion C1/2 in odontoid fractures: indications, technique, and results of transarticular screw fixation. J Spinal Disord 5, 464–475 (1992) 50. Jenkins, J.D., Coric, D., Branch Jr., C.L.: A clinical comparison of one- and two-screw odontoid fixation. J Neurosurg 89, 366–370 (1998) 51. Johnson, J.E., Yang, P.J., Seeger, J.F., et al.: Vertical fracture of the odontoid: CT diagnosis. J Comput Assist Tomogr 10, 311–312 (1986) 52. Julien, T.D., Frankel, B., Traynelis, V.C., et  al.: Evidencebased analysis of odontoid fracture management. Neurosurg Focus 8, e1 (2000) 53. Knöringer, P.: Internal fixation of dens fractures by doublethreaded screws. Orthoped Traumatol 4, 231–245 (1992) 54. Koivikko, M.P., Kiuru, M.J., Koskinen, S.K., et al.: Factors associated with nonunion in conservatively-treated type-II fractures of the odontoid process. J Bone Joint Surg Br 86, 1146–1151 (2004) 55. Koller, H., Zenner, J., Hitzl, W., et  al.: In vivo analysis of atlantoaxial motion in individuals immobilized with the halo thoracic vest or Philadelphia collar. Spine (Phila Pa 1976) 34, 670–679 (2009) 56. Lambotte, A.: L’Intervention Operatoire dans Les Fractures Recentes et Anciennes. In: Relter, R.F. (ed.) Fractures. Henri Lamertin, Brussels (1907) 57. Lennarson, P.J., Mostafavi, H., Traynelis, V.C., et  al.: Management of type II dens fractures: a case-control study. Spine (Phila Pa 1976) 25, 1234–1237 (2000) 58. Maak, T.G., Grauer, J.N.: The contemporary treatment of odontoid injuries. Spine (Phila Pa 1976) 31, S53–S60 (2006). discussion S61 59. Magerl, F., Seemann, P.S.: Stable posterior fusion of the atlas and axis by transarticular screw fixation. In: Kehr, P.,

177 Weidner, A. (eds.) Cervical spine, pp. 322–327. Springer, Wien (1987) 60. Marcotte, P., Dickman, C.A., Sonntag, V.K., et al.: Posterior atlantoaxial facet screw fixation. J Neurosurg 79, 234–237 (1993) 61. Mixter, S.J., Osgood, R.B.: IV. Traumatic lesions of the atlas and axis. Ann Surg 51, 193–207 (1910) 62. Montesano, P.X., Anderson, P.A., Schlehr, F., et al.: Odontoid fractures treated by anterior odontoid screw fixation. Spine (Phila Pa 1976) 16, S33 (1991) 63. Muller, E.J., Schwinnen, I., Fischer, K., et  al.: Non-rigid immobilisation of odontoid fractures. Eur Spine J 12, 522– 525 (2003) 64. Muller, E.J., Wick, M., Russe, O., et  al.: Management of odontoid fractures in the elderly. Eur Spine J 8, 360–365 (1999) 65. Nakanishi, K., Sasaki, T., Tokita, N., et al.: Internal fixation for the odontoid fracture. Orthop Trans 6, 176 (1982) 66. Nourbakhsh, A., Shi, R., Vannemreddy, P., et al.: Operative versus nonoperative management of acute odontoid Type II fractures: a meta-analysis. J Neurosurg Spine 11, 651–658 (2009) 67. Omeis, I., Duggal, N., Rubano, J., et al.: Surgical treatment of C2 fractures in the elderly: a multicenter retrospective analysis. J Spinal Disord Tech 22, 91–95 (2009) 68. Platzer, P., Thalhammer, G., Krumboeck, A., et al.: Plate fixation of odontoid fractures without C1-C2 arthrodesis: practice of a novel surgical technique for stabilization of odontoid fractures, including the opportunity to extend the fixation to C3. Neurosurgery 64, 726–733 (2009). discussion 733 69. Platzer, P., Thalhammer, G., Sarahrudi, K., et al.: Nonoperative management of odontoid fractures using a halothoracic vest. Neurosurgery 61, 522–529 (2007). discussion 529–530 70. Polin, R.S., Szabo, T., Bogaev, C.A., et  al.: Nonoperative management of Types II and III odontoid fractures: the Philadelphia collar versus the halo vest. Neurosurgery 38, 450–456 (1996). discussion 456–457 71. Pratt, H., Davies, E., King, L.: Traumatic injuries of the c1/ c2 complex: computed tomographic imaging appearances. Curr Probl Diagn Radiol 37, 26–38 (2008) 72. Roy-Camille, R., de la Caffiniére, J.H., Saillant, G.: Les traumatismes du rachis cervical superieur C1-C2. Masson et Cie, Paris (1973) 73. Ryan, M.D., Henderson, J.J.: The epidemiology of fractures and fracture-dislocations of the cervical spine. Injury 23, 38–40 (1992) 74. Ryan, M.D., Taylor, T.K.: Odontoid fractures. A rational approach to treatment. J Bone Joint Surg Br 64, 416–421 (1982) 75. Ryan, M.D., Taylor, T.K.: Odontoid fractures in the elderly. J Spinal Disord 6, 397–401 (1993) 76. Saeed, M.U., Dacuycuy, M.A., Kennedy, D.J.: Halo pin insertion-associated brain abscess: case report and review of literature. Spine (Phila Pa 1976) 32, E271–E274 (2007) 77. Sandler, A.J., Dvorak, J., Humke, T., et al.: The effectiveness of various cervical orthoses. An in vivo comparison of the mechanical stability provided by several widely used models. Spine (Phila Pa 1976) 21, 1624–1629 (1996) 78. Sasso, R., Doherty, B.J., Crawford, M.J., et al.: Biomechanics of odontoid fracture fixation. Comparison of the one- and twoscrew technique. Spine (Phila Pa 1976) 18, 1950–1953 (1993)

178 79. Saur, K., Sames, M.: Results of the treatment of odontoid fractures by osteosynthesis with a single axial screw. Acta Chir Orthop Traumatol Cech 75, 48–51 (2008) 80. Schatzker, J., Rorabeck, C.H., Waddell, J.P.: Fractures of the dens (odontoid process). An analysis of thirty-seven cases. J Bone Joint Surg Br 53, 392–405 (1971) 81. Schatzker, J., Rorabeck, C.H., Waddell, J.P.: Non-union of the odontoid process. An experimental investigation. Clin Orthop Relat Res 108, 127–137 (1975) 82. Schneider, A.M., Hipp, J.A., Nguyen, L., et al.: Reduction in head and intervertebral motion provided by 7 contemporary cervical orthoses in 45 individuals. Spine (Phila Pa 1976) 32, E1–E6 (2007) 83. Scott, E.W., Haid Jr., R.W., Peace, D.: Type I fractures of the odontoid process: implications for atlanto-occipital instability. Case report. Neurosurg 72, 488–492 (1990) 84. Seybold, E.A., Bayley, J.C.: Functional outcome of surgically and conservatively managed dens fractures. Spine (Phila Pa 1976) 23, 1837–1845 (1998). discussion 1845–1836 85. Shalayev, S.G., Mun, I.K., Mallek, G.S., et al.: Retrospective analysis and modifications of retractor systems for anterior odontoid screw fixation. Neurosurg Focus 16, 1–4 (2004) 86. Streli, R.: Kompressionosteosynthese bei Fracturen und Pseudoarthrosen des Dens Epistrophei. Z Orthop 119, 675– 676 (1981) 87. Strohm, P.C., Muller Ch, A., Kostler, W., et al.: Halo-fixator vest – indications and complications. Zentralbl Chir 132, 54–59 (2007)

11  Odontoid Process Fractures 88. Stulik, J., Suchomel, P., Lukas, R., et al.: Primary osteosynthesis of the odontoid process: a multicenter study. Acta Chir Orthop Traumatol Cech 69, 141–148 (2002) 89. Subach, B.R., Morone, M.A., Haid Jr., R.W., et  al.: Management of acute odontoid fractures with single-screw anterior fixation. Neurosurgery 45, 812–819 (1999). discussion 819–820 90. Suchomel, P.: Analysis of the causes of failure in the treatment of simple traumatic epidural hematomas. Rozhl Chir 69, 649–654 (1990) 91. Suchomel, P., Taller, S., Lukas, R., et al.: Surgical treatment of fractures of the odontoid process. Rozhl Chir 79, 301–308 (2000) 92. Tashjian, R.Z., Majercik, S., Biffl, W.L., et  al.: Halo-vest immobilization increases early morbidity and mortality in elderly odontoid fractures. J Trauma 60, 199–203 (2006) 93. Tippets, R.H., Alvis, M.A.: Treatment of axis fractures. In: Menezes, A.H., Sonntag, V.K.H. (eds.) Principles of spinal surgery, pp. 871–883. McGraw-Hill, New York (1996) 94. Traynelis, V.C.: Evidence-based management of type II odontoid fractures. Clin Neurosurg 44, 41–49 (1997) 95. Waddell, J.P., Reardon, G.P.: Atlantoaxial arthrodesis to treat odontoid fractures. Can J Surg 26(255–257), 260 (1983) 96. Weller, S.J., Malek, A.M., Rossitch Jr., E.: Cervical spine fractures in the elderly. Surg Neurol 47, 274–280 (1997). discussion 280–271

Fractures of the Ring of Axis (Hangman Type Fractures)

12

P. Suchomel and J. Hradil

Hangman’s fracture [83] is an eponym referring to a bilateral fracture of the C2 partes interarticularis. Schneider’s expressive term was brought to describe a seemingly uniform C2 fracture pattern. Unfortunately, its similarity to the result of proper judicial hanging is misleading. The vast majority of authors observed that contemporary trauma results in a much more graphical complexity of these injuries and is usually a result of much different biomechanical forces. Nonetheless, the term has a long history and has been used for many years. As such, it cannot be formally replaced. Rather than that, there is space for expanding its meaning. The way it is used in contemporary literature abandons original presumptions. It usually encapsulates all radiographic alternatives of the “classical” fracture pattern as well as a full spectrum of associated discoligamentous injuries. Several terminological alternatives that can be traced in the literature: Fracture of the ring of axis (Effendi) Traumatic spondylolisthesis of axis (Garber) Fracture of the middle column of C2 (Roy-Camile) Pedicular fracture of the axis (Borne) Fracture of the neural arch of the axis (Brashear) Fracture of the axis arch (Marar)

12.1 History Hanging as a capital punishment has been practised from biblical times. One of the earliest studies by WoodJones investigates a series of 101 excavated bodies of

P. Suchomel and J. Hradil Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova st. 10, 46063 Liberec, Czech Republic

Nubians executed by Romans in late Roman Byzantine times. The victims, found with rope remnants still around their neck, usually suffered a fracture across the skull base, most probably as a result of “long drop” with subaural knot placement [101]. In England, hanging was introduced by Angles, Saxons, and Jutes around 449 AD. During a reign of Henry VIII, over 72,000 of his subjects were executed. Hanging involved posterior placement of the knot and very short (zero) drop. Victims often struggled violently before succumbing [53]. The same results could, unfortunately, be seen even in more recent executions with subaural knot placement and such cases started intense debates concerning a proper technique of hanging. There is a record of a survivor who was suspended for 15 min [19]. Introduction of a “long drop” dates back to 1784. It was a very effective technique, but several victims were decapitated during public executions. Rev. Prof. Haungton [42] was the first to publish standards on length of the drop and other parameters, quoting on fracture dislocation of second cervical vertebra. Subaural knot placement was a matter of tradition for executioners, and even official recommendations advocating submental placement [62], had no effect on the practice. Colonel Marshall was surprised to find out, that subaural knot technique was, in his own words, “in full swing” as late as in 1913. In fact, things did not change substantially until abolition of capital punishment in United Kingdom in 1965. Due to these circumstances and contrary to common belief, hangman’s fracture caused death in as low as 19% of the hanged convicts [44]. The primary anatomical paper comes from Dr. Frederick Wood-Jones (1913), who investigated bodies of criminals hanged at Rangoon central jail [101]. According to his observations, submental knot placement and a proper technique of “long drop” lead to a quick and uneventful death. Violent hyperextension and

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_12, © Springer-Verlag Berlin Heidelberg 2011

179

180

traction produced a fracture through both partes interacticulares of axis and C2/3 disk disruption. Body of C2 along with dens, atlas, and the head moved cranially, leaving posterior part of the axis attached to the rest of the spine. His paper was the first to show drawings of the fracture and to describe the lethal mechanism in detail. Probably the first real trauma case was published by Clarke [14]. His description of C2 fracture survivor who fell out of the tree suggests it could be a case of hangman’s fracture. Grogono, in 1954, published first radiographs of fractured posterior arch of C2 in a tetraplegic trauma victim [39]. He noticed similarity with the pattern described by Wood-Jones. Later on, several authors also commented on the similarity of their observations [35, 70, 74, 77]. It was Schneider (1965) who provided an expressive eponym “Hangman’s fracture” for eight victims of MVA [83]. French authors adopted the term but they use more precise “la fracture du pendu”, as, it is in fact, a fracture of a “hangee,” not a hangman [78–80]. In 1962, Robert Judet placed first “transpedicular” axis screw (Christian Mazel – personal communication) published by LeConte in 1964 [54] and Découlx (1968) reported on anterior fixation technique for C2/3 space [22]. Cornish (1968) was the first and for a long time the only, advocate of surgical treatment in Englishspeaking literature [18]. The majority of publications reported more or less solitary cases, mostly as part of a larger series of cervical spine trauma. Vichard (1981) found 229 cases of fractures of the pars interarticularis published before 1981 [95]. Treatment was non-surgical in 165 cases and surgical in 64 (25 posterior fusion, 14 anterior fusion, and 25 direct screw fixation of the pars). No concise classification has been proposed nor has there been recognition of any varieties of the injury even in several larger and specific series [7, 18, 23, 60] until 1981, when Francis and Effendi published large sets (123 and 131 cases, respectively) [27, 33]. Francis adopted criteria based on biomechanical cadaveric studies by White and Panjabi [98, 99]. Authors developed a system of five categories. However, it was Effendi’s work that became widely accepted and extensively used in further reports. Effendi et al. divided fractures into three basic categories according to radiological appearance and estimated corresponding (in)stability. The series clearly showed complex fracture patterns including cases of marked asymmetry and infractions into facet joints and C2 body as detected on tomograms. Several traumatic mechanisms were taken into account

12  Fractures of the Ring of Axis (Hangman Type Fractures)

and thoroughly discussed. As in many other papers, the term “hangman’s fracture” was criticized as misleading. Authors suggest “fracture of the ring of axis,” which is probably the most suitable alternative. French authors mostly used a classification of C2 middle column fractures by Roy-Camille [78–80], which includes combination of arch-dens trauma. In 1985, Levine and Edwards modified Effendi’s system by identifying a biomechanically distinct type IIa category [56]. They also specified maximum displacement of type I injury (3 mm). Although based on plain films/tomograms and lacking any direct confirmation of a proclaimed presence/absence of discoligamentous injury, this classification is the most frequently used in recent publications. We also use the Levin and Edward system at our department (Figs. 12.1–12.4). Introduction of CT and MRI confirmed high variability of fracture patterns and soft tissue injury [82] and led to more individual evaluations of stability as well as specific treatment rationales. However, no classification system with respect to CT and/or MRI findings has been put to practice so far. There are over 200 articles, book chapters or other cited sources on the topic [52]. Unfortunately, the majority of these works come from pre-CT and pre-MRI era. Even the most recent sources rely on classifications, results, and recommendations based solely on these reports. The principles derived from these sources were hardly ever analyzed or updated. Nonetheless, they are widely accepted and very seldom questioned.

Fig. 12.1  Levine type I fracture

181

12.2  Classifications

Fig. 12.2  Levine Type II fracture in three different patients. (a) Flexion and dislocation. (b) Extension and dislocation (note fracture of posterior ring of atlas). (c) Angular instability without dislocation

Fig. 12.4  Levine type IIa fracture, kyphotic position susceptible of PLL injury (confirmed by MRI and peroperative discography) Fig. 12.3  CT 3D reconstruction of Levine type II fracture

Spine surgery is heading in the direction of evidence and maximally objective evaluation. Verbal shortcuts of evaluation, such as “good fusion,” “acceptable result,” “no significant pain,” are typical for the majority of articles in the twentieth century and these results need to be reconsidered using contemporary optics. For proper evaluation and evidence of any kind, multicenter-controlled cohort evaluations including CT and MRI findings are absolutely mandatory. Even though there are reports of very good design, the pitfalls of old classifications, traditionalistic setting of treatment options, and short follow-ups do not (and cannot) provide valid evidence.

12.2 Classifications 12.2.1 Effendi Type I: Isolated hairline fractures of the ring of the axis with minimal displacement of the body of C2. The fracture may involve any part of the ring of the axis and may extend anteriorly into the body of C2. The fracture line is then oblique, involving usually one or rarely both postero-inferior corners of the body. The disc space below the axis is normal and stable.

182

Type II: Displacement of the anterior fragment, with an abnormal disc below the axis. The body of the axis may be displaced in extension, flexion or obvious ­forward olisthesis. Type III: Displacement of the anterior fragment with the body of the axis in the flexed position; but in addition, the facet joints at C2-3 are dislocated and locked. A type III lesion must be suspected when the body of the axis is in a position of flexion; it has not been seen when it is in a position of extension or of forward olisthesis.

12.2.2 Francis Grade I displacement of <3.5 mm and angulation <11° Grade II displacement <3.5 mm and angulation >11° Grade III displacement >3.5  mm and <0.5 vertebral width and angulation <11° Grade IV displacement of >3.5 and <0.5 vertebral width and angulation >11° Grade V disc disruption

12.2.3 Levine and Edwards Type I: all nondisplaced fractures and all fractures that showed no angulation and less than 3 mm of displacement (pure hyperextension-axial loading). Type II: Significant displacement (3.5 mm) and angulation (11°) (combination of hyperextension-axial loading with secondary flexion-compression). Type IIa: minimum degree of displacement combined with severe angulation (flexion-distraction). Type III unilateral or bilateral facet dislocation in addition to the posterior element fractures (flexioncompression)

12  Fractures of the Ring of Axis (Hangman Type Fractures)

endplate and upper endplate of C3. The criteria apply even on dynamic radiograph. Type 2: unstable fractures with either more than 2 mm AT or more than 5° of angulation or both. Type 3: middle column fracture with dislocation of C2 facet joints over C3. Type 4: middle column fracture with dens fracture.

12.3 Etiology and Epidemiology Axis fractures consist of approximately 20% of all acute cervical spine fractures. Around 14–16% of axis fractures are combined with atlantal injury. Hangman’s fractures represent 20% of C2 trauma and male/female ratio is approximately 1.6. Most of the cases come from high velocity accidents, i.e., motor vehicle (50– 80%), falls from height (14–25%), diving (1–4%), and other more specific causes [27, 33, 38, 58, 88]. Serious head or chest injury can be seen in up to 43% of cases, polytrauma in general is present in 10–56% admissions [52, 67]. The exact numbers are influenced by population, environment and as such, there can be substantial differences between reports. In general, the spectrum of fractures changes with lifestyle of the population and available diagnostic possibilities. It surely changes in time. Data are closely connected with automotive industry and safety designs involved [55, 102]. The incidence does not seem do decrease with airbag use, it may even be slightly increased due to this otherwise valuable technology [59]. In our series of 40 cases treated surgically [88], there was a male/female ratio of 1.86, mean age 44  years (18–79), and combined fracture with C1 in 17.5%. We found Levine-Edwards type I in 5 cases, type II in 25 cases, and no case of type III fracture.

12.4 Symptoms and Signs 12.2.4 Roy-Camille Classification of C2 Middle Column Fractures Type 1: stable fracture with less than 2 mm of anterior translation (AT) and less than 5° of regional angulation rheumatoid arthritis, which is an angle defined by C2

In isolated Hangman’s fractures, the symptoms are often limited to neck pain, stiffness or transient “electricity-like” whole body irritations. Permanent neurological deficits are very rare [38, 58, 88]. It is appropriate to cite the work of Marar [60] here as it belongs to one of the most misinterpreted of all. Marar claims at least some cord involvement in 11 out of 15 of his patients,

183

12.5  Radiology

Fig. 12.5  Different atypical fracture patterns where the “natural enlargement” of canal by fracture is not true. (a) The arch is not really disconnected – fracture of posterior caudal C2 wall. (b)

Bilateral symmetric anterior spurs, any dislocation can cause compression. (c) Unilateral anterior spur

which is considered an extremely high number by many authors. But he considers purely sensory disturbances, too. This applies not only to six of his patients that recovered fully in 24 h, but also to all other recovering fully until 1 month. In his series, there was no permanent neurological injury! Discrepancy could be at least partially assigned to the level of detail during neurological examination and honest reporting of even slightest signs of cord involvement. DeLorme [23] reports neurological deficit in roughly 1/3 of his 40 cases, and 1/3 of these remained permanent; Muller [67] reports 10.3% neurological deficits, all of them present in patients with type II injury and displacement more than 4 mm. The patients often lack any objective evidence of cord involvement and they are happy to have “just neck pain.” This can create certain bias. In our experience, 50% of the patients recall feelings of “electricity going through their body” with more or less rapid recovery. However, this information needs a direct question. Patients seldom report this actively as these symptoms no longer exist on admission. Unfortunately no study evaluating electrophysiological functions of spinal cord after such an event is available. There are series of radiological examination of fatal craniospinal trauma victims, showing surprisingly high incidence of C2 fractures [1, 21]. Alker [1] found that 39% victims who died solely as a result of neck injury had C2 fracture. Study of Bucholz [9] on over 170 cadavers of multiple trauma victims concludes that severe neurological injury is a frequent complication of Hangman’s fracture, and usually is incompatible with survival. Schneider [83] quoted on “roomingness” of the upper cervical canal and “death-averting decompression of

the upper cervical cord accomplished by lesion itself.” This is true of Hangman’s fracture in its classical form, however, frequent “atypical” patterns (Fig. 12.5) show clear potential to compress on medulla at the moment of injury or in case of significant dislocation at any time thereafter [87]. It is likely that the zone between instant death and symptom-free survival is narrow here and a marked neurological deficit in a survivor is a very rare result under these circumstances.

12.5 Radiology Definition of hangman’s fracture is somewhat misty as it evolves in time and so is the radiological description. There are long discussions about which fracture pattern still represents a hangman’s fracture (“typical” or “atypical”) [45]. The axis ring ruptures vary between two extremes – the fracture through the C2 body (Fig. 12.6) and distant fracture of the arch (Fig. 12.7). The most important radiographic features are usually derived from existing classification systems (see above). Thus, radiology is focused only on films. Although currently inevitable, more advanced modalities (dynamic films/CT/MRI) have not been used for classification purposes up to the present. The initial pattern is a bilateral fracture through pars interarticularis (between superior and inferior facet joints). Many authors report fractures of the “pedicle” [5, 6, 11]. However, pars interarticularis is not identical with the pedicle of the axis as we have

184

emphasized previously. True fractures of the pedicle of axis are very rare. There are only unilateral cases reported with various fracture lines through the other side of the axis ring and only extreme conditions can provide such a result [44].

12  Fractures of the Ring of Axis (Hangman Type Fractures)

Fig. 12.6  Axial CT scan showing the fracture of the C2 ring extending into the verberal body

The combination of original fracture site and symmetry is actually rare (Fig.  12.8). The fracture line most often involves posterior cortex of the axis body, facet joints and it can extend into vertebral foramina (Fig. 12.9). There are numerous reports, which do not quote on such alternative patterns but they clearly show them in presented radiolographic documentation [8, 37, 63, 102]. It can be only hypothesized, how many cases with “atypical” patterns were actually present in the series published in pre-CT era. The most common “alternative” pattern is a fracture line invasion into posterior cortical wall of C2 body [2, 9, 11, 27, 30, 61, 64, 85, 88, 97]. The extent of C2 body infraction depends on trauma mechanism and also on the local anatomy of the posterior wall. The cortex is thick in lateral and inferior areas and gets narrow medially and cranially. The result is an incomplete fracture line above the inferior portion of the posterior axis wall. CT-scan shows disruption of the “inner cortical ring” close to the midline, and a characteristic picture of bilateral “spurs” (Fig. 12.5). These can be of various sizes and they are often asymmetric [85, 97]. The ultimate form is a complete avulsion of the posterior wall [88, 96]. The majority of cases have their true fracture site more anteriorly compared to the pattern described by Wood-Jones and “revived” by Schneider. This often leads to another very common finding: superior facet

Fig. 12.7  Plain lateral film showing distant fracture of the arch (only one side)

Fig. 12.8  CT of a rare symmetric transisthmic pattern of hangman’s fracture

12.5  Radiology

185

Fig. 12.9  CT showing fracture line reaching the FT. (a) Bilateral FT fracture involvement. (b) Another patient fracture through FT seen on sagittal reconstruction

joint involvement [18, 27, 82, 88]. Such patterns are routinely seen on 3D CT reconstructions (Fig. 12.9b), and they were confirmed also on autopsy findings in the past [85]. The extension of fracture line into vertebral artery (VA) foramina is not an exception [18, 88]. However, there are only several reports on serious VA injury [58, 71]. Fracture through superior facet joint should be considered unstable and reposition is likely to be essential for a good, long-term functional outcome [52] as in any other intra-articular fractures in human body. The majority of fractures are essentially asymmetric. The above-mentioned variants can be present unilaterally with the other side harboring a “classical” fracture of the pars. Up to 50% of cases can display such pattern [82] and they are likely to be produced due to rotation before or during the trauma impact. However, asymmetric fractures can very often be produced without any asymmetric loading or rotation in laboratory conditions [90]. It should be mentioned that no classification system deals with asymmetry or rotation. All are 30+ years old and rely on plain films that simply cannot demonstrate such features. The radiographic “span” of hangman’s fracture is limited anteriorly with deeper involvement into axis body, where it can merge with type III dens fractures, according to Anderson-D’Alonso. Posterior border seems to be inferior facet joint. Fractures located bilaterally more posteriorly (neural arch, behind the joint) have no significant effect on C2/3 stability and form a separate category of C2 trauma. Cautions is necessary when evaluating radiographs of children. Normally, neurocentral synchondroses ossify between 3 and 7  years of age [86]; however,

persistent synchondroses or primary spondylolyses (as seen in pyknodysostosis or Crouzon disease) can occur and these can mimic hangman’s fracture. On the other hand, there are reports of traumatic lesions in infants under 1 or 2 years of age [32], mostly as a result of child abuse. Dynamic radiographs, CT, MRI, clinical findings, and history of trauma are necessary to distinguish developmental conditions from their traumatic counterparts. Hangman’s fracture is, of course, not only a fracture but frequently (types II and III) a complex soft tissue injury, too. This aspect, if present, is actually very important for treatment planning. At the time of impact, C2/3 disc is put to high levels of mechanical stress that often lead to its injury, which allows more or less ­dislocation/angulation of both adjacent vertebrae. “Slipping” of the axis over C3 is a base for Garber’s term “traumatic spondylolisthesis.” However, only rare cases of atraumatic axis olistheses and C2/3 dislocations without any fracture have been reported and quoted [15, 29, 68]. In case of significant C2/3 displacement and especially angulation, at least one of anterior or posterior longitudinal ligaments is likely to be injured. Definitions of (in)stability vary widely even in recent literature [57]. The majority of hangman’s fractures result from extension and axial loading (typical pattern in MVA). Injury to anterior longitudinal ligament is the most common finding (Levine type II). Flexion distraction mechanisms, that lead to type IIa and type III patterns, are accompanied by injured PLL. Absence of any support in flexion renders this condition highly unstable (Fig.  12.10). Traction is hazardous in case of type IIa injury [26, 46, 47] as it widens the fracture gap and increases angulation. The soft

186

12  Fractures of the Ring of Axis (Hangman Type Fractures)

89], that can even lead to “spontaneous” fusion of the segment after several years [7]. Those who use external immobilization techniques report significant percentages of fusion with residual dislocation [38].

12.6 Treatment Strategy

Fig.  12.10  Levine type II with distractive dislocation, severe ligamentous injury (no traction!)

tissue injury can involve additional elements, such as ­capsulae of C2/3 facet joints in type III injury, nuchal ligament, and the conditions become far more complex in combined atlanto-dental injuries. Mechanisms able to produce a true hangman’s fracture by traction-extension are very rare today [26, 81, 102]. Contemporary suicidal attempts are usually short drops resulting mostly in reflective cardiac arrest or asphyxia, rather than cervical spine injury [45]. Slipping under the safety belt during a car accident is possible but happens rarely with modern car designs. In addition to static features of the injury, it is essential to emphasize the necessity of dynamic film evaluation. Major classifications include this modality to rule out type II unstable fractures that are spontaneously reduced. They resemble type I injury, which can also be easily overlooked on plain films [99]. Nonetheless, negative radiographs of active and/or passive flexion-extension are not an absolute proof of stability due to possible stabilizing effects of reflex muscle spasms. Dynamic MRI would be of a high value here, assessing soft tissue behavior, but this technology is still far from abundant at the present time. The structural incompetence of anterior and posterior longitudinal ligaments can be verified by C2/3 discography also. Unfortunately, studies with long-term follow-up are very rare. Contrary to findings in short periods of time, authors show degenerative changes in C2/3 level [7,

First of all, it is necessary to emphasize that there is no class I or II evidence for the treatment guidelines or recommendations of hangman’s facture available in the literature. According to reviews by Koller [50, 52], more than 50 authors present approx. 40 different concepts for therapy of hangman’s fracture. Many of them conclude that primary surgical treatment of the fracture is not necessary for successful result [10, 17, 29, 30, 38, 41, 60, 65, 72, 76, 84, 85, 93, 104] and advocate various types of external immobilization. Other authors are in favor of early surgical approach at least in unstable ones [6, 12, 34, 40, 48, 49, 52, 61, 66, 75, 89, 91, 94], reporting favorable result, better fracture alignment, substantial reduction of treatment period, and early mobilization with better quality of life. Historically, conservative treatment dominated the field. Most often, dislocated fractures were reduced by traction first and then patients placed to brace. There are plenty of external rigid [4, 29, 30, 41, 60, 84, 93] or in some cases non-rigid [17, 67] immobilization techniques, ranging from sand-bag support [92] through soft/hard collars, traction devices, plaster supports of various designs, up to SOMI braces and halo systems. Although there are extensive reviews and meta-analyses available [10, 17, 38, 96], the definition of successful treatment of hangman’s fracture is usually quite vague in both older and recent publications and it is most often set equal to fusion (usually, without any definition) [16, 96]. Fusion refers to the site of a fracture, and not to the condition of C2/3 disc space and overall sagittal cervical alignment. However, the fracture site is probably of lesser importance than is the type and extent of displacement and associated soft tissue injury [25]. Perfect Complete reduction of dislocated fracture cannot be achieved by halo immobilization in all cases. Nearly all later publications admit recurrence of anterior translation of up to 60% and angulation of up to 40% of the initial status prior to reduction by

12.6  Treatment Strategy

187

traction [56]. Fusion in displacement of up to 5 mm can be seen with halo-vest immobilization as was documented by Coric [17]. Authors often claim that healing in displaced position is not harmful. However, there is a lack of self-evaluation data to support such statement. Malalignment and sagittal profile distortion represents terrain prone to secondary degenerative changes, including disc osteochondrosis, osteophyte formation, and calcification (Fig.  12.11). Delayed spontaneous C2/C3 fusions based on degeneration can be seen as documented in follow-ups lasting at least several years [7, 89]. Considering a frequent involvement of C1/C2 joint, it is highly questionable to claim this result to be a success. Motion restriction and especially pain are usually fairly interpreted or even completely omitted in many studies on both surgical [6] and conservative treatments [10, 17, 41, 72, 73, 96]. It can only by hypothesized why there is no significant pain and limited range of motion, whereas in many other conditions involving C1/2 osteochondrosis, significant pain can be observed [28, 31, 103]. Ubiquitous lack of longer follow-up may surely be a factor. Only recently, there are studies of a high methodological level providing evidence that malunion in C2 fractures has a strong association with development of atlantoaxial osteoarthritis, significant impact

on clinical outcomes, and both total and atlantoaxial neck rotations [51]. Any type of halo immobilization suffers from inherent problems, such as pin loosening, infection of various degrees including epidural abscess, pressure sores, lack of compliance by patient, breathing problems, pneumonia, and transient or permanent loss of range of motion [36]. Also, the halo-vest competence to sufficiently stabilize the upper cervical spine (UCS) is questionable. Current biomechanical works report the sagittal and coronal “snaking” of UCS fixed in halo [43]. Crutchfield [20], satisfied with his invention, claimed that “traction leaves few, if any, indications for surgery” in cervical spine trauma. This is still perfectly true, as nearly all fractures can be treated by simple traction, immobilization with bedrest, and without necessity of surgery. Obviously, it is not a necessity of surgery, but it is a benefit of the patient, what really matters. Yet another aspect is, choice of the patient. In our practice, an average 6-day hospital stay involving 1-h routine anterior surgery clearly wins over 3 months in halo-vest. Advancements in spine surgery in the past 10–20 years inevitably lead to conclusion that treatment recommendation of old studies (including all ground works for classification systems) cannot be fully taken

Fig. 12.11  Hangman’s type II fracture treated conservatively in halo-vest. (a) Axial CT scan showing healed fracture in malposition. (b) Plain lateral radiogram depicting the deformity, note the

fusion of C2-3. (c) Flexion on plain film confirming fixed deformity (Courtesy of Prof. Robert Veres, Budapest)

188

into account. A fracture-dislocation of C3/4 level in an otherwise healthy person would be treated by anterior surgery and fusion today. It becomes very hard to find a reasonable argument against the use of the same principle for C2/3 intervertebral space. Nonetheless, even in this era, there are authors in favor of primary conservative treatment of all variants including highly unstable fractures [17]. There are reports of successful treatment of complete C2/3 dislocations by halo immobilization [76]. On the other hand, some of the authors with extensive experience with conservative approach were very satisfied with technical accessibility, immediate segmental stability, and excellent results of primary surgery and proposed it at least as a fully acceptable alternative [91]. There are many types and extents of surgical treatment in the literature. Most exotic variants are either insufficient, or overtreating, or both [8]. Four viable options of surgical treatment are available today: Anterolateral approach with bone graft (auto/ allograft) fixed by plate and screws (best bicortical) is the first option. Even back in 1970, Norrel reports five cases of unstable hangman’s fracture treated by anterior dowel fusion and average hospital stay of mere 10 days [69]. We reported average hospital stay of 6 to 8  days (polytrauma excluded) including preoperative fracture-reduction by halo-traction [88]. The wound complications and infection is rare in contemporary cervical spine surgery and mostly equals to zero [50, 88]. Although there are some variants, such as transoral approach for C2-3 fusion [100], proper use of anterolateral technique is straightforward, routine, and hardly ever brings problems or complications [50, 52, 88]. Patients are usually advised to wear hard collar for various periods of time after surgery usually depending on the severity concomitant to soft tissue injury. There are also recent studies showing sufficient biomechanical parameters of fixation using simple anterior plate even with non-constrained monocortical screws [13]. Posterior approach can be primarily used in cases of otherwise irreducible locked facets (type III). The purpose of surgery can be limited to facet joint reposition with subsequent anterior fixation, or posterior fixations of various extents can follow. Posterior C2-C3 screw-rod fixation is biomechanically superior to anterior plate stabilization in lateral bending and axial rotation [24]. Considering available clinical series where

12  Fractures of the Ring of Axis (Hangman Type Fractures)

majority of surgeries were performed with anterior fixation [50, 52, 88, 91], it is again the necessity of significant muscle dissection that renders posterior approach less favorable at least in the treatment of fractures other than type III. Combined approach is reasonable in highly unstable injuries where both anterior and posterior fixation techniques can be associated to provide stable 360° construct; for example, type III injury reduced and fixed posteriorly with subsequent anterior disc removal and fusion. Direct pars fixation, according to Judet, is a specific alternative applicable to cases with limited discoligamentous injury. As in other posterior C2 screw techniques, it is often feared as risky because of potential VA injury. Image guidance is of little use here [3], since the fragments are often dislocated and/or can be displaced during the procedure, thus making the C2 fiducial registration impossible. Posterior transpedicular/ transisthmic screw can be introduced by “free hand” safely if individual anatomy, fracture morphology, and preoperative radiologic workup are respected, as described in the previous chapter. However, posterior screw placement performed under direct CT-guidance is not only safer but in addition, can directly show the desired compression of the fracture [89].

12.7 Our Preference All patients with suspected C-spine injury admitted to our hospital, proceeded directly from the place of accident, through the emergency department to spiral CT first, without plain films. Those referred from other hospitals mostly come with plain films and CT images already obtained. If further algorithm is not changed because of concomitant injuries and a hangman type fracture is diagnosed, we obtain an MRI in all cases to assess the extent of soft tissue damage and establish the status of radiographic neural compromise. Although MRI is considered unnecessary by some authors, particularly in the absence of neurologic deficit, we prefer to obtain it. In type I fractures (non-displaced) the eventual C2-3 disc damage can be confirmed, occasionally the disruption of anterior or posterior longitudinal ligament can be seen, and also the coincident injury to other disc levels can be excluded (Fig. 12.12). This algorithm is easily followed in our hospital where MRI is available 24 h a day.

12.7  Our Preference

189

Fig. 12.13  The same patient as Fig. 12.12. Dislocation of fracture during extension. Notice the posterior atlas ring fracture

Fig. 12.12  MRI in T2 sequence of type I Hangman’s fracture. Clearly visible tear of ALL and anterior disc at C2-3 level and canal compromise by degenerative disc disease at C5-6 and C6-7 levels

Further workup is modified according to the type of injury. There is no need for further investigation in dislocated type II and III fractures as they are directly indicated for adequate traction/reduction followed by surgical stabilization. If an isolated hangman fracture without translational or angular dislocation is confirmed and no compression of spinal cord exists, then the major issue is to decide whether the fracture is stable and well configured and if it can be treated conservatively. Bone abruption of anterior C2 edge and/or concomitant fractures (often, posterior C1 arch) can strengthen our inkling for potential instability. Also, MR evidence of ligamentous or disc injury can raise our suspicion but the true stability cannot be accurately confirmed unless there is movement visible on flexion-extension lateral projections (Fig.  12.13). In cooperating conscious patients without neurological deficit, we always perform a manual surgeon-guided flexion and extension under fluoroscopy (possible

also lying on the side) and document the extreme positions (Fig. 12.14). Not only the patient’s status, but also radiographic findings can be a limitation of dynamic imaging. The patient should be carefully evaluated for presence of previously described anterior spurs and/or incomplete ring fractures, which could cause compression during head manipulation. If the fracture is considered stable and there is less than 3 mm of inter-fragmental distance on initial axial CT scan, we recommend the use of a hard cervical collar for 3  months with regular CT follow-up to monitor fracture healing. The treatment is finished when there is no visible fracture line on axial CT images and confirmed stability on dynamic films. Occasionally, this can take more than 3  months (Fig. 12.15). If, despite dynamic stability, the fracture gap is more than 3 mm on axial CT, we discuss with the patient the option of direct CT-guided posterior osteosynthesis (Fig.  12.16). Depending on his/her decision, we perform the fixation or continue with Philadelphia collar or SOMI brace and careful follow-up. In initially displaced fractures or those with instability on dynamic films, we always recommend surgical treatment. In cases of marked dislocation, we apply halo-ring and traction of 2–5  kg (with

190 Fig. 12.14  Non-displaced fracture with abruption of anterior C2 edge. (a) Neutral position. (b) Flexion. (c) Extension revealing angular instability. (d) Final treatment with anterior graft and plate fixation

12  Fractures of the Ring of Axis (Hangman Type Fractures)

191

12.7  Our Preference

Fig. 12.15  Still apparent facture line after 4 month of Philadelphia collar support

a

exception of total disruption – some type Levine IIa, where more gentle manipulation allows preoperative reduction). The majority of displaced fractures can be reduced overnight. Reducible cases are treated by anterior approach with discectomy, auto/allogenic bone graft, plate and bicortical screw fixation. Posterior approach is reserved for complicated cases irreducible by simple traction and more complex C1-2 injury. Although infrequent, there are borderline cases where the decision for surgery is controversial. In such cases, we usually proceed with operative intervention and check the disc and ligament integrity with intraoperative discography. So far, in all these operated cases, we were able to confirm a contrast leak either through anterior or posterior longitudinal ligament (Fig. 12.17). b

Fig. 12.16  CT guided direct osteosynthesis of hangman’s fracture according to Judet. (a) Initial CT scan. (b) Fracture gap disappearance after tightening of the lag screws

192

12  Fractures of the Ring of Axis (Hangman Type Fractures)

a

b

c

d

Fig. 12.17  Borderline instability of hangman’s fracture. (a) Axial CT showing hairline fracture. (b) Dynamic investigation – ­flexion. (c) Dynamic investigation extension. (d) Peroperative discography confirming the morphological incompetence of PLL

193

References Fig. 12.18  (a) extension , (b) flexion

a

b

a

b

Fig. 12.19  (a) extension, (b) flexion

All patients are followed for at least 2  years to establish long-term stability and the influence on adjacent segments; however, the patient response rate, as in other trauma groups, is limited (Figs. 12.18 and 12.19). We always consider treatment prioritization in polytrauma patients. Without neurological compromise, definite surgical treatment of hangman’s fracture can always be postponed until other life-threatening situations are addressed and patient stabilized.

References   1. Alker Jr., G.J., Oh, Y.S., Leslie, E.V.: High cervical spine and craniocervical junction injuries in fatal traffic accidents: a radiological study. Orthop Clin North Am 9, 1003–1010 (1978)   2. Andrews, R.J.: Hangman’s fractures involving the body of C2. Neurosurgery 27, 845–846 (1990)   3. Arand, M., Hartwig, E., Kinzl, L., et al.: Spinal navigation in cervical fractures–a preliminary clinical study on Judetosteosynthesis of the axis. Comput Aided Surg 6, 170–175 (2001)

194   4. Baumgarten, M., Mouradian, W., Boger, D., et al.: Computed axial tomography in C1-C2 trauma. Spine (Phila Pa 1976) 10, 187–192 (1985)   5. Benzel, E.C.: Anatomic consideration of C2 pedicle screw placement. Spine (Phila Pa 1976) 21, 2301–2302 (1996)   6. Borne, G.M., Bedou, G.L., Pinaudeau, M.: Treatment of pedicular fractures of the axis. A clinical study and screw fixation technique. J Neurosurg 60, 88–93 (1984)   7. Brashear Jr., R., Venters, G., Preston, E.T.: Fractures of the neural arch of the axis. A report of twenty-nine cases. J Bone Joint Surg Am 57, 879–887 (1975)   8. Bridwell, K.H.: Treatment of a markedly displaced hangman’s fracture with a luque rectangle and a posterior fusion in a 71-year-old man. Case report. Spine (Phila Pa 1976) 11, 49–52 (1986)   9. Bucholz, R.W.: Unstable hangman’s fractures. Clin Orthop Relat Res 154, 119–124 (1981) 10. Bucholz, R.D., Cheung, K.C.: Halo vest versus spinal fusion for cervical injury: evidence from an outcome study. J Neurosurg 70, 884–892 (1989) 11. Burke, J.T., Harris Jr., J.H.: Acute injuries of the axis vertebra. Skeletal Radiol 18, 335–346 (1989) 12. Chen, X.S., Jia, L.S., Cao, S.F., et al.: Diagnosis and surgical management of Hangman’s fracture combined with intervertebral disc injury. Zhonghua Wai Ke Za Zhi 42, 712–715 (2004) 13. Chittiboina, P., Wylen, E., Ogden, A., et al.: Traumatic spondylolisthesis of the axis: a biomechanical comparison of clinically relevant anterior and posterior fusion techniques. J Neurosurg Spine 11, 379–387 (2009) 14. Clarke, A.P.: Fracture of the cervical vertebrae. JAMA 3, 390–391 (1884) 15. Colangelo, E.J.: Cervicocranium and the aviator’s protective helmet. Aviat Space Environ Med 46, 1263–1264 (1975) 16. Cooper, P.R., Maravilla, K.R., Sklar, F.H., et al.: Halo immobilization of cervical spine fractures. Indications and results. J Neurosurg 50, 603–610 (1979) 17. Coric, D., Wilson, J.A., Kelly Jr., D.L.: Treatment of traumatic spondylolisthesis of the axis with nonrigid immobilization: a review of 64 cases. J Neurosurg 85, 550–554 (1996) 18. Cornish, B.L.: Traumatic spondylolisthesis of the axis. J Bone Joint Surg Br 50, 31–43 (1968) 19. Crook, G.T.: The complete newgate calendar, vol. 2, p. 181. Navarre Society, London (1926) 20. Crutchfield, W.G.: Skeletal traction in treatment of injuries to the cervical spine. JAMA 155, 29–32 (1954) 21. Davis, D., Bohlman, H., Walker, A.E., et al.: The pathological findings in fatal craniospinal injuries. J Neurosurg 34, 603–613 (1971) 22. Decoulx, P., Decoulx, J., Duquennoy, A., et  al.: Fractures and luxations of the cervical spine. Indications and technic of anterior arthrodesis (especially C.2-C.3)]. J Chir (Paris) 96, 423–437 (1968) 23. DeLorme, T.L.: Axis-pedicle fractures. J Bone Joint Surg Br 49, 1472 (1967) 24. Duggal, N., Chamberlain, R.H., Perez-Garza, L.E., et  al.: Hangman’s fracture: a biomechanical comparison of stabilization techniques. Spine (Phila Pa 1976) 32, 182–187 (2007)

12  Fractures of the Ring of Axis (Hangman Type Fractures) 25. Dussault, R.G., Effendi, B., Roy, D., et  al.: Locked facets with fracture of the neural arch of the axis. Spine (Phila Pa 1976) 8, 365–367 (1983) 26. Edgar, M.A., Fisher, T.R., McSweeney, T., et al.: Tetraplegia from hangman’s fracture: report of a case with recovery. Injury 3, 199–202 (1972) 27. Effendi, B., Roy, D., Cornish, B., et al.: Fractures of the ring of the axis. A classification based on the analysis of 131 cases. J Bone Joint Surg Br 63-B, 319–327 (1981) 28. Ehni, G., Benner, B.: Occipital neuralgia and the C1-2 arthrosis syndrome. J Neurosurg 61, 961–965 (1984) 29. Elliott Jr., J.M., Rogers, L.F., Wissinger, J.P., et  al.: The hangman’s fracture. Fractures of the neural arch of the axis. Radiology 104, 303–307 (1972) 30. Fielding, J.W., Francis, W.R., Hawkins, R.J., et al.: Traumatic spondylolisthesis of the axis. Clin Orthop Relat Res 239, 48–52 (1982) 31. Finn, M., Fassett, D.R., Apfelbaum, R.I.: Surgical treatment of nonrheumatoid atlantoaxial degenerative arthritis producing pain and myelopathy. Spine (Phila Pa 1976) 32, 3067– 3073 (2007) 32. Finnegan, M.A., McDonald, H.: Hangman’s fracture in an infant. Can Med Assoc J 127, 1001–1002 (1982) 33. Francis, W.R., Fielding, J.W., Hawkins, R.J., et al.: Traumatic spondylolisthesis of the axis. J Bone Joint Surg Br 63-B, 313–318 (1981) 34. Fuentes, S., Metellus, P., Dufour, H., et al.: Traumatic spondylolisthesis of the axis: arguments in favor of surgical management after analysis of 8 patients. Neurochirurgie 49, 25–30 (2003) 35. Garber, J.N.: Abnormalities of the Atlas and Axis Vertebrae– Congenital and Traumatic. J Bone Joint Surg Am 46, 1782– 1791 (1964) 36. Garfin, S.R., Botte, M.J., Waters, R.L., et al.: Complications in the use of the halo fixation device. J Bone Joint Surg Am 68, 320–325 (1986) 37. Gerlock Jr., A.J., Mirfakhraee, M.: Computed tomography and hangman’s fractures. South Med J 76, 727–728 (1983) 38. Greene, K.A., Dickman, C.A., Marciano, F.F., et al.: Acute axis fractures. Analysis of management and outcome in 340 consecutive cases. Spine (Phila Pa 1976) 22, 1843–1852 (1997) 39. Grogono, B.J.S.: Injuries of the atlas and axis. J Bone Joint Surg Br 36-B, 397–410 (1954) 40. Guiot, B., Fessler, R.G.: Complex atlantoaxial fractures. J Neurosurg Spine 91, 139–143 (1999) 41. Hadley, M.N., Browner, C., Sonntag, V.K.: Axis fractures: a comprehensive review of management and treatment in 107 cases. Neurosurgery 17, 281–290 (1985) 42. Haughton, S.: On hanging, considered from a mechanical and physiological point of view. Philos Mag J Sci 32, 23–34 (1866) 43. Ivancic, P.C., Beauchman, N.N., Tweardy, L.: Effect of halovest components on stabilizing the injured cervical spine. Spine (Phila Pa 1976) 34, 167–175 (2009) 44. James, R., Nasmyth-Jones, R.: The occurrence of cervical fractures in victims of judicial hanging. Forensic Sci Int 54, 81–91 (1992) 45. Jarolimek, A.M., Coffey, C.C., Sandler, C.M., et al.: Imaging of upper cervical spine injuries – part III: C2 bellow the dens. Appl Radiol 33, 9–21 (2004)

References 46. Jeanneret, B., Magerl, F., Ward, J.C.: Overdistraction: a hazard of skull traction in the management of acute injuries of the cervical spine. Arch Orthop Trauma Surg 110, 242–245 (1991) 47. Jeffreys, E.: Disorders of the cervical spine, p. 60. Butterworths, London (1980) 48. Junge, A., El-Sheik, M., Celik, I., et al.: Pathomorphology, diagnosis and treatment of “hangman’s fractures”. Unfallchirurg 105, 775–782 (2002) 49. Kocis, J., Wendsche, P., Visna, P., et  al.: Traumatic spondylolisthesis of the axis. Acta Chir Orthop Traumatol Cech 70, 214–218 (2003) 50. Koller, H.: The unstable traumatic spondylolisthesis C2/3. Akt Traumatol 35, 183–202 (2005) 51. Koller, H., Acosta, F., Forstner, R., et al.: C2-fractures: part II. A morphometrical analysis of computerized atlantoaxial motion, anatomical alignment and related clinical outcomes. Eur Spine J 18, 1135–1153 (2009) 52. Koller, H., Kathrein, A.: Letter to the editor concerning: a systematic review of the management of hangman’s fractures by Xin-Feng Li et al. Eur Spine J 15, 257–269 (2006). Eur Spine J 15:1415–1418; author reply 1419–1421 53. Laurence, J.: A history of capital punishment, p. 42. Sampson, London (1926) 54. Leconte, P.: Fracture et luxation des deux premieres vertebres cervicales. In: Judet, R. (ed.) Luxation Congenitale de la Hanche. Fractures du Cou-de-pied Rachis Cervical. Actualites de Chirurgie Orthopedique de l’Hospital Raymond-Poincare, vol. 3, pp. 147–166. Masson et Cie, Paris (1964) 55. Lesoin, F., Thomas, C.E., Lozes, G., et  al.: Has the safety-belt replaced the hangman’s noose? Lancet 1, 1341 (1985) 56. Levine, A.M., Edwards, C.C.: The management of traumatic spondylolisthesis of the axis. J Bone Joint Surg Am 67, 217– 226 (1985) 57. Li, X.F., Dai, L.Y., Lu, H., et al.: A systematic review of the management of hangman’s fractures. Eur Spine J 15, 257– 269 (2006) 58. Lohnert, J., Latal, J.: Fracture of the axis–surgical treatment. II. Axial isthmus. Acta Chir Orthop Traumatol Cech 60, 47–50 (1993) 59. Maiman, D.J., Larson, S.J.: Management of odontoid fractures. Neurosurgery 11, 471–476 (1982) 60. Marar, B.C.: Fracture of the axis arch. “Hangman’s fracture” of the cervical spine. Clin Orthop Relat Res 106, 155–165 (1975) 61. Marotta, T.R., White, L., TerBrugge, K.G., et al.: An unusual type of hangman’s fracture. Neurosurgery 26, 848–850 (1990). discussion 850–841 62. Marshal, J.J.: Judicial executions. Brit Med J 2, 779–782 (1888) 63. McCall, I., el Masri, W., Jaffray, D.: Hangman’s fracture in ankylosing spondylitis. Injury 16, 483–484 (1985) 64. Mirvis, S.E., Young, J.W., Lim, C., et al.: Hangman’s fracture: radiologic assessment in 27 cases. Radiology 163, 713–717 (1987) 65. Mollan, R.A., Watt, P.C.: Hangman’s fracture. Injury 14, 265–267 (1982) 66. Moon, M.S., Moon, J.L., Moon, Y.W., et  al.: Traumatic spondylolisthesis of the axis: 42 cases. Bull Hosp Jt Dis 60, 61–66 (2001)

195 67. Muller, E.J., Wick, M., Muhr, G.: Traumatic spondylolisthesis of the axis: treatment rationale based on the stability of the different fracture types. Eur Spine J 9, 123–128 (2000) 68. Nordstrom, R.E., Lahdenranta, T.V., Kaitila, I.I., et  al.: Familial spondylolisthesis of the axis vertebra. J Bone Joint Surg Br 68, 704–706 (1986) 69. Norrell, H., Wilson, C.B.: Early anterior fusion for injuries of the cervical portion of the spine. JAMA 214, 525–530 (1970) 70. Norton, W.L.: Fractures and Dislocations of the Cervical Spine. J Bone Joint Surg Am 44, 115–139 (1962) 71. Okuchi, K., Fujioka, M., Konobu, T., et  al.: A case of Hangman’s fracture associated with vertebral arteriovenous fistula treated with trapping. No Shinkei Geka 22, 55–59 (1994) 72. Pepin, J.W., Hawkins, R.J.: Traumatic spondylolisthesis of the axis: Hangman’s fracture. Clin Orthop Relat Res 157, 133–138 (1981) 73. Pinczewski, L., Taylor, T.K., Ryan, M.D.: Hangman’s fracture: nonoperative management with the halocast. Aust N Z J Surg 53, 71–76 (1983) 74. Ramadier, J.O., Bombart, M.: Fractures and dislocations of the cervical spine without spinal cord lesion. I. Generalities. Lesions of the 21st vertebrae. 52 cases. Rev Chir Orthop Reparatrice Appar Mot 49, 741–764 (1963) 75. Reynier, Y., Lena, G., Diaz-Vazquez, P., et al.: Evaluation of 138 fractures of the cervical spine during a recent 5-year period (1979 to 1983). Therapeutic approaches. Neurochirurgie 31, 153–160 (1985) 76. Roda, J.M., Castro, A., Blazquez, M.G.: Hangman’s fracture with complete dislocation of C-2 on C-3. Case report. J Neurosurg 60, 633–635 (1984) 77. Rogers, W.A.: Fractures and dislocations of the cervical spine; an end-result study. J Bone Joint Surg Am 39-A, 341– 376 (1957) 78. Roy-Camille, R., de la Caffiniére, J.H., Saillant, G.: Les traumatismes du rachis cervical superieur C1-C2. Masson et Cie, Paris (1973) 79. Roy-Camille, R., Saillant, G.: Surgery of the cervical spine. 4. Osteosynthesis of the upper cervical spine. Nouv Presse Med 1, 2847–2849 (1972) 80. Roy-Camille, R., Saillant, G.: Spinal injuries without neurologic complications. Int Orthop 8, 155–162 (1984) 81. Saldeen, T.: Fatal neck injuries caused by use of diagonal safety belts. J Trauma 7, 856–862 (1967) 82. Samaha, C., Lazennec, J.Y., Laporte, C., et al.: Hangman’s fracture: the relationship between asymmetry and instability. J Bone Joint Surg Br 82, 1046–1052 (2000) 83. Schneider, R.C., Livingston, K.E., Cave, A.J., et  al.: “Hangman’s fracture” of the cervical spine. J Neurosurg 22, 141–154 (1965) 84. Seljeskog, E.L., Chou, S.N.: Spectrum of the hangman’s fracture. J Neurosurg 45, 3–8 (1976) 85. Sherk, H.H., Howard, T.: Clinical and pathologic correlations in traumatic spondylolisthesis of the axis. Clin Orthop Relat Res 174, 122–126 (1983) 86. Smith, J.T., Skinner, S.R., Shonnard, N.H.: Persistent synchondrosis of the second cervical vertebra simulating a hangman’s fracture in a child. Report of a case. J Bone Joint Surg Am 75, 1228–1230 (1993)

196 87. Starr, J.K., Eismont, F.J.: Atypical hangman’s fractures. Spine (Phila Pa 1976) 18, 1954–1957 (1993) 88. Suchomel, P., Hradil, J., Barsa, P., et al.: Surgical treatment of fracture of the ring of axis – “hangman’s fracture”. Acta Chir Orthop Traumatol Cech 73, 321–328 (2006) 89. Taller, S., Suchomel, P., Lukas, R., et al.: CT-guided internal fixation of a hangman’s fracture. Eur Spine J 9, 393–397 (2000) 90. Teo, E.C., Paul, J.P., Evans, J.H., et al.: Experimental investigation of failure load and fracture patterns of C2 (axis). J Biomech 34, 1005–1010 (2001) 91. Tuite, G.F., Papadopoulos, S.M., Sonntag, V.K.: Caspar plate fixation for the treatment of complex hangman’s fractures. Neurosurgery 30, 761–764 (1992). discussion 764–765 92. Umebese, P.F., Orhewere, F.A.: Hangman’s fracture in head injury. East Afr Med J 66, 611–614 (1989) 93. Vaccaro, A.R., Madigan, L., Bauerle, W.B., et al.: Early halo immobilization of displaced traumatic spondylolisthesis of the axis. Spine (Phila Pa 1976) 27, 2229–2233 (2002) 94. Verheggen, R., Jansen, J.: Hangman’s fracture: arguments in favor of surgical therapy for type II and III according to Edwards and Levine. Surg Neurol 49, 253–261 (1998). discussion 261–252 95. Vichard, P., Mirbey, J., Pinon, P.: Value of anterior arthrodesis in the treatment of fractures of the pedicles of the axis (author’s transl). J Chir (Paris) 118, 565–572 (1981)

12  Fractures of the Ring of Axis (Hangman Type Fractures) 96. Vieweg, U., Schultheiss, R.: A review of halo vest treatment of upper cervical spine injuries. Arch Orthop Trauma Surg 121, 50–55 (2001) 97. Vlach, O., Leznar, M., Bayer, M.: Diagnosis, classification and treatment of so-called hangman’s fractures. Acta Chir Orthop Traumatol Cech 55, 456–466 (1988) 98. White 3rd, A.A., Johnson, R.M., Panjabi, M.M., et  al.: Biomechanical analysis of clinical stability in the cervical spine. Clin Orthop Relat Res 109, 85–96 (1975) 99. White 3rd, A.A., Moss, H.L.: Hangman’s fracture with nonunion and late cord compression. A case report. J Bone Joint Surg Am 60, 839–840 (1978) 100. Wilson, A.J., Marshall, R.W., Ewart, M.: Transoral fusion with internal fixation in a displaced hangman’s fracture. Spine (Phila Pa 1976) 24, 295–298 (1999) 101. Wood-Jones, F.: The ideal lesion produced by judicial hanging. Lancet 181, 53 (1913) 102. Yarbrough, B.E., Hendey, G.W.: Hangman’s fracture resulting from improper seat belt use. South Med J 83, 843–845 (1990) 103. Zapletal, J., de Valois, J.C.: Radiologic prevalence of advanced lateral C1-C2 osteoarthritis. Spine (Phila Pa 1976) 22, 2511–2513 (1997) 104. Zavanone, M., Guerra, P., Rampini, P., et  al.: Traumatic fractures of the craniovertebral junction. Management of 23 cases. J Neurosurg Sci 35, 17–22 (1991)

Miscellaneous C2 Fractures

13

P. Suchomel and J. Hradil

The category of miscellaneous axis fractures was introduced by Hadley [6] to cover all non-odontoid and nonhangman like fractures. These “non-classifiable C2 injuries” represent about one quarter of all C2 fractures. Apart from this “broad” definition, there are narrower alternatives as well. Some authors distinguish certain groups of fractures as separate categories (namely, axis body fractures and tear-drop fractures), excluding them from “miscellaneous” category. However, any detailed categorization faces problems with terminology, classification and, of course, certain uniqueness of this class of axis injuries. The literature on miscellaneous fractures is scarce and terminology is not consistent. There is no firm evidence to guide the treatment strategies.

13.1 Incidence and Classification For the purpose of this text, we use Hadley’s “broad definition”. Further groups can be identified within this category. These include (a) coronal, (b) sagittal, (c) transverse, and (d) burst fractures of the axis body, (e) tear drop fractures, (f) non-hangman injuries to lamina and spinous process, (g) fractures of the superior facet area, and (h) fractures through the vertebral foramen (transverse process). The borders between the groups cannot be precisely defined and this categorization serves only as a general guide. In reality, every “atypical” axis injury needs a strictly individual approach. It is very difficult to estimate the incidence of miscellaneous axis fractures within the population. The literature

P. Suchomel and J. Hradil Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St.10, 46063 Liberec, Czech Republic

is very limited and certain fracture patterns are clearly underreported, for example severe fractures/dislocations in polytrauma patients with early death of the patient. While mild cases are treated in local hospitals (fracture of spinous process), certain patterns are generally underdiagnosed (impression through the superior facet joint). The reported incidence depends on types of fractures considered as “miscellaneous” and therefore, varies between 19 and 32% of axis fractures. Non-odontiod/ non-hangman fractures represent about 20% of axis trauma referred to tertiary spine center, according to the largest series of Greene. This number does not change in time, according to single institution data covering different periods of time [5–7]. Hadley et al. [6] described characteristics and management of 23 miscellaneous fractures: 8 body fractures, 7 lateral mass fractures, 3 lamina fractures, 2 pedicle fractures, 2 spinous process fractures, and 1 pars interarticularis fracture. The authors applied halovest or SOMI brace for 8–12 weeks. Rigid/soft collar or varying “descendent” combinations of external immobilization were applied in more stable injuries. Further reports from the same institution from 1989 [7] and 1997 [5] with a total of 67 non-odontiod and non-hangman fractures out of 340 axis fractures brought no change in diagnostic or treatment rationale. The reported rate of nonunion leading to delayed surgery was 1.6%, no further outcome measures were provided. Fujimura [3] suggested a classification of axis body fractures based on 31 cases including 17 cases of sagittal fractures. Authors distinguish four categories: avulsion, transverse, burst, and sagittal. The results of conservative treatment were mostly ­satisfactory with the exception of sagittal fractures, where 8 of 17 cases treated by external immobilization required C1-C2 fusion due to neck pain based on early C1/C2 joint degeneration. The authors

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_13, © Springer-Verlag Berlin Heidelberg 2011

197

198

recommended surgical intervention in cases with severe malalignment of the atlantoaxial joint. German et al. [4] reported 21 cases (9.7%) of vertical axis fractures (16 coronal, 5 sagittal) in 208 cases of upper cervical trauma. Some of coronal cases would surely be classified as atypical hangman’s fractures by other authors. The authors claimed good results with conservative management but provide no patient-based outcome measures. Korres et al. [10] published axis body fractures in 11% of 172 cases of axis trauma. Additionally, these authors also distinguish isolated fractures of the “lateral mass” (2%) and “miscellaneous” fractures (19%). The treatment involved skeletal traction in bed for 2–6 weeks followed by immobilization in Minerva or halo-vest. They obtained satisfactory results, however, no functional outcome measures were provided. Another article by the same author [11] reported 14 cases of avulsion fractures due to extension injury treated conservatively by external immobilization. Benzel et al. [1] reviewed 15 cases of axis body fractures (12 coronal and 3 sagittal) and dealt with possible biomechanics involved in traumatic mechanisms. Based on evaluation of these two groups, the authors proposed a three-category classification. They suggested adding type III odontiod fractures and atypical hangman type fractures into the category of C2 body fractures. No recommendations concerning treatment strategy based on such classification have been proposed. Taller et al. [23] included two cases of complete avulsion of posterior axis wall into the series of hangman’s fractures. CT-guided transisthmic posterior fixation using lag screws allowed safe reduction of fracture gap running through C1/C2 joints and provided immediate stability. Burke et al. [2] reports 31 miscellaneous fractures including 21 tear drop fractures. This number stands out in terms of incidence and it is most probably a matter of classification criteria.

13  Miscellaneous C2 Fractures

with approximately 20–25% of them with concomitant craniocerebral injury.

13.3 Radiology Radiological evaluation is similar to other UCS injuries. Plain radiographs are performed in self-presenting patients as a first-line assessment and usually only identify dislocations but never delineate the exact fracture pattern (Fig.  13.1). Again, thin sliced CT with reconstructions is the mainstay when it comes to defining the exact fracture and dislocation morphology (Fig.  13.2). MRI evaluation is mandatory to depict the status of ligamentous structures and to exclude neural compromise and/or other level soft tissue injury (Fig.  13.3). Other more sophisticated investigations (dynamic films in cooperating individuals, dynamic CT and/or MRI, MRA, and CTA) are added if necessary.

13.4 Treatment Strategy and Our Preference Most authors prefer conservative treatment in so called “miscellaneous fractures”; however, their approach slightly differs according to specific fracture pattern.

13.2 Clinical Symptoms As with other UCS injuries, the majority of admitted patients may only have nonspecific neck pain frequently radiating to the occiput and some range of motion limitation. Some of them guard their neck against pain by holding their head up with hands. Neurological deficits are rarely seen, however, if present, can vary from mild cranial nerve palsies to pentaplegia and/or coma. This is more often seen in polytrauma victims

Fig. 13.1  Plain laterogram showing fracture of C2 body with anterior C2-3 dislocation

199

13.4  Treatment Strategy and Our Preference

surgical correction with regard to long-term outcome and correct coronal and sagittal spine profile. Generally, we divide miscellaneous fractures into two main groups – fractures of C2 vertebral body and the others.

13.4.1 Coronal Axis Body Fractures

Fig.  13.2  Axial CT depicting the course of oblique fracture. The same patient from Fig. 13.1

As mentioned previously, our strategy is more aggressive in case of instability visible on admission radiographs or provoked by dynamic films in cooperating patients. Nonreducible fracture dislocations, especially those extending into articular surfaces or compressing neural structures have to be considered for

a

Fig. 13.3  Sagittal MRI in T2 sequence in two different patients. (a) Simultaneous comminutive fracture of C2 body and luxation fracture at the level of C7-T1 causing a transverse spinal cord lesion. (b) Unstable comminution of C2 body endangering the spinal cord

Many authors consider these patterns as “atypical” or “unusual” hangman’s fractures [15, 20, 21, 23]; some separate them as a distinct category of C2-body trauma [1, 4]. Hangman’s fractures often involve at least some part of posterior wall of the C2 body and therefore, a clear border separating both categories cannot be identified. The exact site of the fracture ranges from small infractions of the inferolateral aspects of the axis anterior column (Fig.  13.4) to complete posterior wall avulsions with anterior extensions through superior facet joints, transverse processes/vertebral foramina (Fig. 13.5). The superior fracture line leaves the dens attached to the anterior fragment. The inferior line runs through the inferior endplate. Fractures running through the anterior aspect of the axis should be considered either as a transverse fracture of C2 body or type III fractures of the dens (Fig.  13.6). With the majority of these fractures presenting with asymmetry,

b

200 Fig. 13.4  CT of coronal fracture of C2 with partial involvement of posterior wall (more atypical hangman’s type). (a) Only partial one sided posterior wall involvement. (b) Nearly all the posterior wall is abrupted

13  Miscellaneous C2 Fractures

a

b

a

b

Fig. 13.5  Stable coronal C2 fracture involving important part of posterior body wall (more fracture of the C2 body) which was successfully treated in Philadelphia collar. (a) Sagittal MRI. (b) Axial CT scan

cases with no clear classification can easily be encountered.

13.4.1.1 Our Preference

Fig. 13.6  CT in coronal plane showing transverse C2 body fracture (can be classified as deep type III odontoid fracture)

Vertical fractures of C2 vertebra (i.e., hangman’s fractures, atypical hangman’s fractures, and coronal C2 body fractures) are clearly a group of trauma with smooth spectrum of patterns and ­corresponding biomechanics. Although there is an obvious tendency towards external immobilization in the English-speaking literature, we prefer early anterior surgery in unstable cases [21]. It provides immediate stability and also a chance to reduce any displaced fragments (Fig.  13.7). In stable fractures (disc not injured) with fracture distraction greater than 3  mm on initial CT scans, reduction of

201

13.4  Treatment Strategy and Our Preference Fig. 13.7  Coronal split fracture of C2 body treated with anterior graft and plate fusion. (a) CT sagittal reconstruction demonstrating the fracture dislocation. (b) MRI in T2 sequence showing the traumatic C2-3 disc involvement. (c) Anterior graft and plate fusion. Note bicortical screw purchase

a

b

c

fracture gap can be better achieved by posterior approach using transpedicular lag screws [12]. We [23] showed a very safe method of CT-guided screw placement, however, indications for posterior procedure remain limited (as was described in Chap. 12).

13.4.2 Sagittal Axis Body Fractures Sagittal pattern is not limited to the axis body, but it often involves true axis pedicles and also the area covered by superior facet joint. The fracture is frequently unilateral

and/or oblique (Fig. 13.8). Superior aspect of the fracture plane is located close to the base of dens or within the medial aspect of the superior facet joint whereas inferiorly, the fracture often involves the C2-3 disc space. High velocity axial load through the vertex of the skull is the major causative force and severe craniocerebral trauma is a common-associated injury [4]. However, pure axial loading usually results in Jefferson burst fracture of the atlas. This means that additional shear forces [8], preimpact lateral bending or rotation in C1-2 joints are necessary to create this injury. These forces leave traces such as infractions of the base of the dens, unilaterality of fracture site and anteroposterior displacement of fragments.

202 Fig. 13.8  Sagittally oriented C2 body fracture. (a) Axial CT scan. (b) Vertebral body split visible on coronal CT reconstruction. (c) Fusion in malposition on axial CT scan (d) “Fat C2 body sign” and incomplete spontaneous C2/3 fusion on lateral radiograph 6 months after the injury

13  Miscellaneous C2 Fractures

a

b

c

d

13.4.2.1 Our Preference Frequently, fracture union can be achieved by conservative means [3] as there is a large portion of cancellous bone involved. However, cases of severely malaligned C1-C2 joint should be considered for C1-C2 fusion due to poor functional outcomes (Fig. 13.9).

13.4.3 Transverse Axis Body Fractures “Deep” type III odontoid fractures should be mentioned here as several authors [1, 4] advocate their re-assignment to the group of axis body fractures. These fractures surely run through the superior part of the C2 body, however, they separate the dens from the posterior elements and clinical consequences of this functional issue provide a strong argument for their assignment to the odontoid group. True transverse axis fractures are extremely rare,

Fig. 13.9  Axial CT of oblique sagittal C2 body fracture extending to the upper facet

203

13.4  Treatment Strategy and Our Preference

they usually involve extension mechanisms and leave the odontoid process connected to the posterior elements [9]. The major difference between the groups is a potential for dislocation (and thus, spinal cord compression). Type III odontoid fractures are quite stable, but true transverse body fractures are frequently unstable. Extension mechanism of injury makes them prone to associated discoligamentous disruptions of the C2/3 intervertebral connection. To illustrate the aforementioned diversity of axis fractures, it is interesting to include a case of Maki [14] with a “chance-type” fracture produced by bending over anterior fulcrum (steering wheel). 13.4.3.1 Our Preference As mentioned in Chap. 11, our inclusion criterion for a fracture to be considered a type III odontoid fracture, is that the fracture line has to reach at least one superior articular surface, cranially. However, the caudal border distinguishing between type III odontoid fracture and horizontal fracture of the body has not been established up to now. Perhaps, those injuries, where at least one fracture line is located below the C2 upper facet, can be classified as C2 body fractures as well. Despite a large fracture surface area, they are often unstable and incapable of maintaining sagittal alignment. In unstable injuries where the horizontal fracture line does not reach the C2-3 disc space simultaneously but with lack of substantial amount of bone available at the C2 base for direct odontoid screw purchase, we prefer a posterior fixation and fusion (Fig.  13.10). If there is sufficient inferior bone present at the base of C2 body, then the double screw anterior odontoid a

b

Fig. 13.10  Unstable horizontal fracture of C2 body disconnecting the odontoid from posterior elements treated with posterior graft and transarticular screw fusion. (a) Fracture reduced under

fixation can be performed. Additional anterior graft and plate can supplement the construct in cases of simultaneous C2-3 disc rupture. Limits of previously described procedures are written in the section “Specific techniques” of Chap. 6.

13.4.4 Burst Fractures of Axis Body Multiple fragment injuries are a result of extreme axial loading as a primary force. Axis body and the pedicles are strong and this pattern is rare as most axial loads result in Jefferson fractures of atlas or subaxial spinal injury. As in sagittal patterns, shear forces and/or rotation and/or lateral bending help to transfer enough force through the lateral masses of the atlas. High velocity increases the probability of burst fractures, even in cases with concurrent disruption of atlas ring. 13.4.4.1 Our Preference Conservative treatment using external immobilization is often a reasonable treatment option; however, in polytrauma patients who are unconscious and/or dependent on mechanical ventilation, posterior surgical stabilization can dramatically increase their mobility when compared to restrictions offered by cranial traction or halo-vest fixation. The other important positive factor is much improved respiratory care (Figs. 13.11 and 13.12). Also, in elderly patients, surgical fixation and fusion (Fig. 13.13) can offer a significant advantage over a halo-vest immobilization (See Chap. 6). c

traction. (b) Redislocation of the fracture after traction release. (c) Postoperative status

204

13  Miscellaneous C2 Fractures

c

a

b

Fig. 13.11  Comminution of C2 body extending from odontoid to the vertebral base in an unconscious polytrauma patient with unstable chest on ventilatory support. (a) Axial CT showing the

a

b

Fig.  13.12  Another polytrauma, unconscious patient with C2 body comminution and dominant odontoid type fracture and AA instability. (a) CT in sagittal reconstruction. (b) CT in coronal

comminution. (b) Intraoperative image. (c) C1-3 fixation with modular fixator

c

plane. (c) Permanent fixation C1-3 with prolonged Harms fixator to C3 lateral mass screws

205

13.4  Treatment Strategy and Our Preference Fig. 13.13  Comminuted C2 fracture in old patient who refused the halo-vest offer. (a) Sagittal reconstruction showing C2 body comminution. (b) Coronal reconstruction depicting concurrent left side facet damage. (c) Intraoperative view with fixator in place augmented by posterior interlaminar autologous bone grafts. (d) Postoperative lateral film

a

b

c

d

13.4.5 Tear Drop Fractures Flexion with axial loading or extension with distraction are the major causes of tear drop fractures. Treatment is focused on the discoligamentous injury at the C2/3 segment and C2-3 fixation and fusion is a method of choice (Fig.  13.14). Despite this, some authors prefer external immobilization [11].

stable. The ability to support head is not significantly altered and there is no danger of spinal cord injury. Treatment with a hard collar is sufficient as there is no risk of joint incongruence and later degeneration. These fractures are, probably, highly underreported in the literature as they usually do not require a referral to a specialized center. Their follow-up is also unclear.

13.4.6 Non-Hangman Injuries to Lamina and Spinous Process

13.4.7 Fractures of the Superior Facet Area

Direct trauma is probably the only cause in otherwise healthy individual. The spinous process of axis supports extensive muscular structures, but fractures located posterior to inferior facet joints are considered

Fracture lines often run through the area supporting superior facet surface (some authors speak about “lateral mass,” “articular mass,” or “superior articular process”). Apart from transverse/coronal fracture line

206 Fig. 13.14  Tear drop fracture with posterior dislocation. (a) Preoperative lateral film. (b) Graft and plate fusion on lateral X-rays

13  Miscellaneous C2 Fractures

a

extensions and sagittal/burst local patterns, there is also a distinct type of unilateral impression fracture. Trabecular bone mass underlying the superior facet joint collapses and becomes dense on plain x-ray. This fracture is a result of milder axial loads in cases similar to sagittal/burst fractures. It is likely to be underdiagnosed as a standalone injury; however, there is a strong association with odontoid fractures with lateral displacement, according to the literature [8, 18]. The primary treatment is usually conservative. Cases with odontoid fractures can be treated by anterior odontoid screw fixation. C1-C2 fusion is reserved for cases with painful C1/2 joint due to incongruence and/or subsequent degenerative changes.

13.4.7.1 Our Preference The majority of fractures of superior facets are a part of differently classified injury. As mentioned earlier, type III odontoid and atypical hangman-type fractures often involve the articular surface. However, although rare, there is also an injury pattern where the axial load is transferred symmetrically to the articular pillars of C2 (without atlantal bursting) and fracturing of only the upper facets with eventual body extension (Fig.  13.15). Those can be hardly treated conservatively because stable symmetrical distraction

b

is needed to hold appropriate alignment in coronal plane. In such situations, it is reasonable to consider a temporary C1-3 fixation in order to maintain adequate alignment.

13.4.8 Fractures Through the Transverse Foramen Extensions into the transverse process and vertebral foramen are a frequent finding, especially in coronal fracture patterns. Although mostly asymptomatic, there are reports of severe consequences of a vertebral artery injury [13, 16, 17, 19]. Some authors performed intraoperative angiograms to detect the injury and to increase the safety of fragment reposition [22].

13.5 Combination C1-2 Fractures As it is described in Chap. 14, the treatment choice of combined C1-2 fractures is dictated by the most important instability in each individual case. If the miscellaneous fracture type is the more unstable part of the injury, then the treatment follows the above-mentioned recommendations.

207

References

a

b

d

c

e

Fig. 13.15  Simultaneous injury of C2 and subaxial spine (patient from Fig. 13.3a). (a) Fracture of C2 body and posterior C3 arch on CT reconstructed image. (b) C2 stability in flexion. (c) Extension showing instability – dislocation. (d) CT in coronal plane ­reconstruction clearly documenting both articular pillar fractures. (e) Plain lateral radiograph showing C1-3 temporal fixation in reduced position. The 360° reduction-fixation for luxation fracture at the level C7-T1 performed during the first stage procedure

References   1. Benzel, E.C., Hart, B.L., Ball, P.A., et al.: Fractures of the C-2 vertebral body. J Neurosurg 81, 206–212 (1994)   2. Burke, J.T., Harris Jr., J.H.: Acute injuries of the axis vertebra. Skeletal Radiol 18, 335–346 (1989)   3. Fujimura, Y., Nishi, Y., Kobayashi, K.: Classification and treatment of axis body fractures. J Orthop Trauma 10, ­536–540 (1996)

  4. German, J.W., Hart, B.L., Benzel, E.C.: Nonoperative management of vertical C2 body fractures. Neurosurgery 56, 516–521 (2005). discussion 516–521   5. Greene, K.A., Dickman, C.A., Marciano, F.F., et al.: Acute axis fractures. Analysis of management and outcome in 340 consecutive cases. Spine (Phila Pa 1976) 22, 1843–1852 (1997)   6. Hadley, M.N., Browner, C., Sonntag, V.K.: Axis fractures: a comprehensive review of management and treatment in 107 cases. Neurosurgery 17, 281–290 (1985)

208   7. Hadley, M.N., Dickman, C.A., Browner, C.M., et al.: Acute axis fractures: a review of 229 cases. J Neurosurg 71, ­642–647 (1989)   8. Hahnle, U.R., Wisniewski, T.F., Craig, J.B.: Shear fracture through the body of the axis vertebra. Spine (Phila Pa 1976) 24, 2278–2281 (1999)   9. Jakim, I., Sweet, M.B.: Transverse fracture through the body of the axis. J Bone Joint Surg Br 70, 728–729 (1988) 10. Korres, D.S., Papagelopoulos, P.J., Mavrogenis, A.F., et al.: Multiple fractures of the axis. Orthopedics 27, 1096–1099 (2004) 11. Korres, D.S., Zoubos, A.B., Kavadias, K., et al.: The “tear drop” (or avulsed) fracture of the anterior inferior angle of the axis. Eur Spine J 3, 151–154 (1994) 12. Leconte, P.: Fracture et luxation des deux premieres vertebres cervicales. In: Judet, R. (ed.) Luxation Congenitale de la Hanche. Fractures du Cou-de-pied Rachis Cervical. Actualites de Chirurgie Orthopedique de l’Hospital Raymond-Poincare, vol. 3, pp. 147–166. Masson et Cie, Paris (1964) 13. Lohnert, J., Latal, J.: Fracture of the axis–surgical treatment. II. Axial isthmus. Acta Chir Orthop Traumatol Cech 60, 47–50 (1993) 14. Maki, N.J.: A transverse fracture through the body of the axis. A case report. Spine (Phila Pa 1976) 10, 857–859 (1985) 15. Marotta, T.R., White, L., TerBrugge, K.G., et al.: An unusual type of hangman’s fracture. Neurosurgery 26, 848–850 (1990). discussion 850–841

13  Miscellaneous C2 Fractures 16. Okuchi, K., Fujioka, M., Konobu, T., et  al.: A case of Hangman’s fracture associated with vertebral arteriovenous fistula treated with trapping. No Shinkei Geka 22, 55–59 (1994) 17. Pelker, R.R., Dorfman, G.S.: Fracture of the axis associated with vertebral artery injury. A case report. Spine (Phila Pa 1976) 11, 621–623 (1986) 18. Signoret, F., Feron, J.M., Bonfait, H., et al.: Fractured odontoid with fractured superior articular process of the axis. Report of three cases. J Bone Joint Surg Br 68, 182–184 (1986) 19. Simonsen, J.: Massive subarachnoid haemorrhage and fracture of the transverse process of the atlas. Med Sci Law 16, 13–16 (1976) 20. Starr, J.K., Eismont, F.J.: Atypical hangman’s fractures. Spine (Phila Pa 1976) 18, 1954–1957 (1993) 21. Suchomel, P., Hradil, J., Barsa, P., et al.: Surgical treatment of fracture of the ring of axis – “hangman’s fracture”. Acta Chir Orthop Traumatol Cech 73, 321–328 (2006) 22. Takahashi, T., Tominaga, T., Ezura, M., et al.: Intraoperative angiography to prevent vertebral artery injury during ­reduction of a dislocated hangman fracture. Case report. J Neurosurg 97, 355–358 (2002) 23. Taller, S., Suchomel, P., Lukas, R., et al.: CT-guided internal fixation of a hangman’s fracture. Eur Spine J 9, 393–397 (2000)

Multiple Fractures of Axis and Atlas-Axis Fracture Combinations

14

P. Suchomel and J. Hradil

14.1 Multiple Fractures of the Axis An attempt to classify multiple axis fractures based on traditional approaches to categorizing spine fractures is difficult due to the complexity of this particular vertebra and the number of many different combinations of ­fractures. Some of the “single” fractures (by tradition) involve two separate fracture sites (hangman’s fracture, sagittal body fracture) and some of the single fracture lines run through multiple areas of the axis (coronal fractures involving body, pedicles, superior facet joints, and transverse processes). The exact incidence of multiple fractures, thus, depends on the number of categories distinguished by particular authors. The literature addressing this topic is very limited and there is no evidence concerning treatment methods. All multiple fractures of the axis should, therefore, be evaluated on strictly individual case-by-case basis. The treatment should be tailored according to the opinion and personal experience of the treating surgeon. The majority of authors recommend conservative approach with external immobilization. However, several fracture types allow and benefit from effective surgical intervention. Korres et al. reported a combination of two or even three distinct axis fracture types in a single patient in 5% of their series [14]. There were combinations of hangman’s fracture and tear drop (three cases), ­odontoid fracture and hangman’s fracture (two cases), odontoid fracture and “lateral mass” fracture (two cases), and one case of combined odontoid, tear drop, and hangman’s fractures. Daum and Archer described a combination of

P. Suchomel and J. Hradil Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic

odontoid and hangman’s fracture [3]. Signoret et  al. analyzed possible biomechanical background of odontoid fractures and compressive fractures of the superior facet joints [15]. Lateral bending and lateromedial forces leading to compressive fractures of the area supporting superior facet joint and odontoid fracture are a frequent finding in such cases. Hahnle et al. reported compressive fractures of the superior facet joint area with specific fractures of the odontoid base, resulting in dens tilt off the mid-sagittal plane [12]. Iizuka et al. reported a case of an atypical hangman’s fracture and additional fracture of the spinous process [13].

14.1.1 Our Preference The combination of teardrop fracture and hangman’s fracture is probably the most common finding. Teardrop fracture can change treatment rationale as it is strongly suggestive of disc injury. In our opinion, anterior discectomy, interbody graft fusion, and plate fixation (ACDF) should be performed in such situations (Fig. 14.1). In questionable cases, dynamic lateral radiographs should be obtained in cooperating patients, especially in otherwise “stable” type I hangman’s fractures. The second most common scenario is a combination of odontoid and hangman’s fractures. Since there is an expected high rate of non-union in type II and shallow type III odontoid fractures treated by external immobilization, we are convinced that surgical treatment is the method of choice here. A combination of an odontoid screw and C2-3 ACDF is a viable and simple option in this situation (Fig. 14.2). This approach is also suitable for the treatment of a “triple” hangman-tear drop-odontoid fracture. Other fracture combinations can be seldom seen and therefore, require individual evaluation of the degree of

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_14, © Springer-Verlag Berlin Heidelberg 2011

209

210

14  Multiple Fractures of Axis and Atlas-Axis Fracture Combinations

Fig. 14.1  C2 tear drop fracture with simultaneous hangman type fracture. (a) Lateral radiograph. (b) Postoperative picture showing graft and plate anterior fusion. Note triple screw introduction into C3 body to increase the strength of plate lever arm.

instability, their healing potential, and individual patient characteristics. When conservative treatment is selected in borderline or unclear cases, the need for a frequent and close patient follow-up cannot be stressed enough.

14.2 Combined Atlas-Axis Fractures Virtually, any fracture of axis can be accompanied by atlas fracture and vice versa. The published frequency of combined C1-2 fractures is approximately 3% of all acute cervical spine injuries [5, 9]. Their treatment is strictly individual. Most authors are in favor of external immobilization, but there are cases that may benefit from early surgical solution [1, 2, 4, 9]. Gleizes et al. [7] provided a comprehensive analysis of cervical spine fractures associated with upper cervical spine (UCS) trauma. A total of 784 cervical spine injuries were evaluated and of those, 116 (14.8%) involved the UCS. Nineteen cases (16.4% of UCS injuries) of combinations of C1-C2 fracture were found in this series. Double fractures were found in 17 and triple fractures in the remaining 2 patients. Within the pool of 19 UCS fracture combinations, hangman’s with odontoid fractures were present in 21.0%, odontoid with C1 posterior arch in 31.6%, odontoid with a Jefferson fracture in 10.5%, and odontoid with a C2 superior facet fracture in 10.5%. Remaining 26.4% of cases were of a unique fracture pattern.

Dickman et al. [5] presented 25 cases of atlas-axis combinations that represented 43% of all atlas fractures and 16% of all axis fractures. Atlas fractures included Jefferson type fractures in 40%, posterior arch in 28%, unilateral ring in 24%, and lateral mass in 8%. Axis fractures were represented by odontoid type II injuries in 40%, miscellaneous fractures in 28%, odontoid type III in 20%, and hangman’s fractures in 12%. Neurological deficit was higher than in isolated atlas or axis fractures and reached 12% across all comers with combined C1/2 fractures. A large series of patients with axis traumatic injuries accumulated in the same center by Hadley et al. and Greene et al. did not deal with fracture combinations in detail, nevertheless, they described atlas fractures in 41% of axis injuries [8, 10, 11]. We found 17.5% of atlas fractures in a published series of hangman’s fractures [16]. In his fundamental piece, Effendi registered posterior atlas arch fractures in 6.0% and odontoid fractures in 1.5% together with fractures of the ring of axis [6].

14.2.1 Our Preference The general principles of treatment of combined atlasaxis fractures are not based on any available evidence, series analyses, or outcome measures yet. They are based mostly on individual surgeon’s understanding

14.2  Combined Atlas-Axis Fractures

211

Fig. 14.2  Combination of a shallow odontoid type III and hangman’s fractures treated with single anterior odontoid screw and subsequent C2-3 ACDF. (a) Plain lateral film. (b) Axial CT scan

showing hangman type fracture. (c) Sagittal CT reconstruction depicting odontoid type II fracture Postoperative lateral (d) and AP (e) view of screw and plate fixation

and interpretation of the injury impact on ­biomechanical properties of the bony and discoligamentous apparatus of the AA complex. If both the atlas and axis injuries are evaluated as stable and non-displaced, external bracing can lead to a successful union. However, if the fracture of any one or even both vertebrae is considered to be unstable or displaced, or if the discoligamentous apparatus is not

able to maintain adequate spinal alignment, surgical stabilization (usually following traction reduction) can increase the chance of healing quicker and in an appropriate UCS balance. The location of major instability often guides our treatment approach selection. As most of the combinations include odontoid and hangman type fractures, their fixation, if indicated, has to be done first. If the

212

14  Multiple Fractures of Axis and Atlas-Axis Fracture Combinations

concomitant atlas ring injury is stable (anterior, posterior arch) a hard collar for 6–12 weeks with regular radiological follow-up can be sufficient. However, if the atlas fracture is unstable (transverse atlantal ligament dysfunction etc.) or not maintaining alignment,

Fig. 14.3  Combination UCS injury in a 64 year-old man including shallow odontoid type III, hangman and Jefferson like fractures treated with anterior fixation in single stage. (a) Preoperative plain laterogram. (b) Axial CT scan of hangman like coronal fracture of C2. (c) Plain lateral view of C2-3 graft and plate fixation together with single anterior screw odontoid osteosynthesis and double transarticular screw AA fusion. (d) The same patient on AP film

simultaneous AA fixation can be performed. The combined procedure can often be done from a single anterior approach (Fig. 14.3). Combination of unstable atlas and axis fractures can also be treated with posterior AA or OC fixation and fusion (Figs. 14.4. and 14.5).

14.2  Combined Atlas-Axis Fractures Fig. 14.4  Case of miscellaneous C2 fracture combined with AA instability caused by TAL attachment abruption treated with posterior transarticular AA fusion. (a) Axial CT of C2 lateral pillar fracture. (b) Axial CT showing the fragments of abrupted TAL. (c) dynamic lateral radiogram radioraph in flexion documenting increased AADI confirming the AA instability. (d) Posterior C1-2 transarticular instrumentation supplemented with graft and wire sublaminar fixation

Fig. 14.5  Rotatory atlanto-axial subluxation with concomitant C2 articular pillar and odontoid process type II fracture treated with posterior fusion according to Goel-Harms. (a) Axial CT showing abnormal odontoid position in rotatory C1 subluxation. (b) Coronal plane CT (c) 3D CT anterior view of fracture of C2 articular process. (d) Lateral plain film of posterior fixation

213

214

14  Multiple Fractures of Axis and Atlas-Axis Fracture Combinations

References   1. Agrillo, U., Mastronardi, L.: Acute combination fracture of atlas and axis: “triple” anterior screw fixation in a 92-year-old man: technical note. Surg Neurol 65, 58–62 (2006)   2. Apostolides, P.J., Theodore, N., Karahalios, D.G., et  al.: Triple anterior screw fixation of an acute combination atlasaxis fracture. Case report. J Neurosurg 87, 96–99 (1997)   3. Daum, W., Archer, C.R.: Fracture of the odontoid associated with pedicle fracture of the axis: a previously undescribed entity. J Trauma 17, 381–386 (1977)   4. Dean, Q., Jiefu, S., Jie, W., et al.: Minimally invasive technique of triple anterior screw fixation for an acute combination atlas-axis fracture: case report and literature review. Spinal Cord 48, 174–177 (2010)   5. Dickman, C.A., Hadley, M.N., Browner, C., et al.: Neurosurgical management of acute atlas-axis combination fractures. A review of 25 cases. J Neurosurg 70, 45–49 (1989)   6. Effendi, B., Roy, D., Cornish, B., et al.: Fractures of the ring of the axis. A classification based on the analysis of 131 cases. J Bone Joint Surg Br 63, 319–327 (1981)   7. Gleizes, V., Jacquot, F.P., Signoret, F., et al.: Combined injuries in the upper cervical spine: clinical and epidemiological data over a 14-year period. Eur Spine J 9, 386–392 (2000)

  8. Greene, K.A., Dickman, C.A., Marciano, F.F., et al.: cute axis fractures. Analysis of management and outcome in 340 consecutive cases. Spine (Phila Pa 1976) 22, 843–1852 (1997)   9. Guiot, B., Fessler, R.G.: Complex atlantoaxial fractures. J Neurosurg Spine 91, 139–143 (1999) 10. Hadley, M.N., Browner, C., Sonntag, V.K.: Axis fractures: a comprehensive review of management and treatment in 107 cases. Neurosurgery 17, 281–290 (1985) 11. Hadley, M.N., Dickman, C.A., Browner, C.M., et al.: Acute axis fractures: a review of 229 cases. J Neurosurg 71, 642–647 (1989) 12. Hahnle, U.R., Wisniewski, T.F., Craig, J.B.: Shear fracture through the body of the axis vertebra. Spine (Phila Pa 1976) 24, 2278–2281 (1999) 13. Iizuka, H., Shimizu, T., Hasegawa, W., et al.: Fractures of the posterior part of the body and unilateral spinous process of the axis: a case report. Spine (Phila Pa 1976) 26, 528–530 (2001) 14. Korres, D.S., Papagelopoulos, P.J., Mavrogenis, A.F., et al.: Multiple fractures of the axis. Orthopedics 27, 1096–1099 (2004) 15. Signoret, F., Feron, J.M., Bonfait, H., et al.: Fractured odontoid with fractured superior articular process of the axis. Report of three cases. J Bone Joint Surg Br 68, 182–184 (1986) 16. Suchomel, P., Hradil, J., Barsa, P., et al.: Surgical treatment of fracture of the ring of axis – “hangman’s fracture”. Acta Chir Orthop Traumatol Cech 73, 321–328 (2006)

Acute Traumatic Atlantoaxial Dislocation (AAD) in Adults

15

P. Suchomel and R. Fricˇ

Traumatic atlantoaxial dislocation (AAD) occurs less frequently than AOD. This, usually fatal injury, [1, 2] is commonly found as a consequence of high velocity trauma. Generally, due to traumatic impact, the atlas can be displaced in any direction in respect to C2 vertebra. C1 dislocation is frequently accompanied by a fracture of the odontoid process but other UCS fractures can also be present. However, AAD as a result of pure ligamentous injury is very rare. Traumatic AADs can be divided into three categories: translational (AP and lateral), rotatory, and distractive [34]. The adult traumatic translational AAD is most frequently caused by AA instability related to odontoid or other C2 fractures (Fig.  14.4, Chap. 14) and or atlas fractures (Fig.  10.12, Chap. 10). The pure incompetence of transverse atlantal ligament (TAL) is much less frequent (Figs. 10.3 and 10.11, Chap. 10) [24]. The isolated ligamentous rotatory traumatic dislocations are extremely rare in adults [4, 6, 20, 27, 33] and most of the reported cases are combined with C2 fractures (Fig. 14.5, Chap. 14) [8, 13, 17, 21, 26, 34]. The classification dividing nontraumatic AA rotatory fixations in children and young adults into two (often four) categories [10, 31] can be used to classify the degree of traumatic rotatory displacement in adults as well; however, one has to be aware not to overestimate the rotational displacement as it can be, in fact,

P. Suchomel Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic R. Fricˇ Department of Neurosurgery, Rikshospitalet, Oslo University Hospital, Sognsvannsveien 20, 0027 Oslo, Norway

within normal range of AA joint movement. Recently, Mönckeberg et al. [25] reported CT analysis performed on 40 healthy volunteers (actually, colleagues from author’s own institution) clearly documenting that during maximal voluntary rotation (38° on an average) to one side, 70% of AA joint facet surface is uncovered on an average, and that the full facetal contact is achieved only in neutral position. Often, distractive force can also cause the AAD without vertebral fractures, however, a very important warning case of concomitant fracture was reported by Przybylski and Welsch [30]. They had a patient referred from another hospital with type III odontoid fracture. Quadriplegia developed with 5 lbs of traction used, to reduce the angular odontoid process displacement. Horizontal AA complex disruption simultaneous with type III odontoid fracture revealed on CT was responsible for this significant vertical instability with AAD. The first who described distractive AAD without concomitant fractures were Haralson and Boyd in 1969 [16]. Until 1980, only two other similar cases were published [28, 32]. With improved rescue services and development of modern imaging more cases of pure ligamentous AA disruption were reported later on [7, 11, 18, 19, 22, 29, 36, 37]; nevertheless, this often fatal injury is still a rarity. Such AA distraction is possible only with simultaneous disruption of alar and apical ligaments. Tectorial membrane can be torn and a dural tear can be detected [37]. Although not ruptured, the TAL is always suspected of being also seriously damaged. Cases of dual AOD and AAD injury have been reported even less frequently (only three cases up to now) [14, 15, 22]. A congenital atlanto-occipital coalition, various CVJ anomalies and particularly hypoplastic odontoid

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_15, © Springer-Verlag Berlin Heidelberg 2011

215

216

15  Acute Traumatic Atlantoaxial Dislocation (AAD) in Adults

process are thought to be a predisposing factor for AA instability [35]. In these cases, only mild trauma can cause significant AAD. As mentioned previously, competence of TAL and integrity of odontoid process are crucial factors for maintaining atlantoaxial stability. However, the other ligaments, membranes, and joint capsules also play an important role especially in purely traumatic rotatory, uni- or bilateral AA joint dislocations. If reduced during the emergent investigations after injury, a purely ligamentous injury can be missed and the patient can come back later with chronic AA instability. Adult traumatic AAD represents a substantially different topic from atlantoaxial rotatory subluxation/ fixation commonly diagnosed in children (“cockrobin” posture) usually presenting with orofacial infections, minor trauma, ocular problems, and genetic diseases. Also, other pathologies can lead to AA displacement and if trauma is superimposed onto tumor, RA or infection, the diagnosed AA displacement is certainly also not a typical AAD.

15.2 Clinical Diagnosis

15.1 Etiology and Epidemiology AAD resulting from trauma occurs in only 1–2% of patients admitted to hospitals with acute cervical injuries [5, 12]. The mechanisms responsible for translational and rotatory AAD with or without simultaneous UCS fractures can be complex with some predilection for flexion in anterior dislocations and lateral bending and/or rotation in rotatory displacements. The probable mechanism of distractive AAD is the hyperextension of UCS with consequent rupture of alar, apical, and accessory ligaments, but the tectorial membrane and ALL must, in principle, be damaged also. Until this moment the mechanism of disruption is similar to AOD; however, instead of capsular AO disruption the AA joint capsule and its ligaments are crushed with subsequent distraction and/or dislocation. Exceptionally, both UCS joints can also be involved creating dual AAD/AOD injury. An actual injury mechanism reported in the majority of cases is the sudden hyperextension of relaxed spine not prepared to resist, with typical example of a relaxed pedestrian walking on the street hit by a car from behind.

Probably, half of trauma victims suffering from AAD die at the place of accident; however, the majority of survivors fortunately do not have major neurologic deficit. The symptoms vary from nonspecific pain with blocked UCS movement to fixed head rotation away from anteriorly displaced AA joint in rotational AAD. Some can also have suboccipital neuralgia due to an overstretched C2 nerve root. Exceptionally, the signs of VB insufficiency can be seen if the VA is compromised. As in other UCS injuries, approximately in 20% of admitted patients, the specific clinical picture can be clouded by coincident symptomatology of cerebral injury or polytrauma.

15.3 Radiology Plain films are seldom obtained as a first assessment. Lateral projections are often without signs of pathology (if fracture or posterior atlas dislocation are not present) and transoral pictures can show only asymmetric odontoid position. Dynamic films can reveal AA instability in cooperating patients but are never performed as the primary investigation. Currently, trauma patients pass through the emergency department with a helical CT performed primarily. The diagnosis is thus reached earlier than in the past. It is usually easy to visualize the fractures but the rotatory dislocation needs a specific protocol. The scanner gantry angle adopted to respect the axial plane of atlas and images superimposed by computer are necessary to quantify the amount of AA rotation [3]. Recently, the 3D CT can demonstrate the amount of displacement much better. To assess TAL, alar ligaments as well as the space available for spinal cord, the MRI should be done in all cases. In those with marked AA dislocation, VA can be overstretched and thus CTA can be indicated to elucidate its patency.

15.4 Treatment Strategy The goals of AAD treatment are to restore or prevent possible neurological compromise, to stabilize the dangerous instability and if possible in minor, purely

217

References

ligamentous injuries, to restore normal, pain-free motion of AA joint. In translational injuries with fractures, it is the fracture pattern that determines whether conservative or operative treatment should be undertaken [17, 34]. If TAL is damaged, the majority of surgeons perform posterior AA fusion [9, 24]. The controlled traction – reduction nearly always precedes the final surgical procedure. The pure rotatory dislocations can be treated primarily by manipulation (also finger through mouth pressure was recommended), by traction or their combinations with subsequent hard bracing for 6  weeks–3  months [23]. However, if the conservative approach fails, posterior AA fusion should be considered. In fracture-associated rotational dislocation, the stability of the fracture often is the most important point in the decision process. The indications were described in detail in previous chapters. For example, the odontoid type II fracture should be fixed by direct screw osteosynthesis and the AA instability often heals without further problems. In pure distractive injury with posterior atlas displacement, the reduction often consisting of closed distraction, reduction, and release, with [16, 18, 36, 37] or without [32] posterior fusion was reported as efficient treatment. In one case, the transoral odontoid resection was chosen to release the dislocation [11] and in another the high anterolateral approach and partial odontoid resection with C1 reduction followed by anterior AA screw fixation was performed [19]. During any reduction manipulation, one has to take care not to overdistract the AA joint. Some authors prefer closed maneuvers to reduce the AAD but if this is not easily possible the head manipulation controlled by direct “open” visibility of posterior elements seems to be safer. Also, the open reduction maneuvers can be very effective. Yoon et al. reported that caudally oriented pressure to C2 spinous process under simultaneous traction can be effective in AAD reduction [37]. Certainly, all such manipulations have to be controlled by lateral fluoroscopy and IOM.

15.5 Our Preference As humans are standing and walking beings, the spine head fixation system is not adapted to pure distraction and in any injury where the distractive force is

suspected (e.g., gliding of passenger under the seat belts with submandibular excoriations, violent combat sports, rugby, etc.), full investigation of potential distractive injury has to follow. Despite a very limited experience with this topic we suppose that the CT and/or MRI assessment of AA but also of AO joints should be done under mild distractive force either during the diagnostic process or later at follow-up checks to recognize the inadequate joint fissure distraction. It seems doubtful that the ligaments will heal with sufficient capacity to resist the head traction, especially when considering the age and lifestyle of each individual. As described previously, we are surgically active in cases of unstable fracture dislocations as well as in cases of confirmed TAL damage. In the case of pure rotational ligamentous injury without marked TAL damage, the question remains what extent of rotation exceeds the normal physiological range. Reduction by traction and collar fixation is probably a good solution for majority of these.

References   1. Alker Jr., G.J., Oh, Y.S., Leslie, E.V.: High cervical spine and craniocervical junction injuries in fatal traffic accidents: a radiological study. Orthop Clin North Am 9, 1003–1010 (1978)   2. Alker, G.J., Oh, Y.S., Leslie, E.V., et al.: Postmortem radiology of head neck injuries in fatal traffic accidents. Radiology 114, 611–617 (1975)   3. Bono, C.M., Vaccaro, A.R., Fehlings, M., et al.: Measurement techniques for upper cervical spine injuries: consensus statement of the Spine Trauma Study Group. Spine (Phila Pa 1976) 32, 593–600 (2007)   4. Boos, N., Khazim, R., Kerslake, R.W., et al.: Atlanto-axial dislocation without fracture: case report of an ejection injury. J Bone Joint Surg Br 79, 204–205 (1997)   5. Carroll, E.A., Gordon, B., Sweeney, C.A., et al.: Traumatic atlantoaxial distraction injury: a case report. Spine (Phila Pa 1976) 26, 454–457 (2001)   6. Castel, E., Benazet, J.P., Samaha, C., et al.: Delayed closed reduction of rotatory atlantoaxial dislocation in an adult. Eur Spine J 10, 449–453 (2001)   7. Chaudhary, R., Chaudhary, K., Metkar, U., et al.: Posterior atlantoaxial dislocation without odontoid fracture. Skeletal Radiol 37, 361–366 (2008)   8. Cheng, S.G., Blackmore, C.C., Mirza, S.K., et al.: Rotatory subluxation and fracture at C1-C2. AJR Am J Roentgenol 175, 540 (2000)   9. Dickman, C.A., Mamourian, A., Sonntag, V.K., et  al.: Magnetic resonance imaging of the transverse atlantal ligament for the evaluation of atlantoaxial instability. J Neurosurg 75, 221–227 (1991)

218

15  Acute Traumatic Atlantoaxial Dislocation (AAD) in Adults

10. Fielding, J.W., Hawkins, R.J.: Atlanto-axial rotatory fixation. (Fixed rotatory subluxation of the atlanto-axial joint). J Bone Joint Surg Am 59, 37–44 (1977) 11. Fox, J.L., Jerez, A.: An unusual atlanto-axial dislocation. Case report. J Neurosurg 47, 115–118 (1977) 12. Freeman, B.J., Bisbinas, I., Nelson, I.W.: Traumatic atlantoaxial subluxation and missed cervical spine injuries. Hosp Med 59, 330–331 (1998) 13. Fuentes, S., Bouillot, P., Palombi, O., et al.: Traumatic atlantoaxial rotatory dislocation with odontoid fracture: case report and review. Spine (Phila Pa 1976) 26, 830–834 (2001) 14. Gonzalez, L.F., Klopfenstein, J.D., Crawford, N.R., et  al.: Use of dual transarticular screws to fixate simultaneous occipitoatlantal and atlantoaxial dislocations. J Neurosurg Spine 3, 318–323 (2005) 15. Hamai, S., Harimaya, K., Maeda, T., et  al.: Traumatic atlanto-occipital dislocation with atlantoaxial subluxation. Spine (Phila Pa 1976) 31, E421–E424 (2006) 16. Haralson 3rd, R.H., Boyd, H.B.: Posterior dislocation of the atlas on the axis without fracture. Report of a case. J Bone Joint Surg Am 51, 561–566 (1969) 17. Hopf, S., Buchalla, R., Elhoft, H., et al.: Atypical dislocated dens fracture type II with rotational atlantoaxial luxation after a riding accident. Unfallchirurg 112, 517–520 (2009) 18. Jamshidi, S., Dennis, M.W., Azzam, C., et  al.: Traumatic posterior atlantoaxial dislocation without neurological deficit: case report. Neurosurgery 12, 211–213 (1983) 19. Jiang, L.S., Shen, L., Wang, W., et al.: Posterior atlantoaxial dislocation without fracture and neurologic deficit: a case report and the review of literature. Eur Spine J 28, 28 (2009) 20. Jones, R.N.: Rotatory dislocation of both atlanto-axial joints. J Bone Joint Surg Br 66, 6–7 (1984) 21. Kim, Y.S., Lee, J.K., Moon, S.J., et al.: Post-traumatic atlantoaxial rotatory fixation in an adult: a case report. Spine (Phila Pa 1976) 32, E682–E687 (2007) 22. Kleweno, C.P., Zampini, J.M., White, A.P., et al.: Survival after concurrent traumatic dislocation of the atlanto-occipital and atlanto-axial joints: a case report and review of the literature. Spine (Phila Pa 1976) 33, E659–E662 (2008) 23. Levine, A.M., Edwards, C.C.: Treatment of injuries in the C1-C2 complex. Orthop Clin North Am 17, 31–44 (1986) 24. Miyamoto, H., Doita, M., Nishida, K., et al.: Traumatic anterior atlantoaxial subluxation occurring in a professional rugby athlete: case report and review of literature related to

atlantoaxial injuries in sports activities. Spine (Phila Pa 1976) 29, E61–E64 (2004) 25. Monckeberg, J.E., Tome, C.V., Matias, A., et  al.: CT scan study of atlantoaxial rotatory mobility in asymptomatic adult subjects: a basis for better understanding C1-C2 rotatory fixation and subluxation. Spine (Phila Pa 1976) 34, 1292– 1295 (2009) 26. Moore, K.R., Frank, E.H.: Traumatic atlantoaxial rotatory subluxation and dislocation. Spine (Phila Pa 1976) 20, 1928–1930 (20) 27. Ono, K., Yonenobu, K., Fuji, T., et al.: Atlantoaxial rotatory fixation. Radiographic study of its mechanism. Spine (Phila Pa 1976) 10, 602–608 (1985) 28. Patzakis, M.J., Knopf, A., Elfering, M., et al.: Posterior dislocation of the atlas on the axis: a case report. J Bone Joint Surg Am 56, 1260–1262 (1974) 29. Payer, M., Wetzel, S., Kelekis, A., et al.: Traumatic vertical atlantoaxial dislocation. J Clin Neurosci 12, 704–706 (2005) 30. Przybylski, G.J., Welch, W.C.: Longitudinal atlantoaxial dislocation with type III odontoid fracture. Case report and review of the literature. J Neurosurg 84, 666–670 (1996) 31. Roche, C.J., O’Malley, M., Dorgan, J.C., et al.: A pictorial review of atlanto-axial rotatory fixation: key points for the radiologist. Clin Radiol 56, 947–958 (2001) 32. Sassard, W.R., Heinig, C.F., Pitts, W.R.: Posterior atlantoaxial dislocation without fracture. Case report with successful conservative treatment. J Bone Joint Surg Am 56, 625–628 (1974) 33. Sinigaglia, R., Bundy, A., Monterumici, D.A.: Traumatic atlantoaxial rotatory dislocation in adults. Chir Narzadow Ruchu Ortop Pol 73, 149–154 (2008) 34. Spoor, A.B., Diekerhof, C.H., Bonnet, M., et al.: Traumatic complex dislocation of the atlanto-axial joint with odontoid and C2 superior articular facet fracture. Spine (Phila Pa 1976) 33, E708–E711 (2008) 35. Weiner, B.K., Brower, R.S.: Traumatic vertical atlantoaxial instability in a case of atlanto-occipital coalition. Spine (Phila Pa 1976) 22, 1033–1035 (1997) 36. Wong, D.A., Mack, R.P., Craigmile, T.K.: Traumatic atlantoaxial dislocation without fracture of the odontoid. Spine (Phila Pa 1976) 16, 587–589 (1991) 37. Yoon, D.H., Yang, K.H., Kim, K.N., et al.: Posterior atlantoaxial dislocation without fracture. Case report. J Neurosurg 98, 73–76 (2003)

Posttraumatic Deformity

16

P. Suchomel and R. Fricˇ

Posttraumatic spinal deformity has always been a problem inevitably related to spinal trauma, although relatively small number of spine fractures, with the majority located in the thoracolumbar region, have been described in paleopathological studies [9]. Weber et  al. [26] reported observation of a pseudarthrosis after Anderson-Alonzo type III odontoid fracture in skeletal remnants of a medieval man. Posttraumatic deformity may be encountered in all parts of upper cervical spine (UCS) where fractures and dislocations typically occur, although pseudoarthrosis of the odontoid has been reported most frequently [1, 3, 15, 20, 22, 24, 25]. The deformity of UCS results in sagittal and/or frontal dysbalance and can easily become a significant pain generator due to excessive mobility of UCS. Moreover, healing failure of important stabilizing elements such as the odontoid may represent a critical and potentially life-threatening instability. Patients presenting with intractable pain related to posttraumatic UCS deformity or instability usually have to be treated surgically, although old patients and/or those with high risk of medical complications may be acceptable exceptions. Decompression of neural structures, sagittal and frontal spinal realignment, and stabilization are the primary goals of surgery.

P. Suchomel Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic R. Fricˇ Department of Neurosurgery, Rikshospitalet, Oslo University Hospital, Sognsvannsveien 20, 0027 Oslo, Norway

16.1 Etiology Generally, the UCS deformity is a result of healing failure after trauma. This can be caused by a missed diagnosis, inappropriate treatment, and/or failure of an otherwise appropriate treatment. Some types of fractures can easily be overlooked on initial plain films, in particular those with  hairline appearance and without dislocation (Figs. 11.3 and 11.9, Chap. 11) [4]. Deformity and instability following conservative treatment are often related to insufficient or failed external immobilization (Fig.  11.11, Chap. 11; Fig. 10.9, Chap. 10). If fracture healing is not confirmed by radiological studies and the brace is withdrawn too early, a ­deformity may develop consequently. Another cause might be an inappropriate choice of conservative treatment in cases obviously requiring surgery due to present instability (Fig. 16.1). On the other hand, falsely indicated or poorly performed surgical intervention (Fig. 16.2) may lead to deformity of UCS as well. Even an adequately indicated and preformed treatment of UCS trauma can fail, and posttraumatic spinal deformity or instability may develop anyway [7].

16.2 Clinical Symptoms The clinical picture of posttraumatic UCS deformity is not specific and does not differ from other causes of UCS instability. Patients can be completely asymptomatic. However, the majority of them present with pain dependent on head rotation, occipital pain, neck stiffness, and reduced mobility of the neck. The symptoms relates to the type of UCS injury and the treatment performed. Though not often, myelopathy signs due to prolonged spinal cord compression and/or

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_16, © Springer-Verlag Berlin Heidelberg 2011

219

220 Fig. 16.1  14-year-old boy with Marphan’s syndrome referred to our hospital from abroad after 3 months of unsuccessful treatment with halo vest, presenting with progressive quadrisymptomatology. Unclear history of spine injury 1 year ago. (a) C2-3 kyphotic deformity on plain radiogram. (b) Sagittal MRI scan showing spinal cord compression. (c, d) Pressure sores caused by wearing the halo vest, note the general exhaustion appearance

Fig. 16.2  Inadequately performed Goel-Harms fixation of type II odontoid fracture with posterior AA subluxation and posterior angulation of the odontoid process with spinal cord compression. (a) Midsagittal CT scan. (b) Sagittal MRI scan in T2-weighted images. (c) CT reconstruction in the plane of subluxed AA joint. (d) Postoperative CT after satisfactory transoral decompression

16  Posttraumatic Deformity

a

b

c

d

a

b

c

d

221

16.4  Treatment Strategy Fig. 16.3  AA instability as shown on dynamic lateral radiograms in a case of odontoid pseudarthrosis following treatment of type II fracture with external brace. (a) Extension. (b) Flexion

a

b

a

b

Fig. 16.4  Imaging of the odontoid process pseudarthrosis in two different patients. (a) Coronal CT reconstruction. (b) Midsagittal CT scan in 3D format

vascular compromise may develop even several years after the injury [18, 19].

16.3 Radiology Plain films usually reveal significant deformity or malunion, and dynamic radiographs will show its potential instability (Fig.  16.3). Nonetheless, a CT scan with adequate reconstruction is necessary if the distortion of UCS anatomy shall be clearly depicted (Fig. 16.4). MRI can show the capacity of spinal canal and extent of neural compression (Fig. 16.5).

16.4 Treatment Strategy Conservative treatment with external immobilization, activity restriction, and analgesics can be justified only in mild and stable deformities without any neurological

Fig. 16.5  MRI depicting the spinal cord compression consequent with odontoid pseudarthrosis, the same patient as on Fig. 16.4b

222

symptoms, particularly in patients with sufficient pain relief. Conservative treatment is also an option in elderly patients and/or those with significant medical comorbidities that unacceptably increases the risk of surgical treatment. Nevertheless, the vast majority of patients with symptomatic posttraumatic deformity/ instability are treated surgically. As in case of any other UCS pathology, the decompression of neural structures is the primary goal of surgical treatment, although correction of the deformity and adequate spinal reconstruction with concomitant fixation are the inherent parts of the posttraumatic deformity surgery.

16.5 Odontoid Pseudarthrosis As odontoid pseudoarthrosis appears to be the most frequent reason to surgical intervention among posttraumatic deformities in our department, we wish to focus on this topic in detail. The rate of odontoid pseudarthrosis, as reported in literature, varies between 1 and 64% and depends on the type of fracture and the treatment modality [12]. Failure in bone healing after type II odontoid fracture, according to Anderson and D’Alonso, is the most frequent cause. There are similar reasons to development of pseudoarthrosis as in other treatment failures of UCS injuries. Fractures initially overlooked and those not treated at all are still common [4] and can lead to non-union in up to 100% of cases [2]. Type II odontoid fracture is highly unstable and the treatment with external fixation fail in 30–50% of cases [8, 10, 13, 14, 23]. If patients older than 60 years are included in the series, the failure rate increases dramatically up to 77–86% [10, 21]. Conversely, in fractures adequately treated with anterior compressive osteosynthesis, the malunion occurs only in up to 15% [7]. Diagnosis of odontoid pseudarthrosis is usually based on the history of injury, clinical symptoms, and radiological findings. Cervical spine radiographs in lateral, AP, and Sandberg (transoral) projections along with CT scan with bone windows and 3D reconstructions confirm the diagnosis. Stability of the odontoid process can best be assessed on flexion-extension views (Figs. 16.3 and 16.8). Capacity of the spinal canal can be evaluated by MRI, particularly in cases where hypertrophic callus due to pseudoarthrosis is present. Dynamic MRI can be of value in patients where conservative treatment is considered.

16  Posttraumatic Deformity

Exceptionally, the diagnosis of pseudoarthrosis may be difficult as demonstrated in the case of a 55-year-old man reported by Rudzki et al. [19]. After a long asymptomatic period, signs of cervical myelopathy had developed and pseudoarthrosis following type II fracture was diagnosed 39 years after the original injury. Blauth et al. suggested a classification of odontoid pseudarthrosis based on their extensive experience with spine injuries [1]. In type I, called “fixed pseudarthrosis,” the fracture line is not bridged by osseous fusion but there is no dislocation on flexion-extension radiographs. The patients are usually asymptomatic and can be followed up with serial imaging. In type II, grossly dislocated but “stable pseudoarthrosis,” the proximal fragment is usually ventrally dislocated along with the atlas. Dynamic films usually confirm this finding. Reduction is not possible. As the patient can develop symptoms and risk of neurological deterioration is real, the surgical treatment is indicated. In type III, referred to as “unstable pseudoarthrosis,” dynamic radiographs show marked displacement. The patient can be symptomatic and surgery is always indicated. Type IV, “posttraumatic os odontoideum,” is distinguished by a high degree of instability. It may also be found as an incidental finding. Indication for surgery depends on clinical symptoms, general health, and age of the patient. As pseudoarthrosis of the odontoid process can become a cause of chronic myelopathy and/or acute spinal cord injury, the majority of authors recommend performing AA stabilization even in asymptomatic patients [2, 3, 5, 16]. Nonetheless, some authors advocate the conservative approach, particularly in patient with high risk of complications owing to age, immobilization, or general health [11, 17]. In case of surgical AA stabilization, the posterior C1-2 fixation is usually performed [1, 3, 6, 18]. Use of transoral decompression in cases of hypertrophic fibrous malunion was also reported [3, 5]. In order to preserve AA motion, Ruf et  al. recommended a transorally performed debridement of pseudarthrosis, with cancellous bone grafting accompanied by simultaneous anterior or posterior temporary screw AA fixation for 3–4 months [20].

16.6 Our Preference Depending on the type of deformity and ­instability, surgery is indicated for those patients who can benefit from decompression, reconstruction, and

223

16.6  Our Preference

stabilization. In asymptomatic patients, an estimation of potential risk related to the type of instability is crucial for surgical indication. Neurologically compromised individuals with apparent symptoms of spinal cord compression, but also those suffering from continuous intractable pain as a result of malalignment, should definitely be considered for surgical

Fig. 16.6  The same patient as shown on Fig. 16.1. Simple manual traction allowed lordotization of the deformity from 41° to 20°, thus giving evidence of incompetence of the halo-vest. (a) Lateral fluoroscopy. (b) Manual traction the angle changed from

release and reconstruction. As an example, deformity consequent with conservative treatment of hangman’s fracture causing only temporary pain may be treated conservatively (Fig.  12.11, Chap. 12); however, a similar situation causing neurological deficit represents an absolute indication for surgery in our opinion (Figs. 16.6 and 16.7).

b

a

a

b

c

d

Fig. 16.7  The same patient as on Figs. 16.1 and 16.6. CT scan in 3D reconstruction revealed old C2/3 luxation on the right side which together with elongated pars interarticularis led to diagnosis of type III inveterate hangman’s fracture. Because of general health condition of the patient but also due to partial reducibility of the kyphosis, only a single stage C3 somatectomy

with graft and plate fixation was performed. (a) 3D CT scan in sagittal plane demonstrating C2/3 facet dislocation on the rightside. (b) Postoperative plain laterogram. (c) CT scan of anterior graft and plate fusion. Note bicortical screw insertion. (d) Patient walking with hard collar fixation a week later

224

16  Posttraumatic Deformity

As the actual risk of progressive neurological symptoms in case of odontoid pseudarthrosis is not known and cases of significant clinical deterioration have been reported, we favor an active surgical approach in the majority of cases. We definitely operate on patients with AA instability and those with neurological symptoms related to compression caused by deformity (Fig. 16.8). We prefer stabilization also in patients with stable pseudoarthrosis developed after conservative treatment, particularly in active and young individuals (Fig. 16.9), while conservative approach with careful radiological follow-up can be an alternative when age and/or medical risks do not allow the surgical treatment. In cases where odontoid screw compressive osteosynthesis fails

Fig. 16.8  A case of unstable non-union of the odontoid treated with posterior transarticular screw and graft fusion. (a) Classical tomogram of a ­pseudarthrosis. Dynamic radiograph in flexion (b) and extension (c). (d) Result of fusion one year after surgery

to create continuous bone bridge across the fracture line as confirmed by CT, we also choose an active approach, particularly if the fracture gap is documented (Fig. 6.38, Chap. 6). The hardware failure may have catastrophic consequences. Occasionally, we see patients with inadequately performed odontoid screw fixation where the risk of hardware failure is obvious (Fig. 16.10). We usually perform posterior AA stabilization according to techniques described by Magerl or GoelHarms, supplemented by an interlaminar graft. In rare cases of hypertrophic pseudoarthrosis directly causing anterior compression of neural structures, we perform transoral odontoidectomy with removal of fibrous tissue, followed by posterior AA stabilization (Fig. 16.11).

a

b

c

d

225

16.6  Our Preference Fig. 16.9  Stable odontoid pseudarthrosis in a 40 years old active sportsman, treated with posterior transarticular screw and graft fusion. (a) Sagittal CT reconstruction. (b) Posterior AA fusion

Fig. 16.10  Failure of inadequately performed anterior double-screw osteosynthesis of type II odontoid fracture. Lag screw is not compressing the fracture and the antirotational screw is too short, not passing through but probably distracting the fracture. This instability was treated with posterior fusion according to Magerl with interlaminar autologous graft. (a) Transoral projection. (b) Laterogram

a

a

b

a

b

b

Fig. 16.11  Irreducible odontoid malunion creating a deformity with spinal cord compression; treated with transoral decompression, cage support, and posterior fusion in a single stage surgery. (a) Sagittal CT scan showing the bone deformation. (b) Spinal

c

cord compression as shown on MRI. (c) Anterior Harms cage fixed to C3 caudally and with a fork notch to clivus, posterior O-C3 screw fusion

226

References   1. Blauth, M., Richter, M., Kiesewetter, B., et  al.: Operative versus non operative treatment of odontoid non unions. How dangerous is it not to stabilize a non union of the dens? Chirurg 70, 1225–1238 (1999)   2. Clark, C.R., White 3rd, A.A.: Fractures of the dens. A multicenter study. J Bone Joint Surg Am 67, 1340–1348 (1985)   3. Crockard, H.A., Heilman, A.E., Stevens, J.M.: Progressive myelopathy secondary to odontoid fractures: clinical, radiological, and surgical features. J Neurosurg 78, 579–586 (1993)   4. Cusmano, F., Ferrozzi, F., Uccelli, M., et al.: Upper cervical spine fracture: sources of misdiagnosis. Radiol Med 98, 230–235 (1999)   5. Fairholm, D., Lee, S.T., Lui, T.N.: Fractured odontoid: the management of delayed neurological symptoms. Neurosurgery 38, 38–43 (1996)   6. Finn, M.A., Apfelbaum, R.I.: Atlantoaxial transarticular screw fixation: update on technique and outcomes in 269 patients. Neurosurgery 66, A184–A192 (2010)   7. Fountas, K.N., Kapsalaki, E.Z., Karampelas, I., et al.: Results of long-term follow-up in patients undergoing anterior screw fixation for type II and rostral type III odontoid fractures. Spine (Phila Pa 1976) 30, 661–669 (2005)   8. Fujii, E., Kobayashi, K., Hirabayashi, K.: Treatment in fractures of the odontoid process. Spine (Phila Pa 1976) 13, 604–609 (1988)   9. Gerszten, P.C., Gerszten, E., Allison, M.J.: Diseases of the spine in South American mummies. Neurosurgery 48, 208– 213 (2001) 10. Greene, K.A., Dickman, C.A., Marciano, F.F.: Acute axis fractures. Analysis of management and outcome in 340 consecutive cases. Spine (Phila Pa 1976) 22, 1843–1852 (1997) 11. Hart, R., Saterbak, A., Rapp, T., et al.: Nonoperative management of dens fracture nonunion in elderly patients without myelopathy. Spine (Phila Pa 1976) 25, 1339–1343 (2000) 12. Knoller, S., Jeszenszky, D., Willms, R., et  al.: Transaxial spongiosa-plasty and ventral, temporary atlanto-axial fixation for therapy of dens pseudarthrosis. Z Orthop Ihre Grenzgeb 137, 232–235 (1999) 13. Koivikko, M.P., Kiuru, M.J., Koskinen, S.K., et al.: Factors associated with nonunion in conservatively-treated type-II

16  Posttraumatic Deformity fractures of the odontoid process. J Bone Joint Surg Br 86, 1146–1151 (2004) 14. Lennarson, P.J., Mostafavi, H., Traynelis, V.C., et  al.: Management of type II dens fractures: a case-control study. Spine (Phila Pa 1976) 25, 1234–1237 (2000) 15. Muller, E.J., Wick, M., Russe, O., et al.: Accident-induced pseudarthroses of the dens axis. Etiology, follow-up and therapy. Unfallchirurg 101, 750–754 (1998) 16. Paradis, G.R., Janes, J.M.: Posttraumatic atlantoaxial instability: the fate of the odontoid process fracture in 46 cases. J Trauma 13, 359–367 (1973) 17. Pepin, J.W., Bourne, R.B., Hawkins, R.J.: Odontoid fractures, with special reference to the elderly patient. Clin Orthop Relat Res 193, 178–183 (1985) 18. Platzer, P., Vecsei, V., Thalhammer, G., et  al.: Posterior atlanto-axial arthrodesis for fixation of odontoid nonunions. Spine (Phila Pa 1976) 33, 624–630 (2008) 19. Rudzki, J.R., Lenke, L.G., Blanke, K., et al.: Pseudarthrosis of a thirty-nine-year-old dens fracture causing myelopathy. A case report. J Bone Joint Surg Am 86-A, 2509–2513 (2004) 20. Ruf, M., Welk, T., Muller, M., et al.: Ventral cancellous bone augmentation of the dens and temporary instrumentation C1/ C2 as a function-preserving option in the treatment of dens pseudarthrosis. J Spinal Disord Tech 23, 285–292 (2010) 21. Ryan, M.D., Taylor, T.K.: Odontoid fractures in the elderly. J Spinal Disord 6, 397–401 (1993) 22. Schwarz, N., Bauer, J.: Post-traumatic os odontoideum. Unfallchirurg 98, 483–486 (1995) 23. Seybold, E.A., Bayley, J.C.: Functional outcome of surgically and conservatively managed dens fractures. Spine (Phila Pa 1976) 23, 1837–1845 (1998). discussion 1845–1836 24. Suchomel, P., Stulik, J., Klezl, Z., et al.: Transarticular fixation of C1-C2: a multicenter retrospective study. Acta Chir Orthop Traumatol Cech 71, 6–12 (2004) 25. Wang, G.J., Mabie, K.N., Whitehill, R., et al.: The nonsurgical management of odontoid fractures in adults. Spine (Phila Pa 1976) 9, 229–230 (1984) 26. Weber, J., Vieweg, U., Dollhopf, K.D., et al.: Type III odontoid fracture with pseudarthrosis in a skeleton from the early Middle Ages. Acta Neurochir (Wien) 146, 1379–1381 (2004). discussion 1381

Non Specific Inflammation

17

P. Suchomel and O. Choutka

Pyogenic infections of the UCS are very rare; nevertheless, they can have a major impact on the health of the patient and therefore, we feel that it is important to share our experience with the reader. The purulent destruction of bone and ligaments can lead to lifethreatening AA instability [1]. Whereas the rate of specific spine inflammation is growing in the European countries as a result of increased migration from developing countries, the growing rate of non-specific, infective AA osteomyelitis stems from population aging and overall decreased immunocompetency. This is mainly due to a growing number of immunodeficient people, uncontrolled use of broad-spectrum antibiotics, and an increasing number of people suffering from diseases of our civilization (atherosclerosis, diabetes). Further, the substantial development in diagnostic technology enables us to detect a greater number of pathologies. Last but not the least, a significant amount of people worldwide suffer from HIV and/or are drug users, heavy smokers, and alcoholics [20, 21, 27]. Despite an increased frequency of UCS tuberculosis detected not only in Asia but currently also in those countries where a lot of immigrants have settled (UK), we have not seen such a case so far. Nevertheless, this

P. Suchomel Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic O. Choutka University of Cincinnati, Medical Center, Department of Neurosurgery, Albert Sabin Way 231, Cincinnati, OH 45267-0515, USA

possibility must be entertained in every differential diagnosis of an UCS pyogenic process. During the last 15  years, we have encountered an increasing amount of patients presenting with pyogenic spondylitis. Five of them suffered from a dangerous infection involving the UCS and requiring surgical intervention.

17.1 Incidence When considering the entire skeleton, cervical spine is  affected by pyogenic osteomyelitis relatively ­infrequently (3–6%) [6, 7, 15]. The first report of atlantoaxial osteomyelitis in three patients is credited to Malkins and Abbott, in 1896 [16]. All of the reported patients died because of unavailability of antibiotics at that time. In the modern times, only individual cases [13, 14, 27] or very small series [24, 29] of the UCS pyogenic inflammation were reported. In 1994, Gormley and Rock [8] reviewed 17 case reports published previously and revealed that majority of the reports had not been older than 10 years and thus concluded that UCS osteomyelitis is a growing contemporary problem. Most frequently, AA osteomyelitis occurs as a result of previous orofacial infection (primary or ­secondary e.g.: after tonsillectomy, dental surgery). The infection reaches the UCS bones either directly or  via venous drainage, although hematogenous ­dissemination is also possible. Pathogens detected in UCS conform to those causing osteomyelitis in other parts of axial skeleton. Staphylococcus aureus was confirmed in the majority of reported cases followed by Pseudomonas aeruginosa, Escherichia coli, and Proteus mirabilis.

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_17, © Springer-Verlag Berlin Heidelberg 2011

227

228

17  Non Specific Inflammation

17.2 Clinical Symptoms and Diagnosis Similar to other infective processes, patients can present with general symptoms of fever and fatigue as well as symptoms and signs specific to UCS osteomyelitis. Those may include mechanical neck pain worsening especially in rotation, neck stiffness or swallowing difficulties, and enlargement of painful cervical lymph nodes [8, 12, 24]. With neural compromise, patients may suffer from cranial nerve palsies and/or signs of cord compression. Myelopathy may range from a mild quadriparesis to a severe bedridden status and sometimes, finally, death of the patient [1]. Serum inflammatory markers are usually elevated, with possible increases in ESR, CRP, and leukocytosis a

b

Fig. 17.1  Plain lateral film showing consecutive development of UCS inflammatory process with AA subluxation after purulent pharyngitis (S. aureus). (a) Plain film obtained at the time of first neck pain. Odontoid is not well outlined and retropharyngeal

a

b

Fig.  17.2  Plain lateral radiograph depicting development of osseous C2 destruction and atlas settling within 5 months after inflammatory process in the parotid gland (S. aureus). (a) Picture obtained at the time of the first neck pain, only mild widening of

with shift in differential rate to immature cells. Blood cultures should be obtained on admission; however, treatment usually cannot be delayed until their results are available. The physician should never forget to enquire about history of diseases that compromise the immune system such as diabetes mellitus (DM), AIDS, chronic steroid use, or immunomodulants in allergies and/or about history of heavy smoking or other drug abuse.

17.3 Radiology Plain radiographs will typically reveal AA subluxation (Fig. 17.1) or osseous destruction if present (Fig. 17.2). However, the most common finding is usually only an c

space widened. (b) Lateral film obtained 3 weeks after showing marked AA dislocation. (c) Lateral fluoroscopical view at the time of admittance to our hospital 4 weeks after the first films. Note also development of significant kyphosis

c

retropharyngeal space can be detected. (b) Lost contour of odontoid process visible on lateral view 2 month later. (c) Significant C2 destruction with AA dislocation and settling 5 months after complaint onset

229

17.3  Radiology

increase in the prevertebral space and edema of retropharyngeal tissues [8, 23]. Dynamic plain views are absolutely necessary at the earliest convenience to exclude a potential AA instability in primarily nondislocated cases. Similar picture but with greater detail of bony morphology can be obtained from CT imaging (Fig. 17.3). Contrasted CT can also demonstrate pyogenic membrane of an abscess, if already present [8, 10, 23]. The “hot spot” and also other disease foci can

Fig. 17.3  CT of purulent destruction of odontoid process and left C1 lateral mass. (a) Axial scan. (b) Coronal plane reconstructions

Fig. 17.4  MRI in T2 sequence showing the C2 destruction with pus spreading retropharyngeally

a

be detected on bone scintigraphy in the early phases of infection, although it is neither very specific nor ­anatomically sufficient [9, 19]. The ­mainstay of diagnostic assessment is ­contrasted MRI (Fig. 17.4), possibly supplemented with dynamic positions (Fig. 17.5). MRI is superior in demonstrating liquid abscess (Fig.  17.6), in exclusion of spinal cord compression (Fig. 17.7) and in determination of the extent of prevertebral tissue ­involvement [8, 17, 24].

b

230 Fig. 17.5  Dynamic MRI without increased neural compression in position change. (a) Flexion. (b) Extension

17  Non Specific Inflammation

a

b

Fig. 17.6  Liquid epidural pus replacing odontoid process visible on axial MRI scan

a

Fig. 17.7  Destruction of odontoid process with inflammatory tissue expanding posteriorly and causing spinal cord compression (patient was quadriparetic). (a) Midsagittal MRI in T2 sequence depicting ­inflammatory retroodontoid peg formation. (b) CT showing the odontoid posterior destruction

b

231

17.6  Our Preference

17.4 Differential Diagnosis Occasionally, the only radiographic finding will be a non-specific mass destroying the UCS region without any spread so typical for purulent inflammation. The  differential diagnosis then includes tuberculosis, fungal process, and primary or secondary neoplasm – ­processes where the necrotic center can mimic liquid pus. In those situations, excisional or needle biopsy [18] can help determine the diagnosis and identify the infectious agent. In the vast majority of cases, the odontoid area is affected; however, an infection involving purely the atlas has also been reported [26].

17.5 Treatment Strategy There is a considerable risk to a delayed diagnosis with unrecognized development of dangerous AA instability (Figs.  17.1 and 17.2). Gormley and Rock’s retrospective review of available publications found that the late visibility of osseous destruction on plain films was the main reason for a delay in diagnosis [8]. They recommended an assessment of possible instability as soon as UCS involvement is suspected. Early identification of infectious agent is very important. Needle aspiration biopsy is preferred by many [8, 23] as a first diagnostic step after radiographic evaluation. It serves not only as a diagnostic tool in terms of etiology of the process but also allows for samples to be obtained for determination of antibiotic (ATB) sensitivities. Most authors recommend starting with external halo-vest immobilization in patients without neurological deficit and/or compression, and long-term antibiotic administration [8, 12, 17]. In those cases, surgical intervention is reserved for patients who fail conservative treatment with external immobilization. If instability or malalignment persist, posterior AA fusion is performed. On the other hand, transoral decompression followed by posterior fusion is recommended in patients with direct anterior compression and/or epidural pus spread [8, 14, 28]. Posterior fixation after closed traction reduction in case of a complete inflammatory destruction of the dens was also reported [4]. Some authors prefer an aggressive surgical approach upfront. They argue that conservative therapy results are questionable, especially when it comes to the eventual AA instability at the end of therapy [29].

Long-term antibiotics are always administered in cases of osseous involvement. Once cultures and sensitivities are available, an intravenous, ­microbe-specific ATB regime is instituted for 6–12 weeks followed by  long-term oral suppressive therapy for at least 6 months [24].

17.6 Our Preference In our series of patients with infections of UCS, four out of five patients were initially treated for pharyngitis without any radiographic evaluation. They were referred to our hospital late with significant UCS osseous destruction and/or instability. This was the reason for early surgical intervention in all of them. We choose conservative approach only in cases where there is no MRI-detectable pus. We respect the old surgical rule “ubi pus ibi evacua.” The decision whether soft collar or hard external bracing is required should be based on the presence or absence of instability on dynamic films. We consider Philadelphia or SOMI brace to be hard external braces and do not recommend halo-vest immobilization because of reasons explained in Chap. 11. If either encapsulated (abscess) or spreading (retropharyngeal or epidural empyema) pus is detected on MRI and/or CT, it should be evacuated via the shortest and safest route, in our opinion. Direct needle biopsy with aspiration of purulent material for microscopy, culture and sensitivity should be done as soon as infection is suspected by MRI. The “old fear” of possible tuberculous fistula formation is unfounded nowadays as is documented by many authors [2, 22, 25]. There are two approaches for UCS needle aspiration: fluoroscopically-guided transpharyngeal route or lateral CT-guided needle aspiration. In the absence of UCS instability or neural compromise, surgical evacuation of liquid pus collection followed by external immobilization, long-term ATB administration, and careful radiological follow-up can be sufficient. However, if significant bone inflammatory involvement (spondylitis) exists, we prefer to widely debride any affected tissue (i.e., may involve odontoidectomy) and insert antibiotics locally (gentamycin foam, spheres, etc.). In widespread bone tissue disintegration, we set up a drainage–lavage system. Most importantly, spinal cord has to be decompressed. As the majority of these procedures are performed

232

transorally (Fig. 17.8) and transoral fixation possibilities are rather limited, we usually prefer to perform posterior CVJ fixation restricted to affected motion segment during the same procedure (Fig. 17.9). Pyogenic inflammation in UCS region is very dangerous and can culminate in patient’s death; however,

Fig. 17.8  Intraoperative picture of pus coming out immediately after longitudinal posterior pharyngeal wall incision

Fig. 17.9  Examples of posterior fixation performed in single session after transoral decompression. (a) Transarticular screw fixation according to Magerl. Note the metal clip fixed to the remnant of C1 anterior arch (anterior tubercle resected) as a guiding point for Magerl screws. (b) Occipitocervical fusion in the case of AO and AA joint purulent destruction

17  Non Specific Inflammation

if recognized early and treated properly, it can have a relatively benign course with a very good outcome.

17.7 Remarks on Tuberculosis in UCS Region Spinal tuberculosis (TB) causing paraplegia was first described by Percival Pott in 1779. Later, the old surgeons called it “Pott’s disease” or “caries in spine”. Victor Horsley was probably the first person to perform cervical laminectomy for relief of tuberculous compression caused by “pachymeningitis cervicalis” in 1893 [11]. It is not clear who started to treat TB surgically in UCS but Berchtold Hadra was probably one of the first with his silver wire loops in the beginning of the twentieth century [11]. Nowadays, the spine is affected in approximately 1% of all TB patients [2], with UCS involved in 0.3–1.0% of them [3, 5]. The incidence of newly diagnosed TB in UCS is growing in European countries today. Although not the primary subject of this chapter, it needs to be stressed that TB may be finding its way back to the differential diagnosis of UCS infections with increasing frequency. There is a large body of contemporary experience gained from published series of patients treated in Asia [25].

References

References   1. Ahlback, S., Collert, S.: Destruction of the odontoid process due to atlanto-axial pyogenic spondylitis. Acta Radiol Diagn (Stockh) 10, 394–400 (1970)   2. Behari, S., Nayak, S.R., Bhargava, V.: Craniocervical tuberculosis: protocol of surgical management. Neurosurgery 52, 72–80 (2003). discussion 80-71   3. Bhojraj, S.Y., Shetty, N., Shah, P.J.: Tuberculosis of the craniocervical junction. J Bone Joint Surg Br 83, 222–225 (2001)   4. Busche, M., Bastian, L., Riedemann, N.C.: Complete osteolysis of the dens with atlantoaxial luxation caused by infection with Staphylococcus aureus: a case report and review of the literature. Spine (Phila Pa 1976) 30, E369–E374 (2005)   5. Edwards, R.J., David, K.M., Crockard, H.A.: Management of tuberculomas of the craniovertebral junction. Br J Neurosurg 14, 19–22 (2000)   6. Forsythe, M., Rothman, R.H.: New concepts in the diagnosis and treatment of infections of the cervical spine. Orthop Clin North Am 9, 1039–1051 (1978)   7. Frederickson, B., Yuan, H., Olans, R.: Management and outcome of pyogenic vertebral osteomyelitis. Clin Orthop Relat Res 131, 160–167 (1978)   8. Gormley, W., Rock, J.: Spontaneous atlantoaxial osteomyelitis: no longer a rare case? Case report. Neurosurgery 35, 132–135 (1994). discussion 135-136   9. Handmaker, H., Leonards, R.: The bone scan in inflammatory osseous disease. Semin Nucl Med 6, 95–105 (1976) 10. Heary, R.F., Hunt, C.D., Wolansky, L.J.: Rapid bony destruction with pyogenic vertebral osteomyelitis. Surg Neurol 41, 34–39 (1994) 11. Keller, T.: Victor Horsley’s surgery for cervical caries and fracture. The centennial anniversary. Spine (Phila Pa 1976) 21, 398–401 (1996) 12. Lam, C.H., Ethier, R., Pokrupa, R.: Conservative therapy of atlantoaxial osteomyelitis. A case report. Spine (Phila Pa 1976) 21, 1820–1823 (1996) 13. Leach, R.E., Goldstein, H.H., Younger, D.: Osteomyelitis of the odontoid process. A case report. J Bone Joint Surg Am 49, 369–371 (1967) 14. Limbird, T.J., Brick, G.W., Boulas, H.J., et al.: Osteomyelitis of the odontoid process. J Spinal Disord 1, 66–74 (1988)

233 15. Malawski, S.K., Lukawski, S.: Pyogenic infection of the spine. Clin Orthop Relat Res 272, 58–66 (1991) 16. Malkins, G.H., Abbot, F.C.: On acute primary osteomyelitis of the vertebrae. Ann Surg 23, 510–539 (1896) 17. Noguchi, S., Yanaka, K., Yamada, Y.: Diagnostic pitfalls in osteomyelitis of the odontoid process: case report. Surg Neurol 53, 573–578 (2000). discussion 578-579 18. Ottolenghi, C.E., Schajowicz, F., Deschant, F.A.: Aspiration biopsy of the cervical spine. Technique and results in thirtyfour cases. J Bone Joint Surg Am 46, 715–733 (1964) 19. Palestro, C.J., Kim, C.K., Swyer, A.J., et al.: Radionuclide diagnosis of vertebral osteomyelitis: indium-111-leukocyte and technetium-99m-methylene diphosphonate bone scintigraphy. J Nucl Med 32, 1861–1865 (1991) 20. Sapico, F.L., Montgomerie, J.Z.: Pyogenic vertebral osteomyelitis: report of nine cases and review of the literature. Rev Infect Dis 1, 754–776 (1979) 21. Sapico, F.L., Montgomerie, J.Z.: Vertebral osteomyelitis. Infect Dis Clin North Am 4, 539–550 (1990) 22. Sinha, S., Singh, A.K., Gupta, V.: Surgical management and outcome of tuberculous atlantoaxial dislocation: a 15-year experience. Neurosurgery 52, 331–338 (2003). discussion 338–339 23. Spies, E.H., Stucker, R., Reichelt, A.: Conservative management of pyogenic osteomyelitis of the occipitocervical junction. Spine (Phila Pa 1976) 24, 818–822 (1999) 24. Suchomel, P., Buchvald, P., Barsa, P.: Pyogenic osteomyelitis of the odontoid process: single stage decompression and fusion. Spine (Phila Pa 1976) 28, E239–E244 (2003) 25. Teegala, R., Kumar, P., Kale, S.S.: Craniovertebral junction tuberculosis: a new comprehensive therapeutic strategy. Neurosurgery 63, 946–955 (2008). discussion 955 26. Ueda, Y., Kawahara, N., Murakami, H.: Pyogenic osteomyelitis of the atlas: a case report. Spine (Phila Pa 1976) 34, E342–E345 (2009) 27. Venger, B.H., Musher, D.M., Brown, E.W., et  al.: Isolated C-2 osteomyelitis of hematogenous origin: case report and literature review. Neurosurgery 18, 461–464 (1986) 28. Wiedau-Pazos, M., Curio, G., Grusser, C.: Epidural abscess of the cervical spine with osteomyelitis of the odontoid process. Spine (Phila Pa 1976) 24, 133–136 (1999) 29. Zigler, J.E., Bohlman, H.H., Robinson, R.A., et al.: Pyogenic osteomyelitis of the occiput, the atlas, and the axis. A report of five cases. J Bone Joint Surg Am 69, 1069–1073 (1987)

Rheumatoid Arthritis

18

P. Suchomel, P. Buchvald, and O. Choutka

Rheumatoid arthritis (RA) is defined as a sterile erosive inflammation of the synovial membrane causing polyarthropathy and leading to destruction of ligamentous, cartilaginous, and bony structures with subsequent structural deformity and instability. Pannus formation, articular cartilage invasion, periarticular erosions, and destruction of adjacent structures can all be the result of this devastating process. RA affects the cervical spine in up to 86% of patients with majority of UCS involvement [33, 68]. Patients diagnosed as having RA must fulfill the revised criteria of the American Rheumatism Association [4].

18.1 Etiology and UCS Pathophysiology Although RA is a chronic auto-aggressive (autoimmune) reaction, the exact cause of the disease is still unknown. The era of intensive investigation of the immunologic aspects of RA began following the initial discovery of rheumatoid factor (RF) by Waaler in 1940 [56] and confirmation by Rose et al. in 1948 [66]. A genetic predisposition has also been identified. The combination of genetic susceptibility with yet unidentified inciting events can lead to disease expression [57, 64]. In the UCS, RA process induces AA (atlanto-axial) instability

and subluxation caused by insufficiency of the transverse and alar ligaments. Subsequent morphological destruction of AA and atlanto-occipital (AO) joints can cause vertical instability with the atlas telescoping downward and the odontoid process vertically migrating into the cranial cavity – so called cranial settling. As the odontoid process migrates cranially, it crowds the foramen magnum (FM) and, in the worst case scenario, compresses the brainstem. Neural compromise at this level can, however, be caused by a simple pannus formation and/or posterior C1 arch pressure in cases of AA subluxation [38, 41]. Subsequent progression of the disease that occurs in the majority of cases, thus, gradually advances from relatively simple AA instability to kyphotic deformity of the CVJ. In the final stages of the process, what is initially a mobile deformity ultimately becomes irreducible. Further, erosive fractures of lateral masses and the odontoid process are also known [41]. Subaxial cervical spine RA involvement typically includes multilevel subluxations called “stepladder deformity” caused by destruction of ligaments, facet joints, and discovertebral junctions (Fig. 18.1). However, this problem is less common than the above mentioned CVJ involvement (approx. 20% of all cervical RA).

18.2 History and Incidence P. Suchomel and P. Buchvald Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic O. Choutka University of Cincinnati, Medical Center, Department of Neurosurgery, Albert Sabin Way 231, Cincinnati, OH 45267-0515, USA

The first description of the disease is usually attributed to A.J. Landre-Beauvais [35], but the term rheumatoid arthritis was first used by A.B. Garrod in 1854 [22]. His son A.E. Garrod documented the RA cervical spine impact in 1890 [23]. In a group of 500 patients suffering from RA, he found 36% with affected cervical spine. The first descriptions of the actual destructive pathological changes in AA and

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_18, © Springer-Verlag Berlin Heidelberg 2011

235

236

18  Rheumatoid Arthritis

with RA and radiographic evidence of cervical subluxation [48]. The worst mortality rate was described by Crockard and Grob who reported that half of the RA patients presenting with myelopathy will be dead within one year [18]. However, only 10% of patients with juvenile RA will go on to develop destructive polyarthropathy with the potential for cervical spine problems. The others make a full recovery [16].

18.3 Clinical Symptoms

Fig. 18.1  Sagittal MRI in T2 sequence showing simultaneous RA odontoid pannus formation and subaxial cervical “stepladder deformity”

occipitocervical (OC) regions came from Englander [21] and Davis and Markley [19], who were also the first to describe a case of AA subluxation causing death in a RA patient. The worldwide incidence of RA is around 3 cases per 10,000 and it is two or three times more frequent in women. The prevalence of RA is approximately 1% of the world’s adult population when defined by either the presence of serum rheumatoid factor (RF) or erosive changes on radiographs in a patient with a compatible clinical presentation. The onset is more frequent during the fourth and fifth decades of life with 80% of all patients developing the disease between 35 and 50 years of age [61]. Cervical spine is affected in 44–88% of RA patients [5, 15, 45, 59, 63]. Based on the various studies, between 5% and 73% of RA patients will develop AA subluxation, about 20% will present with significant subaxial cervical spine disease, and approximately 17% will develop neurological sequelae [15, 27, 37, 40, 50]. Mikulowski et al., Paus et al., and Hamilton et al. estimated that 10% of patients with RA may die from brainstem compression that is unrecognized before their sudden death [28, 42, 49]. Oostveen et al. reported an overall mortality rate of 17% in patients

RA activity in the cervical spine begins early and progresses clinically and radiographically simultaneously with the peripheral joint disease. In fact, the severity of the peripheral erosive damage correlates with the degree of structural damage in the cervical spine [11, 54]. Even though most patients are initially asymptomatic, the signs related to cervical spinal abnormalities develop approximately in 60–80% of them [16]. The clinical manifestations can be extremely variable; however, neck pain with radiating occipital headache is often the leading symptom. Also, brachialgia with “frozen shoulder” can be seen more often. Limb paresthesias, weakness, vertigo, cranial nerve palsies, sphincter disturbance, and difficulty to walk could be the manifestations of myelopathy/neuropathy due to compression of spinal cord, brainstem, and cranial nerves. The Ranawat classification is most frequently used to categorize patients with rheumatoid myelopathy based on their clinical history and physical findings [52]. The Ranawat classification of neurologic deficit in RA: Class I – No neural deficit Class II – Subjective weakness, dysesthesias, and hyperreflexia Class IIIA – Objective weakness and long-tract signs; patient remains ambulatory Class IIIB – Objective weakness and long-tract signs; patient no longer ambulatory It is often difficult to distinguish which symptoms are caused by neural compression and which ones by the disease itself. The clinical appearance is also ­frequently modified by the concomitant pharmacological treatment – corticosteroids, nonsteroidal anti-­ inflammatory drugs (NSAIDs), disease-modifying ­anti-rheumatic drugs (DMARDs), biologic medications etc. This can indirectly lead to delayed diagnosis of neural compression and thus late surgical intervention. Marks and Sharp reported an average delay of

18.4  Radiology

31 weeks from the first appearance of neurologic signs to the correct diagnosis of myelopathy [36].

18.4 Radiology Plain radiography is still the first line of investigation when it comes to imaging of the rheumatoid cervical spine. Overall bony alignment, bone quality, and soft tissue swelling should be assessed. AA instability can be documented by flexion/extension views (Fig. 18.2). If any suspicion of RA process is established, then CT and MRI should be obtained to further depict the bone and soft tissue pathologic anatomy. Depending on the stage of the disease and subsequent surgical indication, other imaging modalities such as 3D bone CT, CTA, dynamic MRI or scintigraphy can be added. AA subluxation represents the most common manifestation of rheumatoid involvement of the spine [8]. In the majority of cases, the anterior type of subluxation is present although only 50% of these are symptomatic [50]. Anterior atlantodental interval (AADI) of 4–6 mm indicates early instability and implies transverse atlantal ligament damage or laxity. AADI larger than 6 mm indicates that the alar ligaments are also damaged. Posterior AA dislocation can occur in approximately 7% of all subluxations, if anterior C1 arch is defective or odontoid process eroded. This is, however, usually not accompanied by neural compression. Lateral subluxation is seen in approximately 20% of all AA

Fig. 18.2  Dynamic radiographs showing atlantoaxial instability in RA patient. (a) Flexion. (b) Extension

237

dislocations and usually is a result of lateral mass destruction or rotational deformity. It is defined as a lateral mass shift greater than 2 mm [7, 62]. Vertical subluxation, originally defined as a protrusion of the odontoid tip by more than 7 mm above the McGregor line [39], accounts for 22% of all dislocations. Other craniometric lines and indices obtained from plain films were historically used to establish the vertical odontoid migration [52, 53]. However, more recent publications demonstrated that the accuracy of these methods is substantially limited by the visibility of the anatomical landmarks on plain films, especially if their boundaries are eroded by RA process. It can be concluded that only a combination of plain film measurements can be of some practical value [55]. Currently, CT evaluation dominates in documentation of bone destruction. Reformatted sagittal images can precisely show the position of odontoid process and the amount of AA dislocation (Fig. 18.3). All the joints can be directly visualized and the actual focus (AA or AO joint) of disease defined. The total amount of UCS distortion can be evaluated on three-dimensional CT reconstructions. CT also plays an essential role in showing the exact morphology and amount of bone available for screw placement prior to any surgical fixation. According to contemporary studies of Chen et al. [12] and Myiata et al. [44], one may need to anticipate a high-riding vertebral artery (VA) in 31–70% of RA patients with AA instability or other UCS RA (Fig. 18.4). This translates into an increasing need for CTA evaluation in this situation followed by

238

18  Rheumatoid Arthritis

Fig. 18.3  CT sagittal image showing AA dislocation and odontoid process destruction caused by RA (a), CT 3D reconstruction depicting the AA displacement in another patient (b)

Fig. 18.4  High riding VA in RA patient. (a) Sagittal image showing the AA dislocation. (b) Left enlarged FT of C2 axial scan. (c) Sagittal image showing lack of space for isthmic screw placement

Fig. 18.5  T2 MRI weighted images in neurologically intact RA patient with AA dislocation. (a) Sagittal image, note preodontoid formation of inflammatory synovial cyst. (b) Axial image of the same patient, note spinal cord deformity and its posterolateral dislocation

three-dimensional reconstructions and navigation techniques in patients planned for posterior instrumentation in order to minimize the risk of VA injury. CT myelography can be used if MRI is contraindicated (pacemaker) with acceptable accuracy. However, MRI has become the preferred modality for evaluation of

CVJ [46]. The medulla, brainstem, soft tissue destruction, ligaments, pannus formation, bone swelling or relationship of the odontoid and/or pannus to the neural tissue and change of cervicomedullary angle (CMA) can be directly seen (Fig. 18.5). MRI proves invaluable in cases where soft odontoid pannus is large and

18.5  Treatment Strategy

239

Fig. 18.6  Dynamic MRI of RA patient. (a) Extension showing sufficient space for spinal cord. (b) Flexion causing atlantal forward dislocation with posterior arch ­compressing the cord

compressive while the degree of AA dislocation and/or bone destruction is much less impressive. Recently, dynamic (flexion-extension) MRI (Fig. 18.6) has been shown to be able to delineate the instability and provide further information about the dynamics of CVJ, in particular about the alarming narrowing of posterior subarachnoid space during flexion [3, 48]. MRI is also a very important noninvasive tool capable of monitoring disease progression during patient follow-up.

18.5 Treatment Strategy The exact etiology of RA is not known and thus causal treatment for the disease does not exist yet. Therapy usually involves a combination of medications, patient education, rehabilitation, joint protection, and surgery focused on arthrodesis or joint replacement. It is generally accepted that early treatment of RA results in better outcomes. In clinically asymptomatic individuals with radiographically confirmed dangerous UCS instability or deformity, the goal of any surgical treatment is to establish spinal stability and prevent neurological sequelae of the disease. In symptomatic patients with radiographic compression, with or without concomitant neurological signs or symptoms, decompression with subsequent stabilization is indicated. It is a well-known fact that once myelopathy occurs in RA patients, prognosis is poor and can hardly be altered by intervention [28]. Careful follow-up and early surgical intervention is the key to prevention of neurological decline and potential mortality as the disease is indeed progressive. Currently, surgical treatment is clearly indicated in RA patients with intractable pain and/or neurological

deficit with corresponding morphological background. Even further, it is also clear that some of those without a deficit or major pain syndrome can benefit from preventative surgery. This is especially true in cases of AA dislocation due to the well known natural progression over time. In Smith’s series of RA patients [63], progression of AADI occurred in 55% of patients during a 4.5  year follow-up. It advanced from an initial average distance of 3.5–5.0 mm to 5.0–8.0 mm in 45% and over 8.0 mm in another 10% of patients. However, there is no single parameter capable of predicting the future development of myelopathy and thus most authors prefer various combinations of predicting measures [9, 10, 13, 14, 20, 67]. Schizas et  al. believe that patients with AADI greater than 6 mm in flexion should be treated surgically whereas those with cranial settling should undergo surgery if AADI is more than 3 mm [58]. Boden reported that PADI is more correlative to eventual presence of cord compression and recommended surgery when PADI was less than 14  mm [6]. Shen et al. suggest surgical intervention in those with PADI less than 14 mm measured on dynamic plain films and SAC (space available for cord) less than 13 mm and/or CMA less than 135° measured on MRI. If the AA dislocation is accompanied by cranial settling, they advise to be more aggressive and to perform surgery in presence of any cord compression. However, they accept observational strategy if there is settling without neural compression and clinical symptoms [60]. Although there is no doubt that patients with neurological deficit caused by morphological compression have to be operated on, the questions remain if anterior, posterior or combined procedure should be performed as well as how many segments should be included in the fixation and fusion.

240

For further surgical decision, it is also necessary to determine whether subluxation and/or deformity can be reduced. A reducible lesion is defined as one in which relief of compression of the cervicomedullary neural structures could be obtained by restoring the anatomic relationship of CVJ [41]. Generally, reduction can be accomplished by simple positioning or by traction. Cervical skeletal traction is applied to patients with cranial settling and/or positionally irreducible kyphotic deformity. Depending on surgeon’s experience, traction can be used for shorter or longer periods, preoperatively. Simple mobile AA dislocation is most frequently treated by posterior reduction, fixation, and fusion. The immediately stable, fixation methods are preferred today and thus the majority of patients are treated with posterior C1-2 transarticular screws (Magerl), posterior C1 lateral mass, and C2 pedicle screw construct (GoelHarms) or, in case of high-riding VA, by posterior C1 lateral mass screws connected by rods to C2 laminar screws (Wright). Technical details of previous methods are described in Chap. 6. Due to poor bone quality and limited healing potential in RA, autologous bone grafts have to be used to supplement posterior AA fixation. Asymptomatic patients with predominantly a pannus formation can be treated by simple AA fixation and fusion. It is known that at least part of pannus tissue is provoked to grow by abnormal movement related to mechanical irritation that can be eliminated by fusion. Pannus size decreases or even disappears after AA fusion [26, 43, 70, 72]. AA fusion can also improve the rate of subsequent vertical odontoid migration [25]. In cases of fixed AA dislocation, posterior C1 laminectomy and fusion or transoral decompression and fusion are recommended [41]. If AO joint is affected by the RA process, posterior OC constructs are used to stabilize the entire CVJ. In such cases, the extent of fusion should not exceed the damaged segments [25]. In cases of cranial settling, with or without coincident CVJ kyphosis, it is essential to clarify if the deformity can be reduced by traction or not. If the odontoid process can be successfully drawn out of the FM, then posterior OC fixation and fusion is considered as sufficient [41, 60]. If sufficient reduction cannot be achieved but neural compression is not evident the same treatment can be selected [60]. However, if brainstem is compressed despite adequate reduction attempts, then decompression should precede any fusion procedure. Transoral decompression is indicated in all irreducible dislocations with marked anterior compromise [32]. The first transoral procedure for UCS RA deformity

18  Rheumatoid Arthritis

was performed by Sukoff et al. in 1972 [65]. Usually, anterior decompression is followed by posterior stabilization [17, 25, 41, 60]. Alternatively, as suggested by Harms, anterior plating between C1 lateral masses and C2 vertebral body could be used [29]. However, Kandziora et  al. [31] proved that such construct is not stable enough without posterior AA graft and wire fusion. Other plating systems were subsequently introduced that allowed for stand-alone anterior fixation and, with some, even for reduction of AA kyphosis [31, 69]. Although, the stability provided by these anterior constructs was biomechanically comparable to Magerl’s method [30], only a handful of patient series treated in this manner have been reported [2, 32]. Other options include posterior reduction by forced lordosis of C2 transpedicular screw attached to a rod anchored in occipital plate, as described by Abumi [1]; or a direct distraction of collapsed AA joints with cage and C1-2 plate fixation [24]. Zygmunt et al. [71] found that the majority of failures of OC fusion constructs in RA patients were the result of a progressive subaxial instability (37 of 163 cases), particularly at C3-4 and C4-5 levels. Krause et al. [34] described a 36% overall incidence of subaxial instability after OC fusion. This complication can by minimized by assessment of subaxial spine with dynamic plain films to identify any subaxial instability prior to a planned OC fusion. A more caudal extension of OC construct would then need to be carried out if instability exists at the lower levels. The other important key point is to respect sagittal alignment during any OC fixation to avoid undue overload of adjacent mobile segments [47, 51].

18.6 Our Preference There is a great variability of data evaluating the incidence of cervical spine involvement in RA patients and its consequences. Therefore, for practical purposes, we can assume that more than half of the patients suffering from RA would suffer from cervical spine disease. Of those, approximately 80% would present with UCS involvement and 20% with subaxial cervical spine subluxations. In the UCS, AA subluxation is the most frequently detected abnormality, with the ­majority (70%) being an anterior dislocation. Lateral, ­posterior, and rotational subluxations are much less frequent. Cranial settling of AA and AO joints is seen in about 20% of those with UCS disease. Certainly, any

18.5  Treatment Strategy

c­ ombination of previous dislocations is possible. Approximately 50% of RA patients with cervical spine involvement can expect a radiographic progression of their disease and one quarter of patients will develop neurological compromise over 10 years. Generally, those with radiographic evidence of disease without gross instability, neurological deficit or intractable pain can be followed conservatively. However, once signs of clinical and/or radiological progression are detected, surgical intervention needs to be considered. At our institution, we offer surgical treatment to those patients with AA subluxation where the AADI is more than 6 mm and PADI less than 14mm on plain dynamic films and/or SAC less than 13 mm with neural compression visible on MRI. We do so even in the absence of neurologic deficit and good control of symptoms. Patients with AA instability with intractable pain and/or neurologic deficit are surgical candidates without discussion. The coexistence with cranial settling makes the decision for surgical intervention more imperative in both previously mentioned groups.

241

Reducible AA dislocations are treated surgically with posterior C1-2 fusion. Majority of our patients are treated with transarticular C1-2 fusion according to Magerl supplemented with posterior autologous H  graft fixed to arches with titanium braided wire (Fig. 18.7). In partially reducible AA subluxations, we prefer the use of Goel-Harms C1-2 posterior fixation as it allows proper opening of the AA joint, C1 manipulation, and eventual C1 laminectomy (Fig. 18.8). In cases of high-riding VA, we always prepare a virtual three-dimensional screw trajectory plan. If pedicle or isthmic screw is not safe, we opt for a combination construct of a laminar screw (Wright) on the affected side and a pedicle/isthmic screw on the other. Exceptionally bilateral high-riding VA can be found, then C2 double-crosslaminar screw fixation can be a choice. In positionally irreducible kyphotic deformity localized predominantly to C1-2 segment, we prefer transoral odontoidectomy potentially with AA joint release followed by immediate posterior screw and graft fusion (Fig. 18.9). In such cases, we do not try to reduce the deformity with traction.

Fig.  18.7  Reduced AA dislocation fixed by Magerl screws supplemented with posterior autologous Gallie-type graft fixed by braided titanium wire. (a) Preoperative lateral plain image. (b) Preoperative MRI. (c) Postoperative lateral film

Fig. 18.8  Reduced AA dislocation due to manipulation with C1 lateral mass screws of Goel-Harms construct. (a) Preoperative CT showing AA dislocation and embarking cranial settling.

(b) Postoperative CT depicting the reduction achieved. (c) Harms fixator in the same patient

242

18  Rheumatoid Arthritis

Fig.  18.9  Fixed AA deformity treated with transoral odontoidectomy and joint release followed by posterior reduction and transarticular C1-2 fusion. (a) Plain film in flexion. (b) Plain film in extension. (c) Sagittal CT reconstruction. (d) Postoperative

sagittal CT documenting the extend of TO odontoidectomy. (e) 3D CT showing successful reduction, note the vicinity of the screw to C2 FT

If there is radiographic evidence of rheumatoid pannus compressing the spinal cord/brainstem without neurological symptoms or signs, we opt for a simple posterior fixation and fusion only if voluntary extension of the spine reduces the dislocation without complaints. However, in presence of neurological deficit, we would perform a transoral decompression first. This, perhaps unusual preference of transoral decompression, stems from good long-term results with this approach at our institution. A routine odontoidectomy represents approximately one hour of relatively safe surgery whereas long-term skeletal traction is not only uncomfortable but can also be rather morbid in patients with marked cord compression and common RA-related systemic problems (Fig.  18.10). If any attempt to reduce by traction is planned for AA deformity, then it is done in an anesthetized patient just prior to the planned procedure under electrophysiological monitoring.

If there is evidence of a significant AO joint involvement on CT in patients with CVJ instability, we prefer to extend the fixation and fusion to the occiput. Our strategy of treatment of rare complex CVJ deformities caused by RA is different. Intracranial odontoid migration/basilar invagination often accompanied by CVJ kyphosis can be either reducible or irreducible, with or without neurologic deficit. In neurologically intact patients, skeletal traction should be attempted, in our opinion, even if longer term. If the deformity can be reduced as evidenced by concomitant MRI-proven neural decompression, then posterior OC fusion is usually sufficient for those patients. If reduction is not successful but there is no significant compression on MRI, OC fusion can also be selected (Fig. 18.11). In patients with irreducible deformity and with neurological deficit with radiographic evidence of compression, we prefer direct decompression supplemented by fixation and fusion. Most frequently, we use a

18.5  Treatment Strategy

transpalatopharyngeal anterior approach with posterior OC fusion at the same sitting. In conclusion, it is important to emphasize that wellselected patients suffering from RA with cervical spine involvement can benefit from surgical intervention

243

with an expected improvement in two thirds of them. Therefore, it is of utmost importance to cooperate with referring physicians in early indication of appropriate candidates but also in long-term evaluation of surgical results.

Fig. 18.10  Bedridden patient with quadriparesis caused by cord compression treated by transoral decompression and posterior fusion without initial traction attempt. (a) MRI before the surgery. (b) MRI after surgery

Fig. 18.11  Neurologically intact patient with traction-irreducible cranial settling treated by C1 posterior laminectomy simple OC fixation. (a) T1 MRI sagittal image showing vertical ­intracranial

odontoid migration. (b) CT in the same plane. (c) Plain films of O–C2 pedicle screw fixation. (d) Postoperative MRI

244

18  Rheumatoid Arthritis

Fig. 18.11  (Continued)

References   1. Abumi, K., Takada, T., Shono, Y., et al.: Posterior occipitocervical reconstruction using cervical pedicle screws and platerod systems. Spine (Phila Pa 1976) 24, 1425–1434 (1999)   2. Ai, F., Yin, Q., Wang, Z., et al.: Applied anatomy of transoral atlantoaxial reduction plate internal fixation. Spine (Phila Pa 1976) 31, 128–132 (2006)   3. Allmann, K.H., Uhl, M., Uhrmeister, P., et  al.: Functional MR imaging of the cervical spine in patients with rheumatoid arthritis. Acta Radiol 39, 543–546 (1998)   4. Arnett, F.C., Edworthy, S.M., Bloch, D.A., et  al.: The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 31, 315–324 (1988)   5. Bland, J.: Rheumatoid arthritis of the cervical spine. J Rheumatol 1, 319–342 (1974)   6. Boden, S.D.: Rheumatoid arthritis of the cervical spine. Surgical decision making based on predictors of paralysis and recovery. Spine (Phila Pa 1976) 19, 2275–2280 (1994)   7. Bogduk, N., Major, G.A., Carter, J.: Lateral subluxation of the atlas in rheumatoid arthritis: a case report and post-mortem study. Ann Rheum Dis 43, 341–346 (1984)   8. Bouchaud-Chabot, A., Liote, F.: Cervical spine involvement in rheumatoid arthritis. A review. Joint Bone Spine 69, 141– 154 (2002)   9. Bundschuh, C.V., Alley, J.B., Ross, M., et  al.: Magnetic resonance imaging of suspected atlanto-occipital dislocation. Two case reports. Spine (Phila Pa 1976 17, 245–248 (1992) 10. Bundschuh, C., Modic, M.T., Kearney, F., et al.: Rheumatoid arthritis of the cervical spine: surface-coil MR imaging. AJR Am J Roentgenol 151, 181–187 (1988)

11. Cassar-Pullicino, V.N.: The spine in rheumatological disorders. Imaging 11, 104–118 (1999) 12. Chen, T.Y., Lin, K.L., Ho, H.H.: Morphologic characteristics of atlantoaxial complex in rheumatoid arthritis and surgical consideration among Chinese. Spine (Phila Pa 1976) 29, 1000–1004 (2004). discussion 1005 13. Clark, C.R., Goetz, D.D., Menezes, A.H.: Arthrodesis of the cervical spine in rheumatoid arthritis. J Bone Joint Surg Am 71, 381–392 (1989) 14. Conaty, J.P., Mongan, E.S.: Cervical fusion in rheumatoid arthritis. J Bone Joint Surg Am 63, 1218–1227 (1981) 15. Conlon, P.W., Isdale, I.C., Rose, B.S.: Rheumatoid arthritis of the cervical spine. An analysis of 333 cases. Ann Rheum Dis 25, 120–126 (1966) 16. Crockard, H.A.: Surgical management of cervical rheumatoid problems. Spine (Phila Pa 1976) 20, 2584–2590 (1995) 17. Crockard, H.A., Calder, I., Ransford, A.O.: One-stage transoral decompression and posterior fixation in rheumatoid atlanto-axial subluxation. J Bone Joint Surg Br 72, 682–685 (1990) 18. Crockard, H.A., Grob, D.: Rheumatoid arthritis: upper cervical involvement. In: Clark, C.R. (ed.) The cervical spine, pp. 705–714. Lippincot, Philadelphia (1998) 19. Davis Jr., F.W., Markley, H.E.: Rheumatoid arthritis with death from medullary compression. Ann Intern Med 35, 451–454 (1951) 20. Dickman, C.A., Sonntag, V.K., Papadopoulos, S.M., et al.: The interspinous method of posterior atlantoaxial arthrodesis. J Neurosurg 74, 190–198 (1991) 21. Englander, O., Prague: Non-traumatic occipito-atlanto-axial dislocation. Br. J. Radiol. 15, 341–345 (1942) 22. Garrod, A.B.: On the blood and effused fluids in gout, rheumatism and Bright’s disease. Tran Med Chir Soc Edinb 37, 49 (1854)

References 23. Garrod, A.E.: A Treatise on Rheumatism and Rheumatoid Arthritis. Griffin, London (1854) 24. Goel, A.: Craniovertebral stabilization. J Neurosurg Pediatr 1, 173–175 (2008). author reply 175 25. Grob, D.: Atlantoaxial immobilization in rheumatoid arthritis: a prophylactic procedure? Eur Spine J 9, 404–409 (2000) 26. Grob, D., Wursch, R., Grauer, W., et al.: Atlantoaxial fusion and retrodental pannus in rheumatoid arthritis. Spine (Phila Pa 1976) 22, 1580–1583 (1997). discussion 1584 27. Halla, J.T., Hardin Jr., J.G.: The spectrum of atlantoaxial facet joint involvement in rheumatoid arthritis. Arthritis Rheum 33, 325–329 (1990) 28. Hamilton, J.D., Johnston, R.A., Madhok, R., et al.: Factors predictive of subsequent deterioration in rheumatoid cervical myelopathy. Rheumatology (Oxf) 40, 811–815 (2001) 29. Harms, J., Schmelze, R., Stolze, D.: Osteosynthesen im occipito-cervikalen Übergang vom transoralen Zugang aus. XVII SICOT World Congress Abstracts. Demeter Verlag, Munich (1987) 30. Kandziora, F., Kerschbaumer, F., Starker, M., et  al.: Bio­ mechanical assessment of transoral plate fixation for atlantoaxial instability. Spine (Phila Pa 1976) 25, 1555–1561 (2000) 31. Kandziora, F., Pflugmacher, R., Ludwig, K., et  al.: Biomechanical comparison of four anterior atlantoaxial plate systems. J Neurosurg 96, 313–320 (2002) 32. Kerschbaumer, F., Kandziora, F., Klein, C., et al.: Transoral decompression, anterior plate fixation, and posterior wire fusion for irreducible atlantoaxial kyphosis in rheumatoid arthritis. Spine (Phila Pa 1976) 25, 2708–2715 (2000) 33. Kolen, E.R., Schmidt, M.H.: Rheumatoid arthritis of the cervical spine. Semin Neurol 22, 179–186 (2002) 34. Kraus, D.R., Peppelman, W.C., Agarwal, A.K., et  al.: Incidence of subaxial subluxation in patients with generalized rheumatoid arthritis who have had previous occipital cervical fusions. Spine (Phila Pa 1976) 16, S486–S489 (1991) 35. Landre-Beauvais A.J.: La goutte asthénique primitive (doctoral thesis). Paris (1800) 36. Marks, J.S., Sharp, J.: Rheumatoid cervical myelopathy. Q J Med 50, 307–319 (1981) 37. Mathews, J.A.: Atlanto-axial subluxation in rheumatoid arthritis. Ann Rheum Dis 28, 260–266 (1969) 38. Mathews, J.A.: Atlanto-axial subluxation in rheumatoid arthritis. A 5-year follow-up study. Ann Rheum Dis 33, 526– 531 (1974) 39. McGregor, M.: The significance of certain measurements of the skull in the diagnosis of basilar impression. Br J Radiol 21, 171–181 (1948) 40. Meikle, J.A., Wilkinson, M.: Rheumatoid involvement of the cervical spine. Radiological assessment. Ann Rheum Dis 30, 154–161 (1971) 41. Menezes, A.H., VanGilder, J.C., Clark, C.R., et al.: Odontoid upward migration in rheumatoid arthritis. An analysis of 45 patients with “cranial settling”. J Neurosurg 63, 500–509 (1985) 42. Mikulowski, P., Wollheim, F.A., Rotmil, P., et  al.: Sudden death in rheumatoid arthritis with atlanto-axial dislocation. Acta Med Scand 198, 445–451 (1975) 43. Milbrink, J., Nyman, R.: Posterior stabilization of the cervical spine in rheumatoid arthritis: clinical results and magnetic resonance imaging correlation. J Spinal Disord 3, 308–315 (1990)

245 44. Miyata, M., Neo, M., Ito, H., et al.: Is rheumatoid arthritis a risk factor for a high-riding vertebral artery? Spine (Phila Pa 1976) 33, 2007–2011 (2008) 45. Nakano, K.K., Schoene, W.C., Baker, R.A., et al.: The cervical myelopathy associated with rheumatoid arthritis: analysis of patients, with 2 postmortem cases. Ann Neurol 3, 144–151 (1978) 46. Narvaez, J.A., Narvaez, J., Serrallonga, M., et  al.: Cervical spine involvement in rheumatoid arthritis: correlation between neurological manifestations and magnetic resonance imaging findings. Rheumatology (Oxf) 47, 1814–1819 (2008) 47. O’Brien, M.F., Casey, A.T., Crockard, A., et al.: Histology of the craniocervical junction in chronic rheumatoid arthritis: a clinicopathologic analysis of 33 operative cases. Spine (Phila Pa 1976) 27, 2245–2254 (2002) 48. Oostveen, J.C., Roozeboom, A.R., van de Laar, M.A., et al.: Functional turbo spin echo magnetic resonance imaging versus tomography for evaluating cervical spine involvement in rheumatoid arthritis. Spine (Phila Pa 1976) 23, 1237–1244 (1998) 49. Paus, A.C., Steen, H., Roislien, J., et al.: High mortality rate in rheumatoid arthritis with subluxation of the cervical spine: a cohort study of operated and nonoperated patients. Spine (Phila Pa 1976) 33, 2278–2283 (2008) 50. Pellicci, P.M., Ranawat, C.S., Tsairis, P., et al.: A prospective study of the progression of rheumatoid arthritis of the cervical spine. J Bone Joint Surg Am 63, 342–350 (1981) 51. Pham, X.V., Bancel, P., Menkes, C.J., et al.: Upper cervical spine surgery in rheumatoid arthritis: retrospective study of 30 patients followed for two years or more after CotrelDubousset instrumentation. Joint Bone Spine 67, 434–440 (2000) 52. Ranawat, C.S., O’Leary, P., Pellicci, P., et al.: Cervical spine fusion in rheumatoid arthritis. J Bone Joint Surg Am 61, 1003–1010 (1979) 53. Redlund-Johnell, I., Pettersson, H.: Radiographic measurements of the cranio-vertebral region. Designed for evaluation of abnormalities in rheumatoid arthritis. Acta Radiol Diagn (Stockh) 25, 23–28 (1984) 54. Resinck, D.: Diagnosis of Bone and Joint Disorders, pp. 891–974. Saunders, Philadelphia (2002) 55. Riew, K.D., Hilibrand, A.S., Palumbo, M.A., et  al.: Diagnosing basilar invagination in the rheumatoid patient. The reliability of radiographic criteria. J Bone Joint Surg Am 83-A, 194–200 (2001) 56. Rose, H.M., Ragan, C., et al.: Differential agglutination of normal and sensitized sheep erythrocytes by sera of patients with rheumatoid arthritis. Proc Soc Exp Biol Med 68, 1–6 (1948) 57. Saag, K.G., Cerhan, J.R., Kolluri, S., et al.: Cigarette smoking and rheumatoid arthritis severity. Ann Rheum Dis 56, 463–469 (1997) 58. Schizas, C., de Goumoens, P., Fragniere, B.: Rheumatoid arthritis of the cervical spine: surgical management. Rev Med Suisse Romande 124, 575–578 (2004) 59. Sharp, J., Purser, D.W.: Spontaneous atlanto-axial dislocation in ankylosing spondylitis and rheumatoid arthritis. Ann Rheum Dis 20, 47–77 (1961) 60. Shen, F.H., Samartzis, D., Jenis, L.G., et  al.: Rheumatoid arthritis: evaluation and surgical management of the cervical spine. Spine J 4, 689–700 (2004)

246 61. Silman, A.J., Pearson, J.E.: Epidemiology and genetics of rheumatoid arthritis. Arthritis Res 4(Suppl 3), S265–S272 (2002) 62. Silvaggio, V., Donaldson, W.F., Kraus, D.R.: Surgery of rheumatoid arthritis of cervical spine. In: Bridwell, K.H., DeWald, R.L. (eds.) Textbook of Spinal Surgery, pp. 1435– 1455. Lippincot, Philadelphia (1997) 63. Smith, P.H., Benn, R.T., Sharp, J.: Natural history of rheumatoid cervical luxations. Ann Rheum Dis 31, 431–439 (1972) 64. Stolt, P., Bengtsson, C., Nordmark, B., et al.: Quantification of the influence of cigarette smoking on rheumatoid arthritis: results from a population based case-control study, using incident cases. Ann Rheum Dis 62, 835–841 (2003) 65. Sukoff, M.H., Kadin, M.M., Moran, T.: Transoral decompression for myelopathy caused by rheumatoid arithritis of the cervical spine. Case report. J Neurosurg 37, 493–497 (1972) 66. Waaler, E.: On the occurrence of a factor in human serum activating the specific agglutination of sheep red corpuscles. Acta Pathol Microbiol Scand 17, 172–188 (1940) 67. Weissman, B.N., Aliabadi, P., Weinfeld, M.S., et  al.: Prognostic features of atlantoaxial subluxation in rheumatoid arthritis patients. Radiology 144, 745–751 (1982)

18  Rheumatoid Arthritis 68. Winfield, J., Cooke, D., Brook, A.S., et  al.: A prospective study of the radiological changes in the cervical spine in early rheumatoid disease. Ann Rheum Dis 40, 109–114 (1981) 69. Yin, Q., Ai, F., Zhang, K., et al.: Irreducible anterior atlantoaxial dislocation: one-stage treatment with a transoral atlantoaxial reduction plate fixation and fusion. Report of 5 cases and review of the literature. Spine (Phila Pa 1976) 30, E375–E381 (2005) 70. Young, W.F., Boyko, O.: Magnetic resonance imaging confirmation of resolution of periodontoid pannus formation following C1/C2 posterior transarticular screw fixation. J Clin Neurosci 9, 434–436 (2002) 71. Zygmunt, S.C., Christensson, D., Saveland, H., et  al.: Occipito-cervical fixation in rheumatoid arthritis – an analysis of surgical risk factors in 163 patients. Acta Neurochir (Wien) 135, 25–31 (1995) 72. Zygmunt, S., Saveland, H., Brattstrom, H., et al.: Reduction of rheumatoid periodontoid pannus following posterior occipito-cervical fusion visualised by magnetic resonance imaging. Br J Neurosurg 2, 315–320 (1988)

19

Tumors P. Suchomel, V. Benes, and M. Kaiser

The basic goal of treatment is to achieve decompression of the neural elements, ideally by complete tumor resection, and reconstructing the spine while maintaining normal sagittal balance and minimizing loss of motion segments. This goal is often difficult or even impossible due to anatomical restraints and the proximity or involvement of neurovascular structures. In certain types of primary, benign bone tumors, control or cure can be achieved by partial tumor resection. This, however, is the exception and in most tumors, gross total excision needs to be achieved when technically feasible in order to prevent recurrence or progression. Comprehensive studies of tumors in the UCS region do not exist; however, we do know that, with a few exceptions, the general occurrence of tumors is similar to the subaxial spine. The occurrence of chordoma in the C2 region is a notable exception. In general, C1 is less commonly affected as C2 is affected by tumors more often [60].

19.1 Extradural UCS Tumors The most common types of extradural tumors at the UCS are secondary tumors, at about 90%. Primary tumors are less common but are usually difficult to

P. Suchomel, V. Benes, and M. Kaiser Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova St. 10, 46063 Liberec, Czech Republic

treat or cure [113]. The current approach to treatment of spinal extradural tumors is multidisciplinary. The treatment teams are usually composed of a spine surgeon, oncologist, pathologist, and organ specialist (urologist in renal tumors, dermatologist in melanoma, etc.). This team must evaluate the patient’s general status, establish the diagnosis, the prognosis, and recommend the most effective treatment. This same team theoretically should follow the patient’s course and final outcome, all of which, unfortunately, is not common in practice.

19.1.1 Radiological Remarks The initial evaluation of patients with suspected spinal extradural neoplasms is plain radiographic examination. Plain radiographs require a 30% to 50% demineralization to detect a destructive process within the vertebral body (Fig. 19.1) [32]. Plain radiographs allow localization and determination of the extent of tumor involvement, and with dynamic films, spinal stability can be evaluated. For patients with a known primary cancer, bone scintigraphy has been standard but more recently, positron emission tomography (PET) scan is used to screen for metastatic lesions. Although bone scans are very sensitive in determining the presence of a high bone metabolic turnover state and/or hyperemia, they cannot differentiate whether the process is an infection, a healing fracture or a tumor. CT scanning permits direct visualization of the bone, any destruction, and allows evaluation of the anatomy for planning for eventual stabilization procedures (Fig. 19.2). MRI can provide information about soft tissue, including extent of tumor involvement (Fig.  19.3), neural

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_19, © Springer-Verlag Berlin Heidelberg 2011

247

248

19  Tumors

the vertebral artery (VA) (Fig. 19.4) [53], adjacent neural structures, and the oropharynx (Fig. 19.5) [77].

19.1.2 Therapeutic Remarks Depending on tumor staging, histology, age, and ­general patient status the consequent patient risk ­versus benefit ratio can be evaluated. The final therapeutic protocol can vary from observation with no therapy (life expectancy less than 6 weeks) to radical, ­aggressive surgery followed by radiation and chemotherapy. Given the surgical focus of this book, it is important to emphasize that any excision of tumor may impact the statodynamic properties of the spine and in such cases, necessitates appropriate spinal column reconstruction. Fig. 19.1  Plain lateral film of C2 body destruction by metastasis of breast cancer in a 50 year old woman

19.1.3 Surgical Oncologic Terms

compression, and is also good to determine multilevel spine involvement. The use of gadolinium on T1-weighted images enhances the tumor tissue and allows delineation of normal and abnormal processes. Evaluation of cervical tumors not only requires defining the extent of the lesion but also its relationship to

The terms used to describe the extent of surgical resection are a constant source of confusion. Often, one can read that the resection was performed: “en bloc,” “gross total,” or that there was a “radical resection.” Also, the term “margin” often mixes concept of surgical margin with histological margins.

Fig. 19.2  Destruction of C2 by renal carcinoma metastasis. (a) Coronal plane CT reconstruction. (b) Anterior view 3D CT

19.1  Extradural UCS Tumors

Fig. 19.3  C2 affected by metastasis of renal carcinoma, notice the canal compromise (patient from Fig. 19.2)

249

Fig.  19.5  Displacement of oropharynx in a case of anterior ­prevertebral C2 chordoma spread shown on lateral contrast swallow study

It is difficult to adapt the oncologic definitions used in long bone and soft tissue pathologies for spine ­surgery purposes. Single-piece removal of the whole vertebra in the UCS is not possible due to the spinal cord. Also, definition of an extracompartmental ­margin of resection fails if the tumor reaches the extradural space, which is in continuity through the whole spine. It is clear that the surgical margin is the surgeon’s subjective definition of the border of the tumor formed by either a capsule or pseudocapsule of reactive tissue, or by a layer of healthy tissue surrounding infiltrative neoplasms. Conversely, the histological margin can be objectively confirmed as tumor-free only by the laboratory investigations of the specimen after removal. In concordance with others, we would like to define the meaning of each term [45, 49, 133]. Respecting the surgical margin of the tumor during resection:

Fig. 19.4  Narrowed vertebral artery passing through C2 chordoma with extracompartmental extend depicted on vertebral angiogram

• Intra or endolesional (means the surgical margin of the tumor is disrupted during the procedure) • Marginal (means single-piece tumor removal along its surgical border without margin violation) • Wide (means single-piece removal of tumor surrounded by continuous layer or cuff of healthy tissue)

250

According to the method of resection, one can dis­ tinguish: • Piecemeal resection (means successive, step wise tumor removal - typical example is curettage) • En bloc resection (means, tumor removal in one piece regardless of its margins) • Spondylectomy (means, removal of the whole vertebra regardless of the method used to do it) In summary, if the tumor spreads from the body posterior through both pedicles, the oncologic wide margin resection is not possible. Moreover, if the whole vertebra is considered as the malignant tumor compartment, en bloc resection is not feasible in UCS (but is sometimes possible in lumbar region and may damage the cauda equina, which can be justified in certain lesions e.g.: osteogenic sarcoma or chordoma if the patient accepts such morbidity).

19  Tumors

Classification, Grading, and Staging In order to properly evaluate the prognosis, treatment feasibility and methods to follow outcomes, benign tumors should be divided into similar groups. Therefore, systems classifying tumors are necessary. The oncologic staging of primary bone tumors, based on clinical features, radiographic patterns, and histological behavior described for long bone neoplasms by Enneking et al. [44, 45] has been adopted for benign primary bone spinal tumors [21, 77]. It distinguishes three stages of benign and six stages of malignant primary spine tumors. The classification system is named after Enneking.

E nneking Staging of Primary Benign Spine Tumors SI

first stage, latent, asymptomatic, well-bordered by a true capsule

S II

second stage, slowly growing, symptomatic, thin capsule with reactive tissue

S III

third stage, aggressive, rapid growth, symptomatic, capsule discontinuous – absent, or reactive pseudocapsule, invades neighboring compartments

19.1.4 Primary Bone Tumors of UCS Primary bone tumors of the spine are relatively rare comprising about 4.2% of all spine tumors [1, 16].

19.1.4.1 Benign Primary Bone Tumors Benign lesions are less frequent than malignant lesions in the UCS [14, 43, 77]. They occur predominantly in the second and third decades of life and affect males twice as often as females [77]. The tumors most commonly seen in the region of the cervical spine are osteomas (44%), osteoblastomas (15%), and aneurysmal bone cysts (15%). Less frequently seen are eosinophilic granulomas (12%), giant cell tumors (5%), and osteochondromas (5%) [77]. Much rarer, and described in case reports are: hemangiomas, fibrous dysplasia, and a few cases of Gorham disease [76, 77]. Both, the atlas and axis can be involved; however, C2 due to the larger cancellous bone content is much more frequently affected. Although benign tumors that occur in the cervical spine are from a wide range of histological types, they have many common as well as differentiating features that will be described below.

A reproducible classification system that describes the tumor extent, localization, and surgical accessibility was proposed originally by Weinstein and was modified by Boriani et al. is named WBB surgical staging (Weinstein-Boriani-Biagini surgical classification, WBB) [21, 131].

WBB Surgical Staging WBB system divides the transverse extension of vertebral tumors into 12 radiating zones (numbered 1–12 in a clockwise order with no. 1 on the left side of the spinous process) and 5 concentric layers (A–E concentric zones from the paravertebral extraosseous compartments (A) to the dural involvement (E) and the VA foramen area (F)). These staging systems were tested for observer reliability with very good intra-observer reliability results; however, the inter-observer reliability was only moderate [34]. The Enneking and WBB staging status, together with tumor histology obtained by biopsy can give us an idea of potential treatment possibilities, the patient’s

251

19.1  Extradural UCS Tumors

prognosis and can also allow comparison of treatment strategies, and outcomes between centers. For example, a small osteoma of the right posterior C2 arch can be classified as: osteoma, Enneking S1, and WBB 11 B. This information is comprehensive enough to consult the oncologic center, establish prognosis and treatment strategy. If dimensions of the lesion measured on CT images are added, the information is complete for long-term follow-up and observation in conservatively treated patients.

Clinical Symptoms The clinical symptoms of primary bone tumors are nonspecific. The first symptom can be a constant or episodic neck pain unusual in a young individual. The symptoms are dependent on tumor location, its extent, and the likely eventual pathologic vertebral fracture, which can cause a simple radiculopathy or extend to create significant myelopathy. In osteoid osteomas, nocturnal pain is very characteristic and occurs in about 50% of patients [101]. Extracompartmental tumor spread, such as from aneurysmal bone cyst, can cause torticollis or even a palpable posterior neck mass.

Radiology Aggressive osteolytic neoplasms can easily be seen on plain radiographs, usually as a result of deformity or vertebral body collapse; however, most of the benign tumors are more accurately detected by CT or MRI. The CT scan with bone windows defines bony pathology precisely. The MRI is less effective in evaluating bony anatomy but can clearly depict the relationship of the tumor to the adjacent soft tissue and neural structures. In growing tumors with a reactive pseudocapsule, a radioisotope bone scan can reveal the hot spot leading to further investigations. As in other pathologies of the UCS, all the anatomy must be clearly understood if surgical decompression and reconstruction are planned.

General Treatment Strategy S1 tumors usually do not need treatment until a pathologic fracture or severe pain occurs; however, growing

lesions (S2-3) necessitate early surgical management to prevent clinical symptoms. Depending on the histology and surgical accessibility, radical tumor resection is preferred. Puncture biopsy is often helpful prior to the surgical procedure. In some lesions, endolesional or marginal resection is sufficient; however, in aggressive tumors with a known tendency to recur, radical resection to the extent that is allowed by anatomic UCS complexity should be performed. Wide resections involving the stabilizing structures (body, pedicle, and facets) have to be followed by a single stage UCS reconstruction. As many of the tumors can present in childhood, the growing potential of the immature spine must be considered. In these situations, using largely autologous bone that is fixed without growing restraints (wires often better than screws) is required. All the approaches described in Chap. 4. (posterior midline, high anterolateral, lateral, and transoral) can be used to reach the tumor mass. Direct or indirect embolization can help in the treatment of vascularized tumors. Despite some tumors being benign, there is often a tendency to recur and in some tumors adjuvant therapy is recommended. Although evidence is limited, focused radiotherapy is recommended in osteoblastomas, aneurysmal bone cysts, and LCH by some authors, especially in recurrent tumors [31, 47, 70]. Standard chemotherapy has not shown much effect in these tumors although bisphosphonates are sometimes tried.

Osteoid Osteomas and Osteoblastomas Osteoid-producing tumors such as osteoid osteomas and osteoblastomas are rare lesions in the UCS level. Osteoblastomas are seen less frequently than osteoid osteomas with an incidence of 10–25% of primary osseous spine tumors [6]. They affect young people with a male predominance of 2:1. Osteomas occur predominantly within the posterior elements, the arches, facet joints, and pedicles [18, 77] but are occasionally found in the vertebral body [118]. Both types of tumors are histologically similar, however osteoblastomas are more aggressive, larger (more than 20  mm), often involve both the anterior and posterior vertebral segments and can recur in up to 20% [76]. There is only one report of osteoma conversion to osteoblastoma in the literature [25], and in reviewing the literature there is not a clear definition of the differences between the

252

19  Tumors

two tumor types. Only the extremes, small osteomas and large, spreading osteoblastomas can be distinguished clearly from their radiological appearance. Diagnosis The clinical symptoms are often non-specific. The patients present with neck pain, classically with nocturnal peaks that responds to salicylates and NSAIDs. Osteoblastomas can be often visible on plain X-rays as an osteolytic lesion and can frequently cause vertebral body deformity (Fig. 19.6), however the small osteoid osteomas do not. CT scan is very useful in distinguishing the tumor type. While osteoma can typically be seen as sclerotic bone with a round radiolucent ovoid exophytic mass (Fig. 19.7), osteoblastoma is detected as a multilocular, cavernous lytic structure (Fig. 19.8) or as aggressive large osteoma (larger than 20  mm) without a sclerotic border.

Fig. 19.7  Osteoma of the C2 lamina with an internal exophytic extend shown on axial CT

Treatment Strategy Currently, intralesional excision (curettage) of the nidus is the widely accepted treatment of osteomas, even in the UCS area [89]. Nevertheless, radiofrequency ablation has been successfully used in their treatment [28, 29, 74]. Those who are asymptomatic or

respond to pharmacological management with aspirin or NSAIDs can be simply observed [26, 91]. Osteoblastomas also typically involve the posterior elements of the spine, but due to their larger size and aggressive behavior may extend into the pedicle,

Fig. 19.6  Osteoblastoma of C6 in a 16 year-old female. Osteolytic process led to body collapse – “vertebra plana” appearance

Fig. 19.8  Osteoblastoma of C6 (patient from Fig. 19.6) showing the whole vertebra involvement and honeycomb structure. Axial CT scan

253

19.1  Extradural UCS Tumors

vertebral body, and can compromise the spinal canal [59]. In 1964, Lichtenstein and Sawyer first reported a series of 20 osteoblastomas and concluded that these tumors throughout the skeletal system should generally be treated, conservatively. However, they emphasized that osteoblastomas located in the spinal axis should be surgically decompressed and then irradiated [79]. The recurrence rate for osteoblastomas after simple resection has been reported between 10% and 19% [66, 81]. In the largest review of 306 spinal osteoblastoma cases collected at Mayo Clinic over 17.5 years, complete treatment and clinical follow-up was only available for 75 patients. In this study, intralesional resection had a significant recurrence rate of 19.0% (10/52), marginal resection 5.6% (1/18), and surprisingly en bloc resection 20.0%. However, these radical resections were probably not wide-margin resections. The authors noted that these lesions have the potential for local recurrence after subtotal resection [81]. From the available literature it can be summarized that intralesional resection is recommended for nonaggressive osteoblastomas (Enneking 2) and wide-margin resection for aggressive osteoblastomas (Enneking 3) [59]. The role of radiation therapy in recurrent or incompletely resected tumors is controversial, with the majority of cases showing no advantage, but a minority demonstrating a benefit [134]. Marsh et al. concluded in their review of 197 osteoblastoma cases that “radiotherapy does not alter the course of the disease and appears to be contraindicated” [83]. Adjunctive radiation therapy with primary surgery probably can be beneficial in the case of stage 3 osteoblastomas as well as in recurrent tumors [18, 59]. Chemotherapy has a limited role in recurrent aggressive osteoblastomas and there are only a few case reports in the literature [59].

Our Preference In concordance with the Harrop et al. systematic review [59], we prefer to observe osteomas without clinical symptoms and located in noneloquent statodynamic spine areas. If patient becomes symptomatic, either radiofrequency CT-guided ablation [9] or enucleation of the osteoma is performed (Fig. 19.9). In nonaggressive osteoblastomas (Enneking stage S2), a radical resection is performed if anatomically possible, although endolesional tumor removal can be sufficient. Growing and aggressive tumors (Enneking

Fig. 19.9  Osteoma capsule after high-speed drill bone removal. Intraoperative picture

S3) should be resected as radically as possible. We have not had a patient with an S3 osteoblastoma in UCS region; however, we have had a good outcome after total spondylectomy of C6 osteoblastoma in a 16-year-old girl who has survived more than 7 years without recurrence (Fig. 19.10).

Aneurysmal Bone Cysts Aneurymal bone cysts (ABCs) are non-neoplastic in nature, but are expansile, composed of thin-walled cystic areas filled with blood. They were initially described by Jaffe and Lichtenstein in 1942 [67] and represent approximately 15% of all primary spine tumors but are found most frequently in flat bones (pelvis) [3, 39]. In 10–30% of ABC cases, the spine is involved. Commonly, they are located in the thoracic and lumbar spine [61], but around 25% of the time they can arise in the cervical spine [30, 77]. ABCs generally occur in first two decades of life without gender predilection. They are usually localized to posterior spine elements but circumferential vertebral involvement is not uncommon. They can reach enormous size without clinical consequences. Spontaneous resolution of ABC has been described; however, progression is more common [30, 61, 77, 127].

254

19  Tumors

(Fig. 19.12). Selective angiography can be performed for embolization and can reveal the common arterial feeders to the spinal cord.

Treatment Strategy

Fig. 19.10  Plain radiographs 7 years after total spondylectomy for osteoblastoma showing a good bone fusion without recurrence (patient from Figs 19.6 and 19.8). (a) Lateral projection. (b) AP

Simple curettage with or without bone grafting, complete resection, embolization only, radiation therapy or combination of these methods were used for treatment of ABCs [59]. En bloc resection appears to have the highest rate of cure; however, achieving complete excision is very difficult in ABCs, especially in the UCS. Boriani et al. reported only 2/41 cases in which they performed en bloc resection without adjuvant radiation [19]. Gross total endolesional resection using a highspeed drill has also been reported as effective with longterm cures [3, 19, 40, 61]. However, according to other authors, intralesional, incomplete excision is associated with a relatively high progression rate of up to 25% within the first 2 years after surgery in 90% of patients [19, 30, 61, 128]. Radiation therapy has limited primary indications and remains an adjuvant for patients with inoperable lesions, aggressive recurrent disease, and medical conditions that place them at a high surgical risk or in patients with incomplete excision. Boriani et al reported four patients without local disease progression after radiation alone [19]. Capanna et al. has seen three local progressions in six irradiated patients [30]. The successful use of preoperative, selective arterial embolization has been documented for ABCs of the pelvis and long bones, but its role as the sole mode of therapy in the spine is more questionable [76, 88]. Embolization alone may be considered in patients with a recurrent lesion after previous surgeries, or in patients who cannot medically tolerate surgery and only if pathological fracture, spinal deformity, instability, and neurologic compromise are absent [19]. It is important to note that fatalities have been reported with embolization procedures, particularly with cervical spine lesions [96].

Diagnosis Clinically, ABCs can be silent with only neck pain, but can present with vertebral collapse, deformity or signs of neural compression. Plain radiographs can demonstrate an osteolytic lesion with “ballooning out” of the bone cortex. This is confirmed by a typical CT appearance (Fig. 19.11) of thin-walled cavities deforming the vertebral bony borders filled with blood. MRI demonstrates the relationship to the spinal canal contents

Our Preference As we have only very limited experience with this type of primary bone lesion in the UCS, we have to follow the recommendations of Harrop et al. in their systematic review [59]. The aneurysmal bone cyst should be removed as radically as possible to prevent further disease progression. Due to anatomic restrictions, this goal can be difficult to achieve in the UCS, especially

19.1  Extradural UCS Tumors

255

Fig. 19.11  CT of aneurysmal bone cyst of C2 in a 7 year old girl. (a) Coronal plane reconstruction. (b) Sagittal reconstruction showing extracompartmental extend. (c) Axial scan

if the lesion involves vertebral structures circumferentially. As these lesions often destroy stabilizing structures, reconstruction of the spine is necessary. This has to be done with respect to the age of patient, keeping in mind they frequently are not finished with bone growth. Therefore, dynamically fixed (wires) autologous bone grafts seem to be more reasonable than stable “nongrowing” constructs. Other possibilities can include temporary metal-bone fixation with hardware removal at a later date. Selective arterial or direct percutaneous

puncture embolization can precede the surgery to decrease perioperative blood loss. Adjuvant, focused radiotherapy can be added if radical resection is not achievable and/or in recurrence but its benefit is questionable. The recurrence rate is reported as high as 25%; however, this is probably not true recurrence but progression of residual neoplastic tissue growth left in place after nonradical surgery. Complete cure is reached if the resection zone is ­remineralized and without CT visible tumor at 4 years follow-up.

256

19  Tumors

primary bone tumors with 10% of them affecting the spine [38]. The cervical spine is rarely involved and in the UCS even less frequently. Usually, they originate from the vertebral body but gradually involve the posterior structures as well. Prognosis of GCT in spine is unclear with a high recurrence rate whatever treatment is used.

Diagnosis Diagnosis is often delayed as the initial symptom is often pain with neurological deficit occurring later. GCTs present as purely lytic lesions on plain films and the CT is non-specific showing differing areas of osteolysis and reactive tissue with new bone formation. MRI (Fig. 19.13), usually, does not add more diagnostic information and thus free-hand or CT-guided biopsy is performed. The tissue histologically reveals giant osteoclastic cells with spindle-shaped cells and regions of fibrous tissue [87]. Because of frequent reactive hypervascularization the bone scintigraphy is nearly always positive.

Treatment Strategy All treatment modalities have been used to treat GCTs. Despite some papers describing long-term success of radiation therapy alone [33] or embolization only [62],

Fig.  19.12  MRI of aneurysmal bone cyst (patient from Fig. 19.11) and its relationship to neural structures and external spread. (a) Axial image showing canal compromise. (b) Frontal plane T1 image

Giant Cell Tumors (GCT) GCTs are locally aggressive bone tumors originating from histiofibroblastic elements. Their behavior is unpredictable, transforming to malignant forms in 10% and occasionally having pulmonary metastases [48, 76]. Contrary to other primary bone tumors, they occur more frequently in females with a peak incidence in the third decade. They comprise 4–8% of

Fig. 19.13  MRI of recurrent giant cell tumor (Courtesy of Dr Fricˇ, Rikshospitalet HF, Oslo, Norway)

257

19.1  Extradural UCS Tumors

the combination of radical surgery, when feasible, and adjuvant therapy seem to be the most effective [59]. Radiotherapy can be recommended in partial resections.

L angerhans Cell Histiocytosis (LCH) – Eosinophilic Granulomas, Histiocytosis X This lesion, often called eosinophilic granuloma is a benign osteolytic bone process caused by reticuloendotheliosis of unknown origin. The first description is credited to Otani and Ehrlich and to Lichtenstein and Jaffe [78, 95]. It is more common in the first two decades of life with a male predilection. The incidence of LCH is approximately 1:1,500,000 inhabitants [93]. Spinal localization may be solitary or multiple and usually involves only the vertebral body. Wilner and other authors conclude that in cases of spinal LCH the thoracic vertebrae are involved most often (54%), followed by the lumbar (35%), and cervical vertebrae (11%) [132]. Bertram et al. conducted a meta-analysis of their own patients and 53 other cases of cervical LCH reported in the literature till 2002. The UCS was involved in 54% of adults and in 40% of children. In both groups, the vertebral body was affected in the majority of cases [12].

Diagnosis Cervical pain, limitation of movement, and torticollis are the most frequent symptoms. Only the osteolytic lesion can be seen radiographically and often only biopsy can confirm the diagnosis.

Treatment Strategy Reported treatment of cervical LCH varies from simple immobilization, radiotherapy, and local steroid injections, to surgical removal with anterior spine reconstruction. The outcome was good in most of the published cases independently of the choice of treatment [12]. In general, surgery is indicated only in cases of body collapse with deformity and/or neurological deficit.

Other Benign Tumors and Tumor-Like Lesions Hemangiomas can rarely occur in the UCS region. Their treatment is dependent upon the symptoms and structural involvement. Therapy can vary from

observation to resection and/or filling the lesion with bone cement. Gorham disease is an extremely rare osteolytic process that has spontaneous resolution and has been reported in the UCS area also [76]. Osteochondroma is derived from hamartous tissue and is extremely rare in the UCS; however, if present and growing, can cause neurological compression or even death [76, 106]. Fibrous dysplasia is another hamartomatous condition, which leads to weakening of trabecular bone structure and occasionally to body collapse. It usually resolves spontaneously; however, when causing deformity, surgical correction is indicated.

19.1.4.2 Malignant Primary Bone Tumors Primary malignant bone tumors are more common in older age groups than benign lesions with the peak incidence occurring in the fourth–sixth decades of life. The cervical spine is involved in about 20% of all spine primary bone malignant tumors and approximately 25% of these occur in the UCS region, particularly in the C2 vertebra. Males are affected three times more frequently than females [20].

Diagnosis These tumors are diagnosed often late in the course of the disease and complete vertebral involvement and/or spreading to adjacent compartments is common. This is caused either by slow asymptomatic growth or, alternately, by extremely fast progression. The clinical complaints are nonspecific with neck pain and muscular spasms being most common and the tumor is often not discovered until the adjacent soft and/or neurovascular structures are involved. The patient’s status can dramatically worsen during rapid tumor growth with progressive neurologic deficit, swallowing difficulties in the case of anterior extra-compartmental extension, or intractable pain that forces the patient to support his head manually. Plain films often show only an osteolytic lesion (Fig. 19.14); however, if the tumor spreads anteriorly, pharyngeal (Fig.  19.5) and even tracheal dislocation or a neoplastic mass shadow located prevertebrally can be seen. Nonetheless, for diagnosis, prognostic, and surgical considerations, CT and MRI are essential. Frequently bone scintigraphy or, more commonly, PET are used to evaluate possible multiple

258

19  Tumors

management. Currently, CT-guided needle biopsy through the planned surgical incision is preferred to avoid tumor contamination consequent to the needle biopsy tract [80, 98]. The risk of track contamination is considered so high that if the isolated expansion seems to be easily en bloc resectable (e.g., posterior arch only), a direct surgical procedure (excisional biopsy) is preferred [116]. On the other hand, having tumor diagnosis is still considered crucial before any complex operative decision is made.

Classification, Grading, and Staging

Fig. 19.14  Plain laterogram showing C2 significant osteolytic lesion caused by chordoma

Fig. 19.15  Plain lateral film showing only a mild osteolysis of cranial C3 edge caused by chordoma (same patient in Figs. 19.16 and 19.17)

tumor sites (also, “whole” body CT or spine MRI). CTA or MRI can often demonstrate tumor relationship to the VA, show neovascularization and can help plan for potential preoperative tumor embolization or VA trapping. Percutaneous needle or incisional biopsy to distinguish tumor type and to histologically grade the tumor is usually the first step in operative

As was seen in the benign tumors, the sectorial WBB surgical staging system is also applied to malignant primary tumor staging and helps in planning of operative options. According to Boriani’s original recommendation, tumors spreading within sectors 4–8 and 5–9 can be radically resected by vertebrectomy (i.e., somatectomy); sectors 2–5 and 7–11 by sagittal vertebral resection; and sectors 10–3 by posterior arch removal. Complete radical resection, however, can be achieved only within the limits described previously. The Enneking staging system has also been adopted for the purpose of staging of primary malignant spine tumors [21, 116]: SI

Low grade malignant primary bone tumors (e.g., chordoma, chondrosarcoma)

  Ia

Intraosseous, intravertebral

  Ib

Tumor invades perivertebral compartments having only pseudocapsule

S II

High grade malignant tumors (e.g., Ewing sarcoma, osteogenic sarcoma)

  II a

Intraosseous

  II b

Extraosseous spreading

S III

High grade malignant tumors having secondary focuses (metastases) at presentation

  III a

Intra-compartmental

  III b

Extra-compartmental invasion

Treatment The treatment is often complex and the ­decision-making process needs to be a result of multidisciplinary oncologic team discussion. The patient and their family

259

19.1  Extradural UCS Tumors

must be involved and their compliance ensured, especially if radical surgery with a high frequency of potential complications is the best option. It has been frequently reported that wide tumor resection can dramatically decrease the rate of local recurrence even to zero [35, 49, 69, 75, 104, 119]. Radical surgery is the ideal modality in majority of such tumors;however, the often late diagnosis and UCS location make it impossible in the vast majority of cases. Both, the anatomy of the UCS and associated vital structures, as well as the patient’s general medical status dictate the extent of the surgical procedure. The difficulty of radical resection procedures, as well as the major morbidity and considerable mortality, necessitates the question of whether they are justified [49, 50]. Neoadjuvant and/or adjuvant radiation and chemotherapy are often added, however, long-term survival is more an exception than the rule. For the definition of cure in primary malignant bone tumors, the standard should be considered as disease-free survival of more than 10  years [116]. The most common primary malignant tumors diagnosed in the UCS are chordoma, chondrosarcoma, Ewing sarcoma, and osteosarcoma. Malignant fibrous histiocytoma, synovial sarcoma, and rhabdomyosarcoma are rarities in this region. Also, occasionally found in this region are solitary myeloma and nonHodgkin’s lymphoma.

cord represent a late manifestation, usually indicating extra-compartmental intraspinal tumor growth. Plain films can demonstrate an osteolytic lesion with retropharyngeal space enlargement, or a deformity caused by vertebral destruction (Fig.  19.14), or can be nearly normal if the majority of tumor is located extra-compartmentally (Fig. 19.15). CT delineates the intravertebral osteolytic lesion with frequent extracompartmental tumor expansion (Fig. 19.16). The soft tissue relationships can be easily documented by MRI (Fig. 19.17). The tumor commonly extends anterolaterally from the vertebral body, but the epidural space can also be affected. The VA is often encapsulated by tumor tissue necessitating vertebral CTA, although the tumor itself is often not very vascularized. Correct WBB and Enneking staging done by oncologic team can guide the treatment strategy and estimate the feasibility of radical surgery.

Treatment Strategy Currently, it is understood, and in other parts of spine proven, that only radical surgical removal of the tumor can lead to long-term cure [17, 35, 103, 124]. Chemotherapy is ineffective in chordoma and the benefit of adjuvant radiotherapy is still the subject of

19.1.4.3 Chordoma Chordoma is a slow-growing malignant primary bone tumor derived from primitive notochord remnants and most commonly involves the clivus and the sacrum [15, 17, 41]. Less frequently, they occur in the cervical spine but if affecting the USC the second cervical vertebra is predominantly involved [8, 64, 72]. They are known to be locally invasive, recurrent, and resistant to chemotherapy. The majority of patients are men older than 50 years, although older females can be affected also. Their long term mortality is very high.

Diagnosis Most of the patients present late in the disease course suffering from non-specific neck pain, swallowing difficulties and occipital headache, and have been treated conservatively without radiographs. Neurological signs (quadriparesis) due to direct compression of spinal

Fig. 19.16  Axial CT scan of C3 chordoma expanding retropharyngeally with simultaneous invasion of extradural compartment

260

19  Tumors

involved bone, and thus decrease the rate of recurrence [65]. However, the benefits of such radical procedures must outweigh its risks. An extensive radical resection should be done only when reasonable survival (greater than 2 years) is expected [21, 103, 114].

Our Preference

Fig. 19.17  MRI in sagittal T1 sequence depicting anterior and intraspinal chordoma extend (same patient as on Figs. 19.15 and 19.16)

some debate [4, 115, 117]. Positive results showing better long-term local tumor control have been described only for high energy particle (proton) beam or combined proton–photon irradiation performed after tumor resection [22, 63, 86, 92]. Nevertheless, the extent of tumor removal and the amount of remnant invaded tissue play a substantial role in the effectiveness of any adjuvant radiotherapy. Surgical radical resection in the case of UCS chordoma is, in fact, feasible only in the very early stage of tumor growth; however, in cases where both anterior and posterior vertebral elements are involved (nearly always anatomically and always in oncologic terms), en-bloc resection (total spondylectomy) is the treatment of choice. This has been extensively described in other parts of the spinal column [2, 36, 52, 82, 124] but total spondylectomy of C2 is technically challenging given its anatomical relationship to the vertebral arteries, particularly when their preservation is desired [103, 108, 114] . The goal of such radical procedures is to not only remove the tumor, but also any potentially

Given our experience with early recurrence of chordoma after two-staged radical resection (Fig.  19.18), we feel that only maximally radical single-stage procedure prolongs the disease-free interval. Wide-margin resection is indicated in cases where there is a concern for potential dissemination of tumor cells in the surgical wound. In cases of malignant bone tumors, a total spondylectomy appears to be a better oncologic procedure with improved recurrence rates [2, 36, 108]. Upper cervical spondylectomies are complicated by the complex anatomy and the presence of vertebral arteries. Authors that have reported upper cervical spondylectomies, were usually forced to sacrifice the involved VA [8, 103] in order to achieve wide-margin resection of tumor and potentially involved soft tissues. We have described a case of total endolesional spondylectomy in a 64-year-old man whose chordoma was staged as Enneking I b and WBB A-D,F/ 3–8 [114]. CT showed the intraosseous extent of the tumor (Fig.  19.19); however, MRI documented extradural spreading and left VA involvement (Fig.  19.20). The VA balloon occlusion test was performed preoperatively and was positive with neurologic deficit lasting for 48 hours (Fig.  19.21). The patient was fully informed and decided to proceed with a recommended C2 spondylectomy through a combined transoral-posterior single-stage surgery. The procedure was performed beginning with transoral anterior segment removal and reconstruction (Fig.  19.22) followed by posterior C2 segment removal, sparing the left VA with concurrent middle column reconstruction and occipitocervical fusion (Fig. 19.23). He was irradiated conventionally afterwards and proton irradiation was not applied due to economic reasons. The status of this particular patient was good and on 1-year follow-up was without signs of recurrence on repeated MRIs (Fig. 19.24) and CTs; however, later in his course, a recurrence appeared on the left side. To date, he has had two reoperations with acceptable results and still remains ambulatory.

19.1  Extradural UCS Tumors

261

Fig. 19.18  The same patient (Figs.  19.15–19.17) after a twostage total endolesional spondylectomy with sparing of both VAs. (a) Three months after the surgery lateral plain radiogram without marked signs of recurrence. (b) Nine months later, the

same patient returning intubated with quadriparetic symptoms deemed inappropriate for any surgical intervention. MRI showing recurrence directly compressing spinal cord

Although the primary goal of the procedure is radical resection of chordoma, reconstruction of the created defect poses its own challenges. Previous case reports demonstrate that cervical constructs following spondylectomies are prone to failure [103]. A mesh cage secured to C1 arch, or its remainder anteriorly probably does not provide an adequate biomechanical support of the construct, especially if recurrence can be expected. Further, preservation of the atlas arch anteriorly would make a safe removal of the odontoid tip difficult. We believe that direct fixation to the clivus may represent a more robust alternative to previously described techniques. The use of a navigation system allows for safe placement of the cage and screws and also helps in determining the appropriate screw length. The angle of the clivus and the need for a perpendicular screw entry requires an angle that cannot be achieved through a transoral approach. This problem was solved using a surgically created, submandibular

channel through the floor of the mouth. Reconstruction of the middle column using cages placed between the articular surfaces increases the chance of bony fusion and also potentially offloads the strain on posterior occipitocervical fixation particularly in the case of potential tumor recurrence. To further increase the strength of the construct, all lateral mass screws were placed bicortically. Bony fusion was reinforced through autograft bone placed between occiput and each spinous process. Although recent literature demonstrates that a widemargin, en-bloc resection of chordomas involving the spine provides the patient with the best chance of disease-free survival [8, 35, 52, 103, 124], ours and other previously reported cases [15, 17] demonstrate that this may not always be possible in the cervical spine. We believe that, as the experience with resections for this rare condition widens, reconstructive techniques will be required to match those improvements.

262

19  Tumors

Fig. 19.19  Preoperative CT of the cervical spine demonstrates an osteolytic process – chordoma of the second cervical vertebra with involvement of both anterior and posterior elements. (a, b) Axial plane. (c) Sagittal plane image

19.1  Extradural UCS Tumors

Fig. 19.20  Preoperative MRI of the cervical spine confirms a chordoma (from Fig. 19.19) predominantly affecting the vertebral body of axis and left lateral transverse process. There is

263

epidural and extraverterbal disease displacing the left vertebral artery laterally. (a, b) Transverse plane. (c) Sagittal plane

264

Fig. 19.21  The left vertebral artery was draped over the lateral aspect of the tumor and a diagnostic angiogram demonstrated vessel patency. However, the patient did not tolerate a balloon

19  Tumors

occlusion test. (a) Artist’s illustration of VA tumor relationship. (b) Vertebral angiogram of left VA. (c) Angiogram obtained during balloon test occlusion (x2)

19.1  Extradural UCS Tumors

Fig. 19.22  Anterior reconstruction constituted of a Harms mesh cage filled with autograft and hydroxyapatite paste. The cage was screwed into the clivus cranially and C3 vertebral body caudally. (a) Artist’s illustration of cage position in antero-posterior

265

view. (b) CT reconstruction in midsagittal plane showing exact clival screw purchase (note the angle). (c) The right-angle to clivus achieved by submandibular introduction of the image navigated drill

266

Fig. 19.23  The final construct consisted of the anterior cage, two mesh cages placed between the lateral masses of C1 and C3 on the right and C4 on the left, and was completed by an occipitocervical fixation and fusion extending down to C6. C1 lateral

19  Tumors

mass screws passed through the two mesh cages supporting the lateral masses. (a) Artist’s illustration of the final construct – left side view. (b) Lateral slightly oblique plain film of final construct. (c) Transoral plain film of final construct

267

19.1  Extradural UCS Tumors

Fig. 19.24  MRI performed a year after the surgery showing patent VAs and no tumor recurrence. (a) Sagittal T2 sequence. (b) Coronal image

19.1.4.4 Chondrosarcoma Chondrosarcoma of the spine is rare (in particular, in the cervical region) with only case reports of UCS involvement [13, 42, 71]. Reported prevalence of chondrosarcoma in the spine is more than 6% [27]. Törmä, in his paper describing 250 histologically verified malignant spine tumors, found 11 cases of chondrosarcoma [126]. It is composed of cells having tendency to differentiate to chondrocytes, therefore foci of osteolysis can be mixed with nests of cartilage-like tissue and ossifications. Chondrosarcoma is a slow-growing tumor often located in posterior vertebral elements, spreading outside as extraosseous mass; however, circumferential spread is also possible. The histological grading is very important, as the low grade tumors (Gr I) are surgically controllable for long periods, whereas high grade tumors (Gr II+III) recur early [121]. Metastases are not frequent and if they occur, it is usually late with predilection for the lung. The

secondary malignant change of Paget’s disease, or osteochondroma to chondrosarcoma is also possible [105].

Diagnosis Clinically, symptoms are nonspecific until the neural structures are compromised. In posterior locations, suboccipital neuralgia can dominate. Radiologically, the mass usually appears to be extending out of vertebra and is irregular, lobulated with granular ossifications visible on CT or MRI.

Treatment Strategy Recurrences after incomplete resections are frequent and the patients die after many surgeries and multiple revisions (mean recurrence survival of 20% at 5 years) [27]. As adjuvant therapy is considered to be not very

268

effective, radical tumor removal respecting all the previously mentioned UCS anatomical restrictions is the only reasonable option [20, 112].

19.1.4.5 Ewing Sarcoma (ES) ES is a malignant, aggressive, poorly differentiated tumor that arises in the bone and soft tissues. ES is the second most common primary malignant bone tumor in children [129]. The spine is affected in about 3.5% of patients and UCS involvement is extremely rare. ES can invade the intervertebral disk space and multilevel spread is not exceptional. Metastatic spread is common.

Diagnosis Clinical symptoms can begin with neurological deficits related to spinal cord compression, and can appear early because of rapid tumor growth. Radiological diagnosis is based on detection of osteolytic lesions that do not respect the vertebral borders with invasion of intervertebral disks and often leading to vertebral collapse and deformity.

Treatment Strategy Traditional treatment for ES in long bones has been radical surgery (usually meaning amputation) in conjunction with chemotherapy and radiation. ES is known to be sensitive to chemotherapy. In the past, effective chemotherapy has been associated with an increase in the 5-year survival from 5% to 10%. Nowadays, modern chemotherapy regimen used prior to surgery (neoadjuvant chemotherapy) helped to increase the 5-year survival to 65%–70% [7, 90]. Despite that, a true radical resection is not possible in the region of UCS thus surgical tumor removal should be as radical as possible. According to the recent Sciubba et al. [110] systematic review, it can be concluded that neoadjuvant modern chemotherapy offers significant improvement in local tumor control and also long-term survival. Therefore, it follows that chemotherapy should precede surgical procedure. En bloc resection provides better local control, however, it does not increase the overall survival rate. Adjuvant radiotherapy is recommended after incomplete tumor resection, which in the UCS means almost always.

19  Tumors

19.1.4.6 Osteogenic Sarcoma (OS) OS is the most common type of malignant bone ­cancer, accounting for 35% of primary bone ­malignancies. It is rare in the spine (3.5% of all OS), most often detected in sacrum and described rarely in the UCS [108]. Shives and colleagues found 30 cases in the spine, only 4 of which were cervical [111]. Some cases of osteosarcoma of the spine are secondary, arising from Paget disease or after radiation t­ herapy [10]. The course of the disease is very rapid involving the whole vertebra and metastasizing in the majority of cases. Clinical and radiological appearance is similar to other sarcomas. The fast osteolytic process can rapidly lead to neurological compromise and spine deformity. The mortality is generally very high. Sciubba et al. [110] in their recent review analyzed six studies, and concluded that modern neoadjuvant chemotherapy (given before the surgery) can substantially improve the local control as well as the long-term survival. According to their conclusions, radical resection is also effective in terms of local control and survival rate.

19.1.4.7 Solitary Plasmocytoma There is still an ongoing debate whether plasmocytoma can exist as solitary lesion [85] as it is principally a systemic malignant neoplasm of bone marrow originating from plasmacellular differentiation of B-lymphocytes. Either solitary or multiple tumors can be found in the entire spine and not uncommonly in the cervical region including the UCS [54, 68, 99, 120]. The clinical picture is often non-specific similar to the other tumors with slow growth. CT and MRI will show a circumscribed lucent lesion visible mostly within the C2 body (Figs. 19.25 and 19.26). MRI is also a better tool than bone scintigraphy for exclusion of multiple spinal occurrences. The diagnosis is based on laboratory investigations confirming the presence of paraproteins in blood serum and histological analysis of the bone marrow smear. Also, a CT-guided or free-hand trocar biopsy (for C2 vertebra that means transoral) can provide tissue for histology. Although combined chemo and radiotherapy are the treatment of choice of this neoplasm there can be some specific situations requiring more aggressive approach. This is true in the case of vertebral collapse and deformity causing cord compression. In the case of axis involvement, transoral decompression and AA stabilization may

269

19.1  Extradural UCS Tumors

Fig. 19.25  Solitary plasmocytoma of C2 partially filled by transoral needle bone cement vertebroplasty. (a) Radiolucent area under the C2 lateral mass. (b) Partial bone cement filling supporting the lateral mass

be the right choice [99, 120]. However, wide-margin resection is never indicated. When the odontoid process is encompassed, or the lateral C2 mass pillars are substantially weakened, fortification with needle bone cement vertebroplasty is a reasonable approach. This can be performed either transorally (Fig. 19.25) or, and in our experience better, via the minimally invasive high anterolateral approach and vertebral puncture in odontoid screw-like fashion (Fig. 19.26).

19.1.5 Secondary Bone Tumors Metastatic disease is probably the most important problem for spine surgeons dealing with oncologic presentation in the spine. The problem is so broad and complex that cannot be shortly described in one chapter of a book focused on different topics and therefore, only a brief summary will be presented. The frequency of spine metastatic involvement is very high, occurring in about 70% of all patients suffering from cancer. The cervical spine is the least common site for metastatic spread with a reported incidence of between 8 and 20% of all spinal metastases [23, 100]. The average age range of patients diagnosed with cervical spine metastases is 58–61 years, without gender predominance [5]. If the UCS is involved, often the C2 body and/or arch are affected [97]. More rarely, C1 lateral masses or occipital condyles are affected. Different papers describe varying frequencies of appearance of many

tumor types that occur in the cervical spine without special attention to CVJ region [23, 37, 116]. According to our experience, the UCS most commonly features metastases of pulmonary (Fig. 19.27), renal (Fig. 19.28), thyroid gland, and breast carcinomas. We have seen metastases of melanoma (Fig. 19.29) and gynecologic tumors (Fig. 19.30). To date, we have not seen prostate or gastrointestinal tumor metastases to the UCS, but they are certainly possible.

19.1.5.1 Diagnosis Clinical symptoms may range from local and referred pain, mechanical pain of pathological fracture or instability to neurologic manifestations of nerve root and spinal cord compression [51]. Pain is the predominant symptom in most patients and is present in 90% of cases. This is particularly true in the UCS where pain can be provoked by any direction of head movement. Neurologic dysfunction is estimated to occur in 5–10% of patients with metastatic spine disease. Spinal cord compression with symptoms and signs of myelopathy is more common in the subaxial cervical area as opposed to the atlantoaxial region secondary to the differential size of the spinal canal at these levels [94]. Constitutional symptoms, weight loss, and anorexia may also be present. History of malignancy is always very suspicious and warrants further workup. As for the other tumor types, CT and MRI are essential for diagnosis and evaluation of surgical

270

Fig. 19.26  Solitary C2 plasmocytoma treated with bone cement needle vertebroplasty via high anterolateral mini-invasive approach. (a) Radiolucent areas in C2 body on coronal plane CT reconstruction. (b) MRI showing the intensity change of C2

19  Tumors

body and odontoid process. (c) Lateral plain film depicting bone cement filling of C2. (d) AP view documenting uniform cement spread

19.1  Extradural UCS Tumors

Fig. 19.27  Metastasis of pulmonary carcinoma destroying the C2 body treated with palliative occipitocervical fusion. The patient survived 7.5 months. (a) Preoperative CT 3D reconstruc-

271

tion. (b) MRI in sagittal plane. (c) Occipitocervical fusion skipping the C2 without any bone grafts

272

19  Tumors

Fig. 19.28  Metastasis of renal carcinoma to C3 treated with anterior corpectomy and cage and plate fixation. (a) Preoperative sagittal MRI. (b) Plain lateral film after the surgery

feasibility. In questionable cases, biopsy can help. Whole body workup has to be performed to exclude other tumor primary sites in a solitary UCS mass.

patients are appropriate for surgical intervention [122, 123, 125]. However, no single classification is specifically designed for the UCS and most are utilized to estimate patient survival time.

19.1.5.2 Classification, Grading, and Scoring 19.1.5.3 Therapeutic Strategy In an attempt to provide prognosis, to guide appropriate treatment for each patient, and also to allow for information interchange evaluating outcome, classification systems have been described. Harrington proposed a five-level classification scheme [58]. The Kostuik classification system attempts to identify which lesions will cause mechanical instability and are suitable for surgical intervention [73]. Raycroft and colleagues have proposed a specific classification system applied to the management of cervical metastatic tumors [102]. Tomita et al. and Tokuhashi et al. have suggested scoring systems that may assist in differentiating which

Different conservative treatment methods are available in cervical spinal metastases, including radiotherapy, hormonal therapy, chemotherapy, and high-dose steroid therapy. The indications for surgical intervention in upper cervical metastatic tumors include evidence of gross instability or neurologic compromise caused by malalignment or direct tumor compression. Also, solitary appearance of metastasis of radically treated primary tumor, which is feasible for surgical removal can represent an indication.

19.1  Extradural UCS Tumors

273

Fig. 19.29  Generalized melanoma in a very young man (27 years) with metastasis in right occipital condyle simultaneously with tumor destruction of C4 body treated with a combined approach. This unfortunate patient survived only 6  weeks after surgery.

Retrospectively this indication was more than questionable. (a) Coronal plane reconstruction. (b) Sagittal MRI view showing critical spinal cord imperil. (c) Lateral plain film depicting anterior cage and plate C3-5 fusion and posterior occipitocervical fixation

19.1.5.4 Our Preference

patient’s survival time with good quality of life by the multidisciplinary oncologic team. Good quality of life, in our terms, means ambulatory and painless:

For surgeons deciding whether or not to operate, the UCS represents a unique area of the spine. As we know from the thoracolumbar spine, simple laminectomy performed to decompress the neural structures is useful for pain relief and maintenance of ambulation. In the UCS, however, the progressing myelopathyrelated deficit will kill the patient while fully conscious. This is not acceptable for us in the majority of cases. On the other hand, if one cannot offer an ambulatory survival, the surgery ends up being only a technical exercise adding more stress to the unfortunate patient. About 80% of metastatic tumors are complex, which means that they are spreading extra-compartmentally. In those cases, it is illusory to think that any chance of radical resection is feasible. In those with intraosseous involvement only or purely epidural spread, radical resection can certainly be achieved. Surgical decision making is also imperative in the cases of solitary tumor appearance of unknown origin where diagnosis and treatment may be necessary. In summary, we propose that the most important factor in surgical decision process is the estimation of

• In case of expected survival less than 6 weeks, any surgical intervention makes no sense, in our opinion. • If the survival is estimated to be more than 6 weeks but less than 6 months, we prefer to decompress the cord and as simply as possible stabilize the spine (wires etc.). • If the expected survival is longer than 6 months but less than 1 year, then decompression is followed by spine stabilization (mostly, posterior modular) without bone grafting. • In cases of radically removable tumors and in those with life expectancy longer than 1 year, we attempt to achieve radical resection (if possible) and reconstruct the spine in the same manner as in non-oncologic patients. Certainly, this approach cannot be used in absolute terms but it is sensible to consider each individual patient and their specific situation. General medical status must be considered and attention paid to patient’s wishes and expectations.

274

Fig.  19.30  Gynecological tumor (myxoid leiomyosarcoma) metastasis to the posterior C2 elements was radically marginally resected and posterior occipitocervical fusion supplemented by autologous bone grafts allowed survival 1.5  years after the

19  Tumors

s­ urgery. (a) MRI showing the posterior tumor expansion. (b) Peroperative picture showing capsulated huge tumor. (c) Peroperative picture of OC fusion. (d) Laterogram of the fused cervical spine

19.2  Intradural Tumors (Extramedullary, Intramedullary)

275

19.1.6 Tumors of Spine Surrounding Connective Tissue

19.2 Intradural Tumors (Extramedullary, Intramedullary)

Tumors of all the tissue types surrounding the bony spine can occasionally be seen. Lipomas, chondromas, angiomas, and extra spinal neurinomas on the benign side but also sarcomas of muscles and synovial membrane theoretically also exist. A case of synovial sarcoma affecting the C2-3 joint is presented here (Fig. 19.31)

While this is not the primary focus of this book, intradural tumors of the UCS and CVJ area need to be briefly mentioned. The aim is not to comprehensively cover the principles of microsurgical removal but to inform the readers that these pathologies can present in the UCS region and that sometimes the approach used for their removal can damage the statodynamic system

Fig. 19.31  Synovial sarcoma of C2/3 joint removed by marginal resection via high anterolateral approach. (a) Axial MRI. (b) Frontal plane MRI. (c) Sagittal CT reconstruction. (d) Coronal MRI after tumor removal

276

of the spine and require reconstruction. Rarely, they can also mimic the above-mentioned pathologies and therefore, it is good to know their pathoanatomic appearance. Meningiomas, followed by nerve sheath tumors, are the most common primary spinal canal neoplasms [109]. In fact, meningiomas comprise 38–46% of foramen magnum tumors [55, 56]. Anterior meningiomas are, by definition, those attached to the foramen magnum on both sides of the midline (Fig. 19.32), lateral

19  Tumors

are defined as between the midline and the dentate ligament, and posterior tumors (Fig. 19.33) are attached posterior to the dentate ligament [55, 56]. Based on this definition, nerve-sheath tumors are always lateral (Fig. 19.34), although may exhibit posterior or anterior extension [56]. Tumors of the ventral foramen magnum represent formidable surgical lesions, as they may encase the vertebral, basilar, or their perforating arteries, cranial nerves and be densely adherent to the brainstem. Bony structures of the CVJ may also be invaded

Fig. 19.32  Anterolaterally located CVJ meningeoma. (a) Preoperative sagittal MRI. (b) MRI performed after tumor resection (note CSF pseudocyst)

Fig. 19.33  Retromedullary located CVJ meningeoma. (a) Sagittal MRI. (b) peroperative picture showing the released tumor before total extirpation

19.2  Intradural Tumors (Extramedullary, Intramedullary)

277

Fig. 19.34  Dumbbell neurinoma of C2 root. (a) Preoperative sagittal MRI. (b) Tumor extend on coronal MRI. (c) Axial MRI showing the relationship of the tumor to spinal cord. (d) Postoperative MRI showing total tumor removal

[107]. Radical tumor removal can thus be complicated by these tumor properties, or can be associated with a relatively high rate of complications. However, with careful preoperative planning, fine microsurgical techniques, proper use of skull-base approaches, intraoperative neurophysiological monitoring, and close postoperative management, the rate of complications can be decreased and patient survival prolonged.

Contrary to meningiomas, commonly encountered at the CVJ, this region is an unusual location for intramedullary tumors and presents specific surgical challenges. Only 43 patients with intramedullary tumors in this location were seen during a 10-year period in a highly specialized tertiary center [130]. The most common histological subtype is ependymoma and astrocytoma; other tumors are rare [130]. The goal

278

of intramedullary tumor surgery is gross total resection. However, this must not be achieved at the cost of neurological deterioration. Ependymoma (Fig. 19.35), contrary to infiltrative astrocytoma, is a surgical disease and patients after gross total resection may enjoy cure or at least a long recurrence-free survival (with the exception of rare high-grade lesions) [24, 46, 57,

Fig. 19.35  Huge ependymoma spreading from brain stem to C6 microsurgically removed with survival more than 10 years without recurrence. (a) Sagittal MRI showing the extent of tumor. (b) Postoperative MRI without marks of residual tumor on sagittal view. (c, d) Axial cuts depicting “banana skin” like cord appearance after tumor resection

19  Tumors

84]. On the other hand, astrocytoma (Fig. 19.36) can seldom be cured by surgery, recurrence is always a possibility, and thus surgical strategy must respect this distinction [11]. Other extra- and intramedullary lesions are rare (Fig. 19.37) and detailed discussion would be beyond the scope of this chapter.

19.2  Intradural Tumors (Extramedullary, Intramedullary)

279

Fig. 19.36  Inoperable infiltrative cystic spinal cord astrocytoma

Fig. 19.37  Intramedullary cavernous hemangioma unusually located behind C3 vertebra. (a) Preoperative MRI in T1 sequence. (b) Preoperative MRI in T2 sequence. (c) MRI performed 3 months after surgery without sign of residual lesion

280

References   1. Abdu, W.A., Provencher, M.: Primary bone and metastatic tumors of the cervical spine. Spine (Phila Pa 1976) 23, 2767–2777 (1998)   2. Abe, E., Sato, K., Tazawa, H., et al.: Total spondylectomy for primary tumor of the thoracolumbar spine. Spinal Cord 38, 146–152 (2000)   3. Ameli, N.O., Abbassioun, K., Saleh, H., et al.: Aneurysmal bone cysts of the spine. Report of 17 cases. J Neurosurg 63, 685–690 (1985)   4. Amendola, B.E., Amendola, M.A., Oliver, E., et  al.: Chordoma: role of radiation therapy. Radiology 158, ­839–843 (1986)   5. Atanasiu, J.P., Badatcheff, F., Pidhorz, L.: Metastatic lesions of the cervical spine. A retrospective analysis of 20 cases. Spine (Phila Pa 1976) 18, 1279–1284 (1993)   6. Azouz, E.M., Kozlowski, K., Marton, D., et  al.: Osteoid osteoma and osteoblastoma of the spine in children. Report of 22 cases with brief literature review. Pediatr Radiol 16, 25–31 (1986)   7. Bacci, G., Toni, A., Avella, M., et al.: Long-term results in 144 localized Ewing’s sarcoma patients treated with combined therapy. Cancer 63, 1477–1486 (1989)   8. Bailey, C.S., Fisher, C.G., Boyd, M.C., et al.: En bloc marginal excision of a multilevel cervical chordoma. Case report. J Neurosurg Spine 4, 409–414 (2006)   9. Barsa, P., Suchomel, P., Lukas, R., et  al.: Percutaneous CT-guided radiofrequency ablation in spinal osteoid osteoma treatment. Acta Chir Orthop Traumatol Cech 74, 401–405 (2007) 10. Barwick, K.W., Huvos, A.G., Smith, J.: Primary osteogenic sarcoma of the vertebral column: a clinicopathologic correlation of ten patients. Cancer 46, 595–604 (1980) 11. Benes 3rd, V., Barsa, P., Benes Jr., V., et al.: Prognostic factors in intramedullary astrocytomas: a literature review. Eur Spine J 18, 1397–1422 (2009) 12. Bertram, C., Madert, J., Eggers, C.: Eosinophilic granuloma of the cervical spine. Spine (Phila Pa 1976) 27, 1408–1413 (2002) 13. Blaylock, R.L., Kempe, L.G.: Chondrosarcoma of the cervical spine. Case report. J Neurosurg 44, 500–503 (1976) 14. Bohlman, H.H., Sachs, B.L., Carter, J.R., et  al.: Primary neoplasms of the cervical spine. Diagnosis and treatment of twenty-three patients. J Bone Joint Surg Am 68, 483–494 (1986) 15. Boriani, S., Bandiera, S., Biagini, R., et al.: Chordoma of the mobile spine: fifty years of experience. Spine (Phila Pa 1976) 31, 493–503 (2006) 16. Boriani, S., Biagini, R., De Iure, F., et  al.: Primary bone tumors of the spine: a survey of the evaluation and treatment at the Istituto Ortopedico Rizzoli. Orthopedics 18, 993–1000 (1995) 17. Boriani, S., Biagini, R., De Iure, F., et al.: En bloc resections of bone tumors of the thoracolumbar spine. A preliminary report on 29 patients. Spine (Phila Pa 1976) 21, 1927–1931 (1996) 18. Boriani, S., Capanna, R., Donati, D., et al.: Osteoblastoma of the spine. Clin Orthop Relat Res 278, 37–45 (1992)

19  Tumors 19. Boriani, S., De Iure, F., Campanacci, L., et al.: Aneurysmal bone cyst of the mobile spine: report on 41 cases. Spine (Phila Pa 1976) 26, 27–35 (2001) 20. Boriani, S., Nandiera, S., Weinstein, J.N.: Primary malignant tumors of the cervical spine. In: Clark, C.R., Benzel, E.C., Currier, B.L., et al. (eds.) The cervical spine, vol. 4, pp. 840–857. Lippincott, Philadelphia (2004) 21. Boriani, S., Weinstein, J.N., Biagini, R.: ) Primary bone tumors of the spine. Terminology and surgical staging. Spine (Phila Pa 1976) 22, 1036–1044 (1997) 22. Brada, M., Pijls-Johannesma, M., De Ruysscher, D.: Proton therapy in clinical practice: current clinical evidence. J Clin Oncol 25, 965–970 (2007) 23. Brihaye, J., Ectors, P., Lemort, M., et al.: The management of spinal epidural metastases. Adv Tech Stand Neurosurg 16, 121–176 (1988) 24. Brotchi, J., Dewitte, O., Levivier, M., et al.: A survey of 65 tumors within the spinal cord: surgical results and the importance of preoperative magnetic resonance imaging. Neurosurgery 29, 651–657 (1991) 25. Bruneau, M., Polivka, M., Cornelius, J.F., et al.: Progression of an osteoid osteoma to an osteoblastoma. Case report. J Neurosurg Spine 3, 238–241 (2005) 26. Burn, S.C., Ansorge, O., Zeller, R., et  al.: Management of osteoblastoma and osteoid osteoma of the spine in childhood. J Neurosurg Pediatr 4, 434–438 (2009) 27. Camins, M.B., Duncan, A.W., Smith, J., et  al.: Chondrosarcoma of the spine. Spine (Phila Pa 1976) 3, 202– 209 (1978) 28. Cantwell, C.P., O’Byrne, J., Eustace, S.: Radiofrequency ablation of osteoid osteoma with cooled probes and impedance-control energy delivery. AJR Am J Roentgenol 186, S244–S248 (2006) 29. Cantwell, C.P., Obyrne, J., Eustace, S.: Current trends in treatment of osteoid osteoma with an emphasis on radiofrequency ablation. Eur Radiol 14, 607–617 (2004) 30. Capanna, R., Albisinni, U., Picci, P., et al.: Aneurysmal bone cyst of the spine. J Bone Joint Surg Am 67, 527–531 (1985) 31. Caudell, J.J., Ballo, M.T., Zagars, G.K., et al.: Radiotherapy in the management of giant cell tumor of bone. Int J Radiat Oncol Biol Phys 57, 158–165 (2003) 32. Chabot, M., Herkowitz, H.N.: Spine tumors: Patient evaluation. In: Weisel, S. (ed.) Seminars in spine surgery, vol. 7, pp. 260–268. Saunders, Philadelphia (1995) 33. Chakravarti, A., Spiro, I.J., Hug, E.B., et  al.: Megavoltage radiation therapy for axial and inoperable giant-cell tumor of bone. J Bone Joint Surg Am 81, 1566–1573 (1999) 34. Chan, P., Boriani, S., Fourney, D.R., et al.: An assessment of the reliability of the Enneking and Weinstein-BorianiBiagini classifications for staging of primary spinal tumors by the Spine Oncology Study Group. Spine (Phila Pa 1976) 34, 384–391 (2009) 35. Choi, D., Melcher, R., Harms, J., et  al.: Outcome of 132 operations in 97 patients with chordomas of the craniocervical junction and upper cervical spine. Neurosurgery 66, 59–65 (2010). discussion 65 36. Cohen, Z.R., Fourney, D.R., Marco, R.A., et al.: Total cervical spondylectomy for primary osteogenic sarcoma. Case report and description of operative technique. J Neurosurg 97, 386–392 (2002)

References 37. Constans, J.P., de Divitiis, E., Donzelli, R., et  al.: Spinal metastases with neurological manifestations. Review of 600 cases. J Neurosurg 59, 111–118 (1983) 38. Dahlin, D.C., Cupps, R.E., Johnson Jr., E.W.: Giant-cell tumor: a study of 195 cases. Cancer 25, 1061–1070 (1970) 39. Dahlin, D.C., McLeod, R.A.: Aneurysmal bone cyst and other nonneoplastic conditions. Skeletal Radiol 8, 243–250 (1982) 40. de Kleuver, M., van der Heul, R.O., Veraart, B.E.: Aneurysmal bone cyst of the spine: 31 cases and the importance of the surgical approach. J Pediatr Orthop B 7, 286– 292 (1998) 41. Di Lorenzo, N., Delfini, R., Ciappetta, P., et  al.: Primary tumors of the cervical spine: surgical experience with 38 cases. Surg Neurol 38, 12–18 (1992) 42. Dominguez, C.J., Martin-Ferrer, S., Rimbau, J., et al.: Upper cervical chondrosarcoma. Neurocirugia (Astur) 16, 261–265 (2005). discussion 265 43. Dreghorn, C.R., Newman, R.J., Hardy, G.J., et al.: Primary tumors of the axial skeleton. Experience of the Leeds Regional Bone Tumor Registry. Spine (Phila Pa 1976) 15, 137–140 (1990) 44. Enneking, W.F.: A system of staging musculoskeletal neoplasms. Clin Orthop Relat Res 204, 9–24 (1986) 45. Enneking, W.F., Spanier, S.S., Goodman, M.A.: A system for the surgical staging of musculoskeletal sarcoma. Clin Orthop Relat Res 153, 106–120 (1980) 46. Epstein, F.J., Farmer, J.P., Freed, D.: Adult intramedullary spinal cord ependymomas: the result of surgery in 38 patients. J Neurosurg 79, 204–209 (1993) 47. Feigenberg, S.J., Marcus Jr., R.B., Zlotecki, R.A., et  al.: Megavoltage radiotherapy for aneurysmal bone cysts. Int J Radiat Oncol Biol Phys 49, 1243–1247 (2001) 48. Fidler, M.W.: Surgical treatment of giant cell tumours of the thoracic and lumbar spine: report of nine patients. Eur Spine J 10, 69–77 (2001) 49. Fisher, C.G., Keynan, O., Boyd, M.C., et  al.: The surgical management of primary tumors of the spine: initial results of an ongoing prospective cohort study. Spine (Phila Pa 1976) 30, 1899–1908 (2005) 50. Fourney, D.R., Rhines, L.D., Hentschel, S.J., et al.: En bloc resection of primary sacral tumors: classification of surgical approaches and outcome. J Neurosurg Spine 3, 111–122 (2005) 51. Frank, E., Chamberland, D., Ragel, B.: A proposed technique for intraoperative measurement of cervical spine stiffness. Neurosurgery 39, 147–150 (1996) 52. Fujita, T., Kawahara, N., Matsumoto, T., et al.: Chordoma in the cervical spine managed with en bloc excision. Spine (Phila Pa 1976) 24, 1848–1851 (1999) 53. Gabrielsen, T.O., Seeger, J.F.: Vertebral angiography in the diagnosis of intraspinal masses in upper cervical region. Neuroradiology 5, 7–12 (1973) 54. Gebes, S., Winking, M.: Plasmacytoma involving the atlas and axis. Neurochirurgia (Stuttg) 32, 187–188 (1989) 55. George, B., Lot, G., Boissonnet, H.: Meningioma of the foramen magnum: a series of 40 cases. Surg Neurol 47, 371– 379 (1997) 56. George, B., Lot, G., Velut, S., et al.: French language Society of Neurosurgery. 44th Annual Congress. Brussels, 8-12 June

281 1993. Tumors of the foramen magnum. Neurochirurgie 39(1), 1–89 (1993) 57. Guidetti, B., Mercuri, S., Vagnozzi, R.: Long-term results of the surgical treatment of 129 intramedullary spinal gliomas. J Neurosurg 54, 323–330 (1981) 58. Harrington, K.D.: The use of methylmethacrylate for vertebral-body replacement and anterior stabilization of pathological fracture-dislocations of the spine due to metastatic malignant disease. J Bone Joint Surg Am 63, 36–46 (1981) 59. Harrop, J.S., Schmidt, M.H., Boriani, S., et al.: Aggressive “benign” primary spine neoplasms: osteoblastoma, aneurysmal bone cyst, and giant cell tumor. Spine (Phila Pa 1976) 34, 39–47 (2009) 60. Hastings, D.E., Macnab, I., Lawson, V.: Neoplasms of the atlas and axis. Can J Surg 11, 290–296 (1968) 61. Hay, M.C., Paterson, D., Taylor, T.K.: Aneurysmal bone cysts of the spine. J Bone Joint Surg Br 60, 406–411 (1978) 62. Hosalkar, H.S., Jones, K.J., King, J.J., et al.: Serial arterial embolization for large sacral giant-cell tumors: mid- to longterm results. Spine (Phila Pa 1976) 32, 1107–1115 (2007) 63. Hug, E.B., Fitzek, M.M., Liebsch, N.J., et al.: Locally challenging osteo- and chondrogenic tumors of the axial skeleton: results of combined proton and photon radiation therapy using three-dimensional treatment planning. Int J Radiat Oncol Biol Phys 31, 467–476 (1995) 64. Hyun, S.J., Rhim, S.C., Riew, K.D.: A combined posterior, lateral, and anterior approach to ventrolaterally situated chordoma of the upper cervical spine. Surg Neurol 72, ­409–413 (2009). discussion 413 65. Jackson, R.P.: Recurrent osteoblastoma: a review. Clin Orthop Relat Res 131, 229–233 (1978) 66. Jackson, R.P., Reckling, F.W., Mants, F.A.: Osteoid osteoma and osteoblastoma. Similar histologic lesions with different natural histories. Clin Orthop Relat Res 128, 303–313 (1977) 67. Jaffe, H., Lichtenstein, I.: Solitary unicameral bone cyst with emphasis on the roentgen picture, the pathologic appearance and the pathogenesis. Arch Surg 44, 1004–1025 (1942) 68. Kaibara, T., Hurlbert, R.J., Sutherland, G.R.: Transoral resection of axial lesions augmented by intraoperative magnetic resonance imaging. Report of three cases. J Neurosurg 95, 239–242 (2001) 69. Kaiser, T.E., Pritchard, D.J., Unni, K.K.: Clinicopathologic study of sacrococcygeal chordoma. Cancer 53, 2574–2578 (1984) 70. Khan, D.C., Malhotra, S., Stevens, R.E., et al.: Radiotherapy for the treatment of giant cell tumor of the spine: a report of six cases and review of the literature. Cancer Invest 17, ­110–113 (1999) 71. Kingdom, T.T., Nockels, R.P., Kaplan, M.J.: Transoraltranspharyngeal approach to the craniocervical junction. Otolaryngol Head Neck Surg 113, 393–400 (1995) 72. Konya, D., Ozgen, S., Gercek, A., et  al.: Transmandibular approach for upper cervical pathologies: report of 2 cases and review of the literature. Turk Neurosurg 18, 271–275 (2008) 73. Kostuik, J., Weinstein, J.N.: Differential diagnosis and surgical treatment of metastatic spine tumors. In: Frymoyer, J. (ed.) The adult spine: principles and practice, pp. 861–888. Raven, New York (1991)

282 74. Laus, M., Albisinni, U., Alfonso, C., et al.: Osteoid osteoma of the cervical spine: surgical treatment or percutaneous radiofrequency coagulation? Eur Spine J 16, 2078–2082 (2007) 75. Lee, F.Y., Mankin, H.J., Fondren, G., et al.: Chondrosarcoma of bone: an assessment of outcome. J Bone Joint Surg Am 81, 326–338 (1999) 76. Levine, A.M., Boriani, S.: Benign tumors of cervical spine. In: Clark, C.R., Benzel, E.C., Currier, B.L., et al. (eds.) The cervical spine, vol. 4, pp. 816–839. Lippincott, Philadelphia (2005) 77. Levine, A.M., Boriani, S., Donati, D., et al.: Benign tumors of the cervical spine. Spine (Phila Pa 1976) 17, 399–406 (1992) 78. Lichtenstein, L., Jeffe, H.L.: Eosinophilic granuloma of bone: with report of a case. Am J Pathol 16(595–604), 593 (1940) 79. Lichtenstein, L., Sawyer, W.R.: Benign osteoblastoma. Further observations and report of twenty additional cases. J Bone Joint Surg Am 46, 755–765 (1964) 80. Lis, E., Bilsky, M.H., Pisinski, L., et  al.: Percutaneous CT-guided biopsy of osseous lesion of the spine in patients with known or suspected malignancy. AJNR Am J Neuroradiol 25, 1583–1588 (2004) 81. Lucas, D.R., Unni, K.K., McLeod, R.A., et al.: Osteoblastoma: clinicopathologic study of 306 cases. Hum Pathol 25, 117– 134 (1994) 82. Marmor, E., Rhines, L.D., Weinberg, J.S., et  al.: Total en bloc lumbar spondylectomy. Case report. J Neurosurg 95, 264–269 (2001) 83. Marsh, B.W., Bonfiglio, M., Brady, L.P., et  al.: Benign osteoblastoma: range of manifestations. J Bone Joint Surg Am 57, 1–9 (1975) 84. McCormick, P.C., Torres, R., Post, K.D., et al.: Intramedullary ependymoma of the spinal cord. J Neurosurg 72, 523–532 (1990) 85. McLain, R.F., Weinstein, J.N.: Solitary plasmacytomas of the spine: a review of 84 cases. J Spinal Disord 2, 69–74 (1989) 86. Munzenrider, J.E., Liebsch, N.J.: Proton therapy for tumors of the skull base. Strahlenther Onkol 175(Suppl 2), 57–63 (1999) 87. Murphey, M.D., Andrews, C.L., Flemming, D.J., et al.: From the archives of the AFIP. Primary tumors of the spine: radiologic pathologic correlation. Radiographics 16, 1131–1158 (1996) 88. Murphy, W.A., Strecker, E.B., Schoenecker, P.L.: Transcatheter embolisation therapy of an ischial aneurysmal bone cyst. J Bone Joint Surg Br 64, 166–168 (1982) 89. Nagashima, H., Nishi, T., Yamane, K., et  al.: Case report: osteoid osteoma of the C2 pedicle: surgical technique using a navigation system. Clin Orthop Relat Res 468, 283–288 (2010) 90. Nesbit Jr., M.E., Gehan, E.A., Burgert Jr., E.O., et  al.: Multimodal therapy for the management of primary, nonmetastatic Ewing’s sarcoma of bone: a long-term follow-up of the First Intergroup study. J Clin Oncol 8, 1664–1674 (1990) 91. Neumann, D., Dorn, U.: Osteoid osteoma of the dens axis. Eur Spine J 16(Suppl 3), 271–274 (2007)

19  Tumors 92. Noel, G., Feuvret, L., Calugaru, V., et al.: Chordomas of the base of the skull and upper cervical spine. One hundred patients irradiated by a 3D conformal technique combining photon and proton beams. Acta Oncol 44, 700–708 (2005) 93. Nyholm, K.: Eosinophilic xanthomatous granulomatosis and Letterer-Siwe’s disease. Acta Pathol Microbiol Scand Suppl 216, 211+ (1971) 94. Ono, K., Ebara, S., Fuji, T., et al.: Myelopathy hand. New clinical signs of cervical cord damage. J Bone Joint Surg Br 69, 215–219 (1987) 95. Otani, S., Ehrlich, J.C.: Solitary granuloma of bone: Simulating primary neoplasm. Am J Pathol 16(479–490), 477 (1940) 96. Peraud, A., Drake, J.M., Armstrong, D., et al.: Fatal ethibloc embolization of vertebrobasilar system following percutaneous injection into aneurysmal bone cyst of the second cervical vertebra. AJNR Am J Neuroradiol 25, 1116–1120 (2004) 97. Phillips, E., Levine, A.M.: Metastatic lesions of the upper cervical spine. Spine (Phila Pa 1976) 14, 1071–1077 (1989) 98. Pierot, L., Boulin, A.: Percutaneous biopsy of the thoracic and lumbar spine: transpedicular approach under fluoroscopic guidance. AJNR Am J Neuroradiol 20, 23–25 (1999) 99. Prasad, V.S., Raju, B.S., Sundaram, C.: Plasmacytoma of dens as a cause of atlanto-axial instability. Spinal Cord 36, 661–663 (1998) 100. Rao, S., Badani, K., Schildhauer, T., et  al.: Metastatic malignancy of the cervical spine. A nonoperative history. Spine (Phila Pa 1976) 17, 407–412 (1992) 101. Raskas, D.S., Graziano, G.P., Herzenberg, J.E., et  al.: Osteoid osteoma and osteoblastoma of the spine. J Spinal Disord 5, 204–211 (1992) 102. Raycroft, J.F., Hockman, R.P., Southwick, W.O.: Metastatic tumors involving the cervical vertebrae: surgical palliation. J Bone Joint Surg Am 60, 763–768 (1978) 103. Rhines, L.D., Fourney, D.R., Siadati, A , et al.: En bloc resection of multilevel cervical chordoma with C-2 involvement. Case report and description of operative technique. J Neurosurg Spine 2, 199–205 (2005) 104. Rich, T.A., Schiller, A., Suit, H.D., et  al.: Clinical and pathologic review of 48 cases of chordoma. Cancer 56, 182–187 (1985) 105. Robins, S.L.: Pathology, 3rd edn, pp. 1649–1650. Saunders, Philadelphia (1967) 106. Rose, E.F., Fekete, A.: Odontoid osteochondroma causing sudden death. Report of a case and review of the literature. Am J Clin Pathol 42, 606–609 (1964) 107. Samii, M., Klekamp, J., Carvalho, G.: Surgical results for meningiomas of the craniocervical junction. Neurosurgery 39, 1086–1094 (1996). discussion 1094–1085 108. Sar, C., Eralp, L.: Transoral resection and reconstruction for primary osteogenic sarcoma of the second cervical vertebra. Spine (Phila Pa 1976) 26, 1936–1941 (2001) 109. Schellinger, K.A., Propp, J.M., Villano, J.L., et  al.: Descriptive epidemiology of primary spinal cord tumors. J Neurooncol 87, 173–179 (2008) 110. Sciubba, D.M., Okuno, S.H., Dekutoski, M.B., et al.: Ewing and osteogenic sarcoma: evidence for multidisciplinary management. Spine (Phila Pa 1976) 34, 58–68 (2009)

References 111. Shives, T.C., Dahlin, D.C., Sim, F.H., et al.: Osteosarcoma of the spine. J Bone Joint Surg Am 68, 660–668 (1986) 112. Shives, T.C., McLeod, R.A., Unni, K.K., et  al.: Chondrosarcoma of the spine. J Bone Joint Surg Am 71, 1158–1165 (1989) 113. Simmons, E.D., Zheng, Y.: Vertebral tumors: surgical versus nonsurgical treatment. Clin Orthop Relat Res 443, 233–247 (2006) 114. Suchomel, P., Buchvald, P., Barsa, P., et  al.: Single-stage total C-2 intralesional spondylectomy for chordoma with three-column reconstruction. Technical note. J Neurosurg Spine 6, 611–618 (2007) 115. Suit, H.D., Goitein, M., Munzenrider, J., et  al.: Definitive radiation therapy for chordoma and chondrosarcoma of base of skull and cervical spine. J Neurosurg 56, 377–385 (1982) 116. Sundaresan, N., Boriani, S., Okuno, S.: State of the art management in spine oncology: a worldwide perspective on its evolution, current state, and future. Spine (Phila Pa 1976) 34, 7–20 (2009) 117. Sundaresan, N., Galicich, J.H., Chu, F.C., et  al.: Spinal chordomas. J Neurosurg 50, 312–319 (1979) 118. Suttner, N.J., Chandy, K.J., Kellerman, A.J.: Osteoid osteomas of the body of the cervical spine. Case report and review of the literature. Br J Neurosurg 16, 69–71 (2002) 119. Talac, R., Yaszemski, M.J., Currier, B.L., et al.: Relationship between surgical margins and local recurrence in sarcomas of the spine. Clin Orthop Relat Res 397, 127–132 (2002) 120. Thiry, S., Steenebruggen, A., Hotermans, J.M., et  al.: Complete destruction of the body of the axis by a myeloplasmacytoma. Resection by the transoral route and reconstruction of the vertebral body by a bone graft. Presentation of a case 2 years after intervention. Neurochirurgie 14, 799–808 (1968) 121. Thomson, A.D., Turner-Warwick, R.T.: Skeletal sarcomata and giant-cell tumour. J Bone Joint Surg Br 37-, 266–303 (1955) 122. Tokuhashi, Y., Matsuzaki, H., Oda, H., et al.: A revised scoring system for preoperative evaluation of metastatic spine tumor prognosis. Spine (Phila Pa 1976) 30, 2186–2191 (2005)

283 123. Tokuhashi, Y., Matsuzaki, H., Toriyama, S., et al.: Scoring system for the preoperative evaluation of metastatic spine tumor prognosis. Spine (Phila Pa 1976) 15, 1110–1113 (1990) 124. Tomita, K., Kawahara, N., Baba, H., et  al.: Total en bloc spondylectomy. A new surgical technique for primary malignant vertebral tumors. Spine (Phila Pa 1976) 22, ­324–333 (1997) 125. Tomita, K., Kawahara, N., Kobayashi, T., et  al.: Surgical strategy for spinal metastases. Spine (Phila Pa 1976) 26, 298–306 (2001) 126. Torma, T.: Malignant tumours of the spine and the spinal extradural space; a study based on 250 histologically verified cases. Acta Chir Scand Suppl 225, 1–176 (1957) 127. Verbiest, H.: Giant-cell tumours and aneurysmal bone cysts of the spine. With special reference to the problems related to the removal of a vertebral body. J Bone Joint Surg Br 47, 699–713 (1965) 128. Vergel De Dios, A.M., Bond, J.R., Shives, T.C., et  al.: Aneurysmal bone cyst. A clinicopathologic study of 238 cases. Cancer 69, 2921–2931 (1992) 129. Weber, K.L.: Current concepts in the treatment of Ewing’s sarcoma. Expert Rev Anticancer Ther 2, 687–694 (2002) 130. Weiner, H.L., Freed, D., Woo, H., et al.: Intra-axial tumors of the cervicomedullary junction: surgical results and longterm outcome. Pediatr Neurosurg 27, 12–18 (1997) 131. Weinstein, J.M.: Spine neoplasms. In: Weinstein, J.M. (ed.) The pediatric spine, pp. 887–916. Raven, New York (1994) 132. Wilner, D.: Radiology of bone tumors and allied disorders, vol. 2. Saunders, Philadelphia (1982) 133. Yamazaki, T., McLoughlin, G.S., Patel, S., et al.: Feasibility and safety of en bloc resection for primary spine tumors: a systematic review by the Spine Oncology Study Group. Spine (Phila Pa 1976) 34, 31–38 (2009) 134. Zileli, M., Cagli, S., Basdemir, G., et al.: Osteoid osteomas and osteoblastomas of the spine. Neurosurg Focus 15, E5 (2003)

Congenital and Developmental Abnormalities

20

P. Suchomel and O. Choutka

Developmental anomalies of CVJ were first discovered by anatomists and pathologists. Ackerman is credited for being the first to recognized the basilar impression in 1790 as cited by McGregor [26]. According to Gladstone and Erickson, the case of “occipital vertebrae” was described by Meckel in 1815 [13]. Rokitanski, a Czech pathologist working in Vienna, was probably the first one to described the basilar invagination as a developmental deformity in 1844, as cited by Ebenius [9]. Boogaard measured the angle of the clivus related to foramen magnum (FM) diameter on cadavers in 1865 [26]. In 1880, Grawitz contributed with anatomic description of six skulls with basilar impression [5]. Homén discovered, in 1901, the narrowing of FM by the odontoid and cephalic bulge of clivus as a possible cause of death, documented by a drawn picture [18]. The first diagnosis of basilar impression in a living patient documented by a radiograph was reported by Schüler in 1905 [31]. Basilar invagination is the most common abnormality (38–74%) of CVJ followed by atlas assimilation, atlantoaxial (AA) dislocation, Klippel-Feil syndrome, hindbrain herniation, and hydromyelia [8, 10, 14, 27]. In the majority of symptomatic and operated cases, a combination of various abnormalities is observed, however.

P. Suchomel Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova st. 10, 46063 Liberec, Czech Republic O. Choutka University of Cincinnati, Medical center, Department of Neurosurgery, Albert Sabin way 231, Cincinnati, OH 45267-0515, USA

20.1 Etiology A detailed description of embryological failures of CVJ is beyond the scope of this chapter. A significant ­number of CVJ deformities occur due to failure in development and correct connection of CVJ structures before the birth. As examples of these congenital anomalies, proatlas segmentation failure, basilar invagination, atlas assimilation, condylar hypoplasia, odontoid aplasia, hemivertebrae, and segmentation failures may be mentioned [27]. Anomalies developing during the period of unfinished bone growth postnatally are referred to as developmental, including os odontoideum, basilar impression, and syndromal abnormalities [27]. Both congenital and developmental anomalies are considered primary CVJ abnormalities [2, 26]. Failure in formation of UCS and CVJ during childhood can be caused by a variety of reasons including genetic or secondary factors. The distorted CVJ development can be a part of rare syndromes like osteogenesis imperfecta, skeletal dysplasia, Goldenhar’s syndrome, Conradi syndrome, Down syndrome, and spondyloepiphyseal dysplasia, with possible combination of pre- and postnatal developmental failures and later, decompensation caused by axial load [27]. The abnormalities described above, often, present in various combinations with neural anomalies. Chiari malformation and syringomyelia are frequently reported as a part of complex congenital or developmental CVJ deformity. Some deformities of CVJ can arise secondarily as a result of disease affecting the bone: hyperparathyroidism, osteomalacia, osteoporosis, Paget’s disease, rheumatoid arthritis (RA), trauma, and tumors [2, 30].

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_20, © Springer-Verlag Berlin Heidelberg 2011

285

286

20.2 Clinical Appearance As we deal only with adult patients in our department, failures in CVJ development presenting clinically in children will not be discussed in this chapter. There is a great variability in clinical presentation of developmental disorders of CVJ. Menezes stated that the most interesting feature of clinical presentation is its diversity [27]. Besides pain, most other symptoms are related to compression of neural structures, e.g., spinal cord, brainstem, cerebellum, and lower cranial nerves. Some clinical signs can be related to vascular compromise. Motor and/or sensory abnormalities and sphincter dysfunction related to myelopathy; nystagmus, ataxia, dysmetria caused by brainstem and cerebellar compression; dysphagia, speech irregularity, and tongue atrophy due to inferior cranial nerve compromise may be present. The clinical symptoms usually progress slowly, although acute deterioration is possible, particularly if exacerbated by trauma [14]. Starting with mild symptoms like vertigo, torticollis, and/or ataxia, the patients may finally suffer from serious quadrusymptomatology including respiratory disturbances. Risk of sudden death cannot be excluded [29]. Pain, common to other diseases of UCS, is present nearly in all cases, but is usually less dominant in adult patients as they have been living with their deformity for a long time. The patients with combined CVJ anomalies often have short neck, limited neck movement, and low hairline (typical of Klippel-Feil syndrome). Other developmental abnormalities such as cleft palate, face asymmetry, scoliosis, urinary tract abnormities etc. may also be present.

20.3 Radiology Since Schüler’s report in 1905 [31] describing an X-ray of a living patient with basilar invagination, it was necessary to develop a standardized algorithm to diagnose the vertical migration of UCS and other CVJ anomalies. Chamberlain [5], in his classical work, described a line drawn between hard palate and posterior FM margin (opisthion) on lateral X-rays and called it “the basal line.” He also suggested modifying the occipital X-rays to so called “fronto-vertex” projection in order

20  Congenital and Developmental Abnormalities

to see directly the FM and its deformity on radiograms. Chamberlain’s idea was practised in his four patients and confirmed by autopsy in two of them who later died, one following posterior surgical decompression. According to his original paper, the atlas and the whole axis should be located below the basal line in healthy individuals. Analyzing 100 plain lateral radiographsfrom normal adults, Saunders later added the numerical value to Chamberlain’s statement [30]. He reported that the tip of the odontoid process should be located below the Chamberlain’s basal line on lateral projections in mean distance 1.0 mm (SD = 3.6 mm). This value is still valid even for MRI and CT craniometry with the only difference in larger range of acceptable variability (3–7 mm) [2, 23, 34]. McGregor recognized that it is not always easy or even possible to correctly establish the posterior FM lip (opisthion) on lateral films and suggested to use the lowermost occipital skull point as another end of the line drawn from the hard palate [26]. He used 204 lateral films from mixed race South African population for his evaluation and after statistical analysis he stated that the odontoid tip should not lie more than 4.5 mm above the base line. He found 7 mm as the maximal acceptable normal value. Not frequently cited but a very important part of his work, however, is the measurement of the basal angle formed by intersection of clival line and anterior skull base line. He measured mean angle 134° (range 121°–148°), which was comparable to the values obtained by Brailsford, earlier [3]. McRae studied the CVJ anomalies with tomography (“laminagraphy”) and described 25 cases of atlas assimilation. He defined the line connecting the basion with opisthion as an important parameter when locating the FM. Many more parameters describing the normal and pathological radiological anatomy of CVJ derived from plain radiograms or tomograms were later developed. Only some of them are still used today in the era of CT and MRI, when used, however, their significance is similar [34]. Commonly used Wackenheim’s clivus line drawn along the posterior clivus should cross over the odontoid tip [37]. The angle formed by Wackenheim’s clivus line and line along posterior C2 wall called “clivuscanal angle” should not be less than 150° in flexion and is reaching up to 180° in extension. The abnormal flattening of the skull base (platybasia) is considered to be present when the basal angle (as described previously) exceeds 140°. The atlantooccipital (AO) joint angle in

287

20.7  Basioccipital Hypoplasia

coronal plane ranges between 124°–127°. The condylar hypoplasia may be expected in cases of more obtuse angle [34]. Also, other parameters originally used for atlanto-occipital dislocation (AOD) can be applied in CVJ craniometrics. As an example, the basion-dens interval (BDI) of Harris, which normally measures 7.4 mm (SD = 4.3 mm) can be distorted [17]. Cervicomedullary angle (CMA) is an angle between anterior borders of medulla oblongata and spinal cord at the level of FM as measured on MRI [4, 28]. This angle should be between 135°and 175° in normal subjects and measured less than 135° is considered as predictive of myelopathic progression [4, 24]. However, Abumi et  al. [1] obtained values from 50 healthy Japanese individuals and established the angle of 163° as a normal value (range 154°–179°). Nowadays, plain radiography serves only for primary screening of CVJ and UCS anomalies. MRI and CT are the mainstay methods, showing exactly the pathological anatomy in case of CVJ anomaly. Nevertheless, the more sophisticated modalities of contemporary imaging techniques are often requested by a surgeon in order to recognize the reducibility or instability of the deformity. Dynamic MRI directly shows the location and extent of neural compression [27] whereas dynamic CT exactly depicts the instability as well as the reducibility of bone dislocation during traction and/or flexion and extension [16]. MR- or CT-angiography contribute to identifying the frequently present abnormality in vascular supply [38]. If surgery is indicated, the 3D-CT modeling is useful for planning of trajectory of fixation screws and help to avoid an unintended surgical injury to neural or vascular structures.

20.4 Anomalies of the Occiput Abnormal development of occipital bone is often associated with decreased height of skull base and basilar invagination.

20.5 Condylus Tertius This anomaly is a result of failed integration of proatlas (the fourth occipital sclerotome) to occipital bone. The anteriorly located remnants may form the third condyle composed of one piece of bone in majority of cases

Fig. 20.1  Condylus tertius, partial posterior atlas assimilation, pseudoarthrosis of anterior C1 arch

(Fig. 20.1), although multiple ossicles can be present. It can create a joint or pseudojoint with odontoid or anterior C1 arch. Currently, CT or MRI techniques will display this abnormity, often presenting together with os odontoideum. If large enough, the condylus tertius can limit flexion movement in the AO joint.

20.6 Condylar Hypoplasia Insufficient development of one or both condyles, logically, decreases the odontoid distance from FM and basilar invagination is therefore often present, simultaneously. Also, condylar dysplasia is often combined with atlas assimilation. In the past, this deformity was diagnosed due to distortion of the AO joint angle on AP films. Today, CT scan in frontal plane will show exactly the shape and the size of condyles (Fig. 20.2). Consequently, the movement in AO joint is limited, resulting in restricted flexion and extension of the head.

20.7 Basioccipital Hypoplasia Dysplastic changes in basiocciput can be mild or severe, resulting in shortening of the clivus. Because of frequent odontoid invagination and/or synostosis, it is very difficult to reveal this anomaly from plain radiographs. MRI and CT (Fig.  20.3) will show it

288

20  Congenital and Developmental Abnormalities

Fig. 20.2  Condylar aplasia on the left side and hypoplasia with atlas assimilation on the right

Fig.  20.3  Basiocciput hypoplasia with fused odontoid process as shown by (a) CT in sagittal reconstruction and (b) MRI in T2-weighted images

exactly as well as its relation to neural structures. This combination of basiocciput bulge and odontoid invagination narrowing the FOM is probably similar to that drawn on picture by Homen in 1901 (Fig. 20.4).

20.8 Atlantooccipital Assimilation This is probably the second most frequent pathology of CVJ development, often accompanying other abnormalities. Atlas can be assimilated to the occiput partially or completely (Fig.  20.5). The AA joint is overloaded and regularly subluxated. The simultaneous synostosis (Klippel-Feil) of C2-3 is common.

Fig. 20.4  Bulging of the basiocciput to the FM as seen on 3D CT (similar to picture of Homén 1901)

289

20.11  Persistent Ossiculum Terminale

[19]. The isolated atlas arch disconnections, only rarely, are of some clinical importance. The pre-existence of arch defects, however, becomes much more important in the case of trauma and/or surgery scheduled because of other reasons. Under these conditions, arch defects may affect the UCS stability and without posterior atlas arch as an anchor, the surgery might be more challenging. It can, sometimes, be difficult to distinguish whether the arch cleft is present as a developmental failure or as a result of fracture.

20.10 Axis Anomalies

Fig.  20.5  Partial atlas assimilation depicted on 3D CT (the fused posterior arch on 2D CT sagittal reconstruction called “comma sign”)

20.9 Atlas Anomalies Apart from AO assimilation, the vast majority of C1 anomalies are various clefts and/or aplasias and hypoplasias of the arch. They are often detected incidentally without other concomitant anomalies, thus not affecting the stability. The posterior midline cleft is most frequently detected. This so called “posterior rachischisis” was observed in 4% of adult autopsy specimens [34]. Anterior clefts are rare and only a few cases of surgically treated combined anteroposterior atlas schisis (“split atlas”) (Fig.  20.6) were reported

Fig. 20.6  Split atlas on CT scan in axial plane

Except for complex fusion deformities, the anomalies of C2 are frequently isolated failures confined to odontoid process development.

20.11 Persistent Ossiculum Terminale Ossiculum terminale called Bergman’s ossicle results from a failure of fusion of apical ossification center of the odontoid process. It may mimic a rare type I odontoid fracture and as a rule, does not have any influence on UCS stability (Fig. 20.7).

Fig. 20.7  Ossiculum terminale (Bergman’s ossicle)

290

20.12 Odontoid Hypoplasia and Aplasia Dysplastic changes of the odontoid process are not frequent and can vary from hypoplasia (Fig. 20.8) to total aplasia (Fig. 20.9). Depending on the degree of hypoplasia and the relationship to transverse atlantal ligament, these changes can significantly affect the stability of AA complex.

20.13 Os Odontoideum Os odontoideum is defined as an ossicle with smooth circumferential cortical margins located at the place of odontoid process but without any osseous continuity with the C2 body. It was first described by Giacomini in 1886 [12].

Fig.  20.8  Odontoid process hypoplasia as a part of complex deformity on CT scan in coronal plane

Fig. 20.9  Odontoid aplasia as seen on transoral film

20  Congenital and Developmental Abnormalities

It continues to be debated whether the fragment disconnection occurs prenatally as a result of a developmental failure of the odontoid or is acquired due to perinatal or postnatal trauma. This anomaly is very often asymptomatic and os odontoideum is found incidentally during radiological investigation due to other reasons. If symptomatic, then a variety of clinical signs can be observed. The patients may suffer from suboccipital pain and headache only; some of them, nevertheless, present with transient or progressive signs of myelopathy. The plain X-ray in lateral and transoral projections is still a main diagnostic tool clearly showing the pathology. Os odontoideum can be radiologically classified into two anatomic variants. In the orthotopic type (Fig. 20.10), the ossicle moves together with anterior C1 arch. In the dystopic type (Fig. 20.11), it is fused to the basion and moves together with it, thus possibly subluxating anteriorly to the C1 arch. The anterior C1 arch may often appear hypertrophic and rounded (instead of “moon” appearance). Different shapes of odontoid ossicle were described and classified [25]. Associated bone anomalies like assimilation of atlas or C1 ring defects are observed quite often. Plain dynamic films of the cervical spine will reveal AA instability if present, e.g., if the PADI is less than 14 mm and SAC 13 mm, which are considered critical values for potential development of myelopathy. On the other hand, there is no evidence of correlation between the degree of instability and neurological status [32, 35, 40]. Advanced radiological techniques were not considered important for diagnosis in the past; availability of CT and MRI today, however, enables a better visualization of bone anatomy including possibly associated bone abnormalities (CT). MRI reveals a possible presence of fibrous pannus and its relationship to the spinal cord. In reality, only MRI can directly visualize the SAC in neutral and in critical positions, particularly if dynamic sequences are performed. The AA complex is stable on dynamic films in approximately 20% of patients with os odontoideum [11]. Majority of cases of AA instability due to presence of os odontoideum are represented by anterior C1 dislocations, although posterior instability was also reported [11, 32]. However, if os odontoideum is a part of more complex deformity, the AA subluxation can be irreducible and creates a significant

291

20.14  Our Preference Fig. 20.10  Orthotopic type of os odontoideum (fixed behind C1 anterior arch). (a) Plain lateral film. (b) Sagittal tomogram. (c) Sagittal MRI on sagittal plane of the same patient. (d) AA instability treated by posterior transarticular C1-2 screws ad modum Magerl

compression of neural structures either directly anteriorly or posteriorly by the subluxated posterior C1 arch. As a consequence of usually good clinical course without any progressive neurological deficit even in cases of unstable subluxation, as well as of unfavorable results from surgical interventions, most authors tended to prefer conservative approach in the past [7, 35]. On the other hand, deterioration in initially morphologically stable and clinically intact patients was also described [6, 22]. Advancement in imaging technologies and namely significant improvement of surgical fixation techniques has resulted in much better outcome after surgical AA stabilization. Klimo et al. [22] published the largest series of 78 patients treated with posterior transarticular fusion by Magerl’s technique with excellent results and 100% fusion rate. Based on logical analysis of AA biomechanics, they

stated that the risk of neural injury caused by possible trauma is unacceptable for patients with diagnosed os odontoideum and that all of them have to be offered a surgery.

20.14 Our Preference There is no doubt that the patients with symptomatic and unstable os odontoideum have to be treated surgically. We agree with Klimo et al. [22] that the risk of destabilization due to trauma in initially stable patients is high, particularly in population with active lifestyle where traffic and sport injuries are common. We, therefore, tend to offer these patients a surgery. We prefer the “wait and see” approach only in elderly patients with incidental finding

292

20  Congenital and Developmental Abnormalities

Fig. 20.11  Dystopic type of os odontoideum (connected to basion and moving with the head). (a) CT scan in sagittal plane showing the bony connection to basion. (b) MRI of the same patient. (c) CT scan in sagittal plane after transoral disconnection and resection of os odontoideum. (d) C1-2 fusion ad modum Goel-Harms sparing the intact C0-1 joint

of os odontoideum without instability and clinical symptoms. If the AA dislocation is reducible and without fibrous tissue pannus, we perform the posterior C1-2 fusion by technique of Magerl and/or Goel-Harms, supplemented by posterior bone graft in patients with orthotopic type of os odontoideum (Fig. 20.10). If the fibrous tissue surrounding the ossicle creates pressure on the spinal cord in neutral or reduced position as seen on MRI, then we perform the posterior fusion with transoral decompression. In the case of dystopic form, the odontoid ossicle connected to basion follows movement of the head. If posterior approach alone is selected, then the OC fusion must be performed, which means a sacrifice of movement in the OC joint with significant reduction of flexion-extension movement in UCS. Therefore, we prefer a combined approach with transoral disconnection of the ossicle from the basion and its removal, and posterior fusion of C1-2 only (Fig. 20.11).

20.15 Basilar Impression, Invagination The terms basilar invagination, basilar impression, cranial settling, and vertical odontoid migration have been used in patients where the UCS is located abnormally high. However, there is still some confusion in nomenclature of the deformity, due to mixture of description of different pathologies. Crockard [2] suggested to make the term ‘impression’ interchangeable with ‘invagination’ but to strictly distinguish whether the deformity is a result of a developmental failure (primary BI) or has occurred secondarily to disease affecting the bone quality in the CVJ (secondary BI) as was originally proposed by Saunders [30]. Menezes et  al. [27] recommended considering reducibility of the deformity prior to surgical decision making. Smith et  al. strictly differentiated basilar invagination as developmental anomaly from basilar impression as acquired deformity caused by secondary softening of the bone of skull base [33]. According to

20.15  Basilar Impression, Invagination

relationship of the odontoid process to FM, Goel [14] logically reclassified basilar impressions into two groups. In both groups, the odontoid tip is located above Chamberlain’s and McGregor’s lines. The principal difference was that in group A, the odontoid migrated inside the FM, thus becoming visible above McRae’s line as well as above Wackenheim’s clival line, whereas in group B, the whole UCS is migrated above the level of hard palate due to skull base deformity (basiocciput) without invagination of the odontoid inside the FM (Fig. 20.12). The group A was first reported as a fixed AA deformity, but later again Goel et al. recognized the possible vertical AA instability and potential reducibility of settled AA joint (often accompanied by atlas assimilation) allowing the reduction of the odontoid from the FM by simple extension [16]. The previous idea of Goel to classify only those with odontoid appearance inside FM as basilar invagination was principally accepted also by Kovero et  al. who exactly described all possible variants of CVJ anomalies accompanying the vertical UCS migration [23]. Patients with symptomatic neural compression are clearly the candidates for surgery. Also, those without symptoms but with documented morphological progression of instability should be considered for intervention. The etiology, reducibility, direction of the compression, and status of bone growth-age are other factors important for decision making regarding surgical tactic [27]. In reducible deformities, the traction and/or head positioning followed by posterior fixation can be a sufficient treatment [15]. If the deformity cannot be reduced, different techniques for release of neural structure and realignment of the CVJ were reported in the literature. Historically, the posterior decompression was performed regardless of the site of compression. The adverse outcome occurred in approximately 35–40% of patients [27]. Advancement in imaging facilities but mainly the dramatic evolution in surgical techniques and spinal implants led to more targeted approaches. The safety of fixation techniques has also improved. Today, the simple or extended direct anterior transoral decompression is followed by posterior OC fusion when anterior compression is dominant [27, 36]. If posterior fossa decompression is necessary (Chiari malformation), it may be performed separately. Another possibility of achieving anterior decompression is, to renew the shape of CVJ and thus indirectly decompress the spinal cord and the brainstem. Goel

293

et al. [14, 16] recommended to distract AA joint from posterior approach with intraarticularly interposed graft or cage and to fixate the joint with screw and plate. Abumi et  al. [1] suggested to use the posterior lever arm reduction of CVJ kyphosis by monoaxial screw

Fig.  20.12  Thirty-three year old patient with symptomatic basilar impression after two unsuccessful transoral decompressions. (a) Sagittal 3D CT reconstruction showing basilar impression and platybasia. (b) Sagittal 2D CT reconstruction depicting the atlas assimilation and status after partial odontoid resection. (c) Sagittal T2-weighted MRI clearly documenting the neural compression with sharp cervicomedullary angle and Chiari malformation

294

with U-shaped head tightened to the rod that is fixed to occiput. Wang et al. proposed the transoral release of AA joint without odontoidectomy prior to posterior reduction according to Abumi. Having treated 33 cases of irreducible atlantoaxial dislocations (IAAD), the authors stated that most of them can be converted to mobile and reducible state by this technique [39]. In theory, the anterior plating with relordotization can be used for this purpose as was documented in RA patients with vertical AA subluxation and kyphosis [20, 21]. Nevertheless, to our best knowledge, this method has been used in only one case of developmental IAAD so far [41].

20.16 Our Preference In summary, the experience with congenital CVJ anomalies is limited, particularly in countries where systematic prenatal screening is established, or where congenital anomalies are detected and treated early before they become decompensated and irreducible. A lack of continuous experience may result in a large variability of treatment modalities among different countries and surgeons. Also, our experience is limited by the small number of patients treated, almost all of them being adults. To make the nomenclature more transparent, we use the term “basilar invagination” when the odontoid occupies the FM (Goel’s group A), and the term “basilar impression” when the entire UCS is migrated upward with flattening of skull base but where the odontoid is not inside the FM (Goel’s group B). We use both these terms for developmental anomalies regardless of their time of occurrence and etiology. “Platybasia” and “vertical odontoid migration” are descriptive terms, only commenting the anatomical situation and/or its change in time. We use the term “cranial settling” in RA patients, usually meaning not only vertical migration of the odontoid process but also telescoping (downward dislocation) of atlas on axis. In fact, any intracranial position of the odontoid is possible only if structures holding the head weight

20  Congenital and Developmental Abnormalities

above C2 (including the lateral masses of C2) are missing, underdeveloped or destroyed by disease. We offer surgery to those patients with basilar anomalies who suffer from pain and/or neurological deficit provided there is an appropriate morphological background explaining their symptoms. Asymptomatic patients are treated in case of threatening or progressing compression of neural structures due to instability and/or deformity as detected by radiological investigation. In those cases where some degree of reducibility can be expected we start the treatment with traction for a few days and if favorable reduction is achieved and verified by MRI, then the posterior fusion is performed. If the reduction is not good enough or not achievable at all, then we prefer to decompress and stabilize. We perform the decompression always from the side of compression, which means that we most often start with anterior simple or extended (Fig. 20.13) transoral approach. The anterior AA joint release is attempted in order to make the deformity reducible. If extensive bone removal diminishes the load-bearing capacity of the anterior spine, we occasionally support the anterior column with mesh cage fixated to the clivus cranially and anchored to the first vertebral body enabling strong enough support to cage, caudally (usually, C3) (Fig. 20.14). The posterior OC fusion is then added as a single session surgery with or without posterior fossa decompression and C1 laminectomy, depending on the presence of Chiari malformation. When turning the patient to prone position, a halo-vest could be of advantage; however, even with a free-hand manipulation, monitoring of somatosensory and motoric evoked potentials (EP) should be mandatory. Turning the patient during the surgery is the most dangerous part of the procedure and without electrophysiological monitoring, a possible injury to the spinal cord would occur unnoticed. Angular reduction, if necessary for cervical spine alignment and sagittal balance, is performed manually with the head fixed in a halo ring and, again, under EP monitoring. Alternatively, posterior distraction or angular correction of the AA joint can be achieved by means of C2 pedicle screw levers or distraction forceps.

20.16  Our Preference

Fig. 20.13  The same patient as in Fig. 20.12, anterior transoromaxillar approach. (a) Extent of maxillotomy. (b) Both side Crockard deep retropharyngeal distractors. (c) Cosmetic result of extended maxillotomy

295

Fig. 20.14  The same patient as in Fig. 20.12, result of extended anterior decompression, anterior mesh cage support, posterior C1 laminectomy and posterior fossa decompression with OC instrumented fusion. (a) Plain laterogram. (b) MRI documenting the extent of decompression and change of cervicomedullary angle

296

References   1. Abumi, K., Takada, T., Shono, Y., et al.: Posterior occipitocervical reconstruction using cervical pedicle screws and platerod systems. Spine (Phila Pa 1976) 24, 1425–1434 (1999)   2. Bhangoo, R.S., Crockard, H.A.: Transmaxillary anterior decompressions in patients with severe basilar impression. Clin Orthop Relat Res 359, 115–125 (1999)   3. Brailsford, J.F.: The Radiology of Bones and Joints, 3rd ed., J&A Churchill (London), p: 257 (1945)   4. Bundschuh, C., Modic, M.T., Kearney, F., et al.: Rheumatoid arthritis of the cervical spine: surface-coil MR imaging. AJR Am J Roentgenol 151, 181–187 (1988)   5. Chamberlain, W.E.: Basilar impression (platybasia): A bizarre developmental anomaly of the occipital bone and upper cervical spine with striking and misleading neurologic manifestations. Yale J Biol Med 11, 487–496 (1939)   6. Clements, W.D., Mezue, W., Mathew, B.: Os odontoideum– congenital or acquired? – that’s not the question. Injury 26, 640–642 (1995)   7. Dai, L., Yuan, W., Ni, B., et al.: Os odontoideum: etiology, diagnosis, and management. Surg Neurol 53, 106–108 (2000). discussion 108–109   8. Di Lorenzo, N., Fortuna, A., Guidetti, B.: Craniovertebral junction malformations. Clinicoradiological findings, longterm results, and surgical indications in 63 cases. J Neurosurg 57, 603–608 (1982)   9. Ebenius, B.: The roentgen appearance in four cases of basilar impression. Acta Radiol 15, 652–656 (1934) 10. Erbengi, A., Oge, H.K.: Congenital malformations of the craniovertebral junction: classification and surgical treatment. Acta Neurochir (Wien) 127, 180–185 (1994) 11. Fielding, J.W., Hensinger, R.N., Hawkins, R.J.: Os Odontoideum. J Bone Joint Surg Am 62, 376–383 (1980) 12. Giacomini, C.: Sull esistenza dell os odontoideum nell uomo. Gior Accad Med Torino 49, 24–28 (1886) 13. Gladstone, J., Erickson-Powell, W.: Manifestation of occipital vertebra and fusion of atlas with occipital bone. J Anat Physiol 49, 190–199 (1914–1915) 14. Goel, A.: Treatment of basilar invagination by atlantoaxial joint distraction and direct lateral mass fixation. J Neurosurg Spine 1, 281–286 (2004) 15. Goel, A., Kulkarni, A.G.: Mobile and reducible atlantoaxial dislocation in presence of occipitalized atlas: report on treatment of eight cases by direct lateral mass plate and screw fixation. Spine (Phila Pa 1976) 29, 520–523 (2004) 16. Goel, A., Shah, A., Rajan, S.: Vertical mobile and reducible atlantoaxial dislocation. Clinical article. J Neurosurg Spine 11, 9–14 (2009) 17. Harris Jr., J.H., Carson, G.C., Wagner, L.K.: Radiologic diagnosis of traumatic occipitovertebral dissociation: 1. Normal occipitovertebral relationships on lateral radiographs of supine subjects. AJR Am J Roentgenol 162, 881–886 (1994) 18. Homén, E.A.: Deformationen der Schädelbasis und der basalen Schädelhyperostosen. Dtsch Z Nervenheilkd 20, 3–15 (1901) 19. Hu, Y., Ma, W., Xu, R.: Transoral osteosynthesis C1 as a function-preserving option in the treatment of bipartite atlas deformity: a case report. Spine (Phila Pa 1976) 34, 418–421 (2009)

20  Congenital and Developmental Abnormalities 20. Kandziora, F., Kerschbaumer, F., Starker, M., et  al.: Biomechanical assessment of transoral plate fixation for atlantoaxial instability. Spine (Phila Pa 1976) 25, 1555–1561 (2000) 21. Kerschbaumer, F., Kandziora, F., Klein, C., et al.: Transoral decompression, anterior plate fixation, and posterior wire fusion for irreducible atlantoaxial kyphosis in rheumatoid arthritis. Spine (Phila Pa 1976) 25, 2708–2715 (2000) 22. Klimo Jr., P., Kan, P., Rao, G., et al.: Os odontoideum: presentation, diagnosis, and treatment in a series of 78 patients. J Neurosurg Spine 9, 332–342 (2008) 23. Kovero, O., Pynnonen, S., Kuurila-Svahn, K., et  al.: Skull base abnormalities in osteogenesis imperfecta: a cephalometric evaluation of 54 patients and 108 control volunteers. J Neurosurg 105, 361–370 (2006) 24. Krauss, W.E., Bledsoe, J.M., Clarke, M.J., et al.: Rheumatoid arthritis of the craniovertebral junction. Neurosurgery 66, A83–A95 (2010) 25. Matsui, H., Imada, K., Tsuji, H.: Radiographic classification of Os odontoideum and its clinical significance. Spine (Phila Pa 1976) 22, 1706–1709 (1997) 26. McGregor, M.: The significance of certain measurements of the skull in the diagnosis of basilar impression. Br J Radiol 21, 171–181 (1948) 27. Menezes, A.H.: Craniocervical developmental anatomy and its implications. Childs Nerv Syst 24, 1109–1122 (2008) 28. Reijnierse, M., Bloem, J.L., Dijkmans, B.A., et al.: The cervical spine in rheumatoid arthritis: relationship between neurologic signs and morphology of MR imaging and radiographs. Skeletal Radiol 25, 113–118 (1996) 29. Rosomoff, H.L.: Occult respiratory and autonomic dysfunction in craniovertebral anomalies and upper cervical spinal disease. Spine (Phila Pa 1976) 11, 345–347 (1986) 30. Saunders, W.M.: Basilar Impression: the position of the normal odontoid. Radiology 41, 589–590 (1943) 31. Schüler, A.: Die Schädelbasis in Röntgenbilde. Fortschr.a.d. Geb.d.Röntgenstrahlen, Erg.Bd.11. Gräfe u. Sillem, Hamburg (1905) 32. Shirasaki, N., Okada, K., Oka, S., et  al.: Os odontoideum with posterior atlantoaxial instability. Spine (Phila Pa 1976) 16, 706–715 (1991) 33. Smith, J.S., Shaffrey, C.I., Abel, M.F., et  al.: Basilar Invagination. Neurosurgery 66, A39–A47 (2010) 34. Smoker, W.R.: Craniovertebral junction: normal anatomy, craniometry, and congenital anomalies. Radiographics 14, 255–277 (1994) 35. Spierings, E.L., Braakman, R.: The management of os odontoideum. Analysis of 37 cases. J Bone Joint Surg Br 64, 422–428 (1982) 36. Subin, B., Liu, J.F., Marshall, G.J., et al.: Transoral anterior decompression and fusion of chronic irreducible atlantoaxial dislocation with spinal cord compression. Spine (Phila Pa 1976) 20, 1233–1240 (1995) 37. Wackenheim, A.: Roentgen Diagnosis of the Craniovertebral Region, pp. 82–83. Springer, New York (1974) 38. Wang, S., Wang, C., Liu, Y., et  al.: Anomalous vertebral artery in craniovertebral junction with occipitalization of the atlas. Spine (Phila Pa 1976) 34, 2838–2842 (2009) 39. Wang, C., Yan, M., Zhou, H.T., et  al.: Open reduction of irreducible atlantoaxial dislocation by transoral anterior

References atlantoaxial release and posterior internal fixation. Spine (Phila Pa 1976) 31, 306–313 (2006) 40. Watanabe, M., Toyama, Y., Fujimura, Y.: Atlantoaxial instability in os odontoideum with myelopathy. Spine (Phila Pa 1976) 21, 1435–1439 (1996)

297 41. Yin, Q., Ai, F., Zhang, K., et al.: Irreducible anterior atlantoaxial dislocation: one-stage treatment with a transoral atlantoaxial reduction plate fixation and fusion. Report of 5 cases and review of the literature. Spine (Phila Pa 1976) 30, ­375–381 (2005)

Degenerative Disorders

21

P. Suchomel and P. Barsa

Anatomically, atlantoaxial (AA) articulation represents a three-joint complex: the atlantodental articulation and two lateral mass joints. All three may be involved in a degenerative osteoarthritic process that is commonly described as osteoarthrosis. Morphology of atlantoaxial osteoarthritis (AAOA) does not differ from degenerative changes found elsewhere in the spine. In clinical practice, it is commonly associated with peripheral osteoarthritis or degenerative spine disease, predominantly in older population. Osteoarthritis of the AA articulation, although mostly asymptomatic, may cause severe neck pain, occipital neuralgia, and even myelopathy. Progression of AAOA can lead to ligamentous laxity, instability, and rarely even to compression of the spinal cord. Degenerative changes of the facet joints occur with radiological prevalence of 4.8%. However, age-related progression has been shown. While it is documented in 5.4% of patients in their sixth decade, in ninth decade of life 18.2% of patients suffer from this problem [13]. The prevalence of atlantodental osteoarthritis has, so far, not been the subject of a systematic study.

21.1 History The first reference to a neuralgic syndrome in occipital region was made by Bruno y Lantijo and Ramosin in 1821 [9]. Since then, many authors identified cervical arthritis as being responsible for certain suboccipital

headaches [7, 10]. Ehni and Benner in 1984 were, however, the first to describe AAOA as a cause of pain in the occipital region and first who attempted to treat it surgically [2]. Ghanayem et  al. published the first large series of AAOA patients treated successfully with posterior AA fusion [4]. The largest published cohort of 35 patients treated for symptomatic AAOA using transarticular screw AA fusion is that of Grob et al. [5]. Finn et al. [3] presented a group of 26 patients suffering from AAOA treated either with posterior C1-2 transarticular fusion or with combined transoral and posterior fusion procedure [3].

21.2 Etiology Osteorthritic changes are of unknown etiology in many of the patients. A careful history, occasionally, points to a traumatic event years prior to presentation. Similar to other spinal locations, degeneration of AA joint corresponds to overload, osteophyte, and synovial cyst formation, may reflect segmental instability and joint complex laxity with subchondral sclerosis and may be a result of subchondral osseous healing due to repeated (micro) trauma. Weather-sensitivity of pain in CVJ osteoarthritis may be explained by inflammatory changes of soft tissue and involvement of the sympathetic nervous system in the process of pain generation.

21.3 Clinical Symptoms P. Suchomel and P. Barsa Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova st. 10, 46063 Liberec, Czech Republic

Pain is the principal symptom in patients with AAOA. Typically, it ascends unilaterally to the occiput, parietal

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_21, © Springer-Verlag Berlin Heidelberg 2011

299

300

region, and may radiate frontally into ipsilateral eye. Patients may report visual problems often leading to an ophthalmologic workout. Painful and sometimes audible crepitation during head rotation have indeed been described [12]. Pain quality is described as stabbing by patients. It is of mechanical characteristics in point of accentuation with axial rotation of the neck and in some patients, may lead to hand support during head movement or to prophylactic cervical collar wearing. Dreyfuss showed that an experimental increase in lateral AA joint pressure leads to similar patterns of pain as seen in AAOA and thus confirmed the role of lateral AA joint as a pain generator [1]. The pain syndrome may be weather-sensitive and low temperatures may contribute to joint stiffness as seen in other joints affected by arthritis. Although degenerative changes are common in both asymptomatic and symptomatic people, the clinical question is whether patient’s symptoms truly correspond to abnormalities seen on radiographic evaluation. Specific clinical symptom, which could be attributed to arthritic process in CVJ does not exist and laboratory studies (C-reactive protein, rheumatoid factor, ESR) together with clinical investigation

21  Degenerative Disorders

should be completed to rule out any systemic inflammatory disease as well as other arthropathies (psoriatic arthropathy, Reiters syndrome, enteropathic arthropathies). There is, however, no specific laboratory marker referring to AAOA. Atlantodental osteoarthritis may lead to degenerative pannus (Fig. 21.1) or synovial cyst (Fig. 21.2) formation, which may cause medial suboccipital pain and myelopathy due to spinal cord compression at the C1/C2 level [11].

21.4 Radiology The standard radiographic investigation including transoral and lateral view represents a starting point for evaluation of arthritic changes in CVJ. The open-mouth projection should provide an unobstructed view of the skull base, the odontoid, as well as C1-C2 articulation. Most often, the narrowing or obliteration of the AA joint with or without subchondral sclerosis can be seen on transoral films (Fig. 21.3a). Also, marked osteophytosis of the joint can support AAOA suspicion. To exclude

Fig. 21.1  Preodontoidal degenerative pannus in the patient with ankylosing spondylitis. (a) T2 MRI sequence. (b) The same patient T1 MRI sequence

21.4  Radiology

301

Fig. 21.2  Calcified synovial cyst 14  mm in diameter protruding from right C1-2 AA joint retropharyngeally. (a) T1 MRI. (b) Sagittal CT reconstruction. (c) Coronal CT reconstruction. (d) Axial CT in bone window

Fig. 21.3  Patient with left AAOA. (a) Transoral picture showing osteoarthritis with asymmetric left AA joint deformation. (b) CT scan in coronal plane of the same patient

302

potential AA instability, dynamic lateral radiographs have to be obtained. It is not unusual to be able to get only a limited information from such an evaluation due to a slightly oblique orientation of the facet joints and the overlap of facial bony ­structures in some. Coronal and/ or parasagittal reconstructions of the CT scan in bone window will therefore provide more precise information on the morphology of C1/C2 joint (Fig. 21.3b). A CT scan is not only helpful in diagnostics but inevitable if

21  Degenerative Disorders

any AA-stabilizing surgery is considered. Edema of ­surrounding osseous structures, possible synovial cysts, or inflammatory reaction and the precise anatomical ­picture of vertebromedullary relationship should be assessed by an MRI (Fig. 21.4). Diagnostic bone scan may also be added. However, presence of a “hot spot” is not a very specific finding indicative of active arthritic process and its presence implicates closer morphological ­evaluation. A typical morphologic picture of

Fig. 21.4  Another patient with AAOA. (a) Coronal CT showing left AAOA. (b) MRI of the same patient. (c, d) Dynamic lateral pictures showing AA instability caused by AAOA

21.5  Treatment Strategy

303

Fig.  21.5  Accidentally found degenerative intradental cyst in a patient investigated for odontoid fracture. (a) Coronal view. (b) Sagittal reconstruction

osteoarthritis located in lateral joints includes AA joint space narrowing with subchondral sclerosis. Both cartilage and bone may become eroded, producing cyst-like pockets in the bone, called “geodes.” Irregularities of margins of the facet can be demonstrated on coronal or sagittal computed tomography scan reconstructions. Subchondral edema, fibrosis or sclerosis may be delineated on MR images. Productive changes of the bone (osteophyte formation) are consistent with diagnosis of ­osteoarthritis but they are not very typical for arthritis in general. The atlantodental (central) osteoarthritis involves degenerative pannus formation eventually associated with synovial or juxtafacet cysts, which may arise from  the degenerative synovial lining of the joint. Intraodontal osseous cysts (Fig. 21.5) may also be a part of this morphological entity [3] if other etiology is not confirmed.

21.5 Treatment Strategy Methods of conservative treatment include nonsteroidal anti-inflammatory medications, gentle cranial traction, and external temporary immobilization. Although, the efficacy of such measures in reduction of AAOA pain has not been evaluated by prospective randomized

trials, they may, nonetheless, lead to pain reduction to a tolerable level. If these conservative therapies fail, the patient may be referred for steroid injection into the AA joint and a soft collar. Injection therapy may be very successful for a limited period of time [8]. Diagnostic facet block of C1-C2 to confirm the pain generator is recommended by some authors in preoperative decision making. Primary conservative treatment should be started in all AAOA patients and should continue as long as it is effective and symptoms are tolerable. Surgery is indicated when pain becomes intractable. Two methods of surgical treatment have been described in the literature. Ehni and Banner [2] performed C2 rhizotomy in three out of their seven patients. They were influenced by the fact that pain distribution resembled C2 dermatome and indeed their treatment results were described as good. Other, more commonly indicated surgical treatment is AA fusion. As described in other chapters, it may be achieved by various methods. The choice of a specific surgical fusion method is a subject of each individual patient’s anatomical situation and surgeon’s own experience and preference. Surgical outcome is favorable regardless of the type of fusion [4]. Fusion rate of posterior C1-C2 transarticular fixation in AAOA patients has been reported as high as 100% with marked pain improvement in 90% of the patients [3, 5, 12]. Transoral

304

decompression of CVJ supplemented with posterior stabilization is a method of treatment in patients who developed myelopathy due to compression caused by degenerative pannus formation. Similar to rheumatoid pannus, AA stabilization may be sufficient treatment of suboccipital pain in those patients who developed degenerative pannus without myelopathy signs [3, 6].

21.6 Our Preference When assessing patients with atypical headache, occipital one in particular, we should focus also on its concordance with head position and movement. In situations where head rotation or axial load reproduces occipital headache, transoral and UCS dynamic radiographs should be obtained with special attention

21  Degenerative Disorders

to the morphology of AA segment. If AA instability is suspected, further diagnostic workup as previously described is indicated. The majority of the patients referred to a spine surgeon have been treated conservatively for a long time. Nevertheless, the treatment is usually nonspecific to the AA joint. We therefore usually add at least local intra-articular injection therapy under direct CT guidance. This minimal intervention serves not only as a treatment modality but also as a diagnostic tool. In cases where conservative methods fail, surgical AA immobilization is offered. We prefer methods that provide immediate stability. Especially in AAOA patients, where ­mechanical pain is the leading symptom, high primary stability of the construct should be the goal as it will provide immediate pain relief for the patient. Most of our patients are treated by Magerl or Goel-Harms method of fusion (Fig. 21.6). The clinical success rate is impressive.

Fig. 21.6  Patient (Fig. 21.4) treated with AA fixation according to Goel-Harms. (a) Lateral picture. (b) AP projection

References

21.7 Practical Conclusion AAOA is an uncommonly diagnosed disorder of AA junction with suboccipital, typically unilateral pain as a leading symptom. Conservative therapy is the first line treatment as long as myelopathy or instability is not present. This may then be followed by facet block or immobilization. Finally, AA fusion represents an ultimate, highly effective therapeutic intervention in patients who fail nonsurgical treatment.

References   1. Dreyfuss, P., Michaelsen, M., Fletcher, D.: Atlanto-occipital and lateral atlanto-axial joint pain patterns. Spine (Phila Pa 1976) 19, 1125–1131 (1994)   2. Ehni, G., Benner, B.: Occipital neuralgia and C1-C2 arthrosis. N Engl J Med 310, 127 (1984)   3. Finn, M., Fassett, D.R., Apfelbaum, R.I.: Surgical treatment of nonrheumatoid atlantoaxial degenerative arthritis ­producing pain and myelopathy. Spine (Phila Pa 1976) 32, 3067–3073 (2007)

305   4. Ghanayem, A.J., Leventhal, M., Bohlman, H.H.: Osteoarthrosis of the atlanto-axial joints. Long-term follow-up after treatment with arthrodesis. J Bone Joint Surg Am 78, 1300–1307 (1996)   5. Grob, D., Bremerich, F.H., Dvorak, J., et al.: Transarticular screw fixation for osteoarthritis of the atlanto axial segment. Eur Spine J 15, 283–291 (2006)   6. Grob, D., Wursch, R., Grauer, W.: Atlantoaxial fusion and retrodental pannus in rheumatoid arthritis. Spine (Phila Pa 1976) 22, 1580–1583 (1997). discussion 1584   7. Horton, B.T., Macy Jr., D.: Treatment of headache. Med Clin North Am 30, 811–831 (1946)   8. Chevrot, A., Cermakova, E., Vallee, C., et al.: C1-2 arthrography. Skeletal Radiol 24, 425–429 (1995)   9. Perelson, H.N.: Occipital nerve tenderness: a sign of headache. South Med J 40, 653–656 (1947) 10. Pollock, L.J.: Head pain: differential diagnosis and treatment. Med Clin North Am 25, 3–13 (1941) 11. Sato, K., Senma, S., Abe, E.: Myelopathy resulting from the atlantodental hypertrophic osteoarthritis accompanying the dens hypertrophy. Two case reports. Spine (Phila Pa 1976) 21, 1467–1471 (1996) 12. Schaeren, S., Jeanneret, B.: Atlantoaxial osteoarthritis: case series and review of the literature. Eur Spine J 14, 501–506 (2005) 13. Zapletal, J., de Valois, J.C.: Radiologic prevalence of advanced lateral C1–C2 osteoarthritis. Spine (Phila Pa 1976) 22, 2511– 2513 (1997)

Surgical failures

22

P. Suchomel and O. Choutka

No field of medicine that involves any kind of intervention is without complications. Complications can be related to the patient, the procedure, anesthesia, or long-term follow-up; they can be immediate, early, or late. However, the full scope of complications related to a patient with UCS and CVJ pathology is beyond the scope of this chapter that focuses on some of the common and most dangerous, immediate intraoperative complications and their solution. Reconstructions of UCS and CVJ are no different from other fields of medicine and all surgeons must face a certain rate of complications related to their surgical procedure, more so during their learning curve [2, 6]. This eloquent area is certainly not very forgiving to complications that can be dangerous with surgical failure being rather “expensive” in the terms of possible further solutions. Any structural damage to neural tissue is irreversible and it, therefore, remains of utmost surgical importance to protect the spinal cord, medulla, and nerves during any reconstructive procedure at the CVJ. Vascular injury of large neck arteries can cause secondary brain ischemia with similar consequences, although some protection is offered by the multi-source brain and spinal cord vascular supply. Nevertheless, one must not rely on this “freedom” especially if not confirmed by angiography, preoperatively. Venous bleeding and/ or air embolism may appear insignificant but can also

P. Suchomel Department of Neurosurgery, Neurocenter, Regional Hospital Liberec, Husova st. 10, 46063 Liberec, Czech Republic O. Choutka University of Cincinnati, Medical center, Department of Neurosurgery, Albert Sabin way 231, Cincinnati, OH 45267-0515, USA

cause harmful complications, occasionally leading to abrupt, early termination of the procedure without execution of the planned stabilization. A misplaced screw with incorrect anchorage or trajectory may appear as a relatively benign error intraoperatively, especially if only bone or extravertebral tissue is involved. However, a poor anchorage can lead to a delayed instability endangering the previously mentioned essential structures [1, 4, 5]. Another problem of erroneous screw purchase is that it is very often not easy to find the alternative route for salvage screw because of limited bone stock available in the UCS area. An experienced surgeon must be able to recognize an intraoperative complication early and always be prepared to find a safe solution to the newly created problem. In our opinion, preoperative planning is one of the most important steps in complication avoidance. In CVJ, this entails a thorough scrutiny of the patient’s clinical situation and radiographic workup. A careful assessment of all bony and vascular anatomy and morphology is essential in planning of an UCS construct. With 3D computer modeling, the construct can be planned virtually (Chap. 7) but should also include alternative plans of salvage solutions to potential intraoperative events. For example, during atlantoaxial fixation and fusion, one has to consider that drilling, tapping, and placement of transarticular C1-2 screws may be unsuccessful and the surgeon should be prepared to alter the plan to a construct involving pedicle or laminar screws. Occasionally, a simple posterior wire and graft fixation can rescue a situation when no other construct is possible due to intraoperative complications. In general, we can divide the specific intraoperative UCS reconstruction complications to those related to approach, decompression, reduction, or hardware insertion.

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5_22, © Springer-Verlag Berlin Heidelberg 2011

307

308

22.1 Complications of Approach An incorrect position of the patient on the operating table can make the entire procedure complicated from the beginning. Either the target structures are impossible to reach or the required stabilization cannot be achieved. Improper positioning may also result in either excessive venous bleeding due to a dependent position of the surgical field or in air embolism if the field is too high relative to the heart. Neurosurgeons are particular about bloodless exposure of the spine. This aids in an easy identification of anatomical structures and thus avoids injury to the essential ones. Operating field covered in blood can substantially decrease the visibility of important structures and subject them to unnecessary risk. The final goal of the approach is to clearly expose the spine in an anatomical fashion. One must use all preoperative imaging to their advantage and identify any potential anatomical variants (see Chaps. 1 and 6). One key structure to identify and avoid during posterior approaches to the cervical spine is the vertebral artery (VA). Wanibuchi et  al. studied injected cadaveric heads and defined a simple threestep approach to identification of the V3 segment of the VA that was on average 19.1  mm lateral to the C1 tubercle [7]. The artery could be injured during a simple subperiostal exposure of C1 posterior arch in the case of its ponticular covering. An exceptionally rare, persistent first intersegmental artery could be injured during C2 isthmus exposure if not identified pre- or intra-operatively. Anomalous vessels are more common in syndromic patients, e.g., Down’s syndrome [8].

22.2 Complications of Direct Decompression Adequate decompression again requires an intimate knowledge of anatomical variants identified on preoperative imaging but also the relationship of vital structures to the compressive pathology. When addressing tumor resections at the CVJ, it is not only the relationship to neural structures that is important but also their vascular supply. It is not unusual for the last part of decompression to be the most delicate part requiring the utmost attention to careful removal of tumor in direct contact with neural

22  Surgical failures

tissue and doing so at the end of a long tumor resection. Such situation is a setup for a complication and the surgeon must be mentally ready to handle such challenges. Whenever the dura is opened and the arachnoid torn, CSF will escape the dural tube. This represents another unintentional complication. In the majority of cases, a watertight suture is not possible and dural substitutes with biological glue have to be applied. Many cases require temporary CSF diversion (e.g., lumbar drain) to avoid CSF fistulae. With CSF leak, the infection risk increases. Surgical tools, when used inappropriately, will result in complications. When performing bony decompression with high speed drills, we prefer to use diamond drill bits and operate at a high speed. Higher speed and shaving movement without pressure directed to dangerous tissue allows for better tactile feedback of bone remnants on dura or vessels. More recently, various bone ultrasonic aspirators that target only osseous structures without damage of soft tissues appear to be a promising idea [3]. Finding and respecting the natural cleavage planes with sparing of vessels not supplying the tumor is another important point in tumor surgery. We feel also that surgical microscope and electrophysiological monitoring have to be a part of the armamentarium whenever working in the UCS and CVJ area as the microscope aids in early identification of structures, and thus avoiding their unnecessary injury, while electrophysiological monitoring helps in early diagnosis of otherwise unforeseen events (e.g., during positioning).

22.3 Complications of Reduction (Indirect Decompression) In cases of indirect release of deformity by instrumented reduction, one has to be aware of the distance of the spinal cord from the segment being reduced and also the amount of safe free space. Other problems that may arise when attempting to achieve a correct alignment of the spine with good sagittal and coronal balance are both an overcorrection (Fig.  16.2, Chap. 16) and hypocorrection (Fig. 22.1). In the majority of cases, the realignment errors are minor without the need for corrective procedures; however, if postoperative deficit or painful syndrome exists directly related to the malalignment, a revision procedure should be

309

22.3  Complications of Reduction (Indirect Decompression)

a

b

c

Fig. 22.1  Reducible odontoid pseudarthrosis fixed in suboptimal reduction by transarticular C1-2 screws. (a) Flexion showing AA dislocation. (b) Complete reduction in extension. (c) Fixation according to Magerl in suboptimal position

undertaken. In ligamentous damage or injury, one has to be aware of possible overdistraction.

22.4 Complications of Hardware Insertion The most dangerous complications can occur during the reconstruction phase of the procedure. Sometimes, it can be difficult to find an appropriate screw trajectory for sufficient and safe bone anchorage. Especially, in brittle bone of osteoporotic patients or excessively hard bone of degenerated spine, alternative solutions may need to be found. In porotic spine, a bi- or quadri-cortical screw will minimize toggling and screw pull out. When the bone is too hard and does not allow the screw to pass easily through cancellous bone, direct drilling and tapping may be necessary. a

Fig. 22.2  Broken screws during tightening in odontoid type II fracture. Right one is cannulated titanium screw and left one 3.5 mm stainless steel screw. (a) AP radiogram showing healed fracture after 2 years. (b) Lateral film of the same patient

Such a situation may require a more complex reconstruction than was initially planned and alternative solutions must be prepared for. Complications related to low quality of implants are much less frequent nowadays, particularly in Western world. Nevertheless, we feel it is important to draw the reader’s notice to this potential problem as it may be encountered in many places due to economic reasons. In the past, we have encountered screw breakage during surgery of UCS. For example, strong final tightening of odontoid screws can lead to their breakage, especially if the tip is not drilled through and tapped (Figs. 22.2 and 22.3). Odontoid screw fixation is often a subject of surgical errors clearly related to a lack of sufficient experience (Fig.  22.4). Although not very frequent, an error during insertion of a transarticular C1-2 screw can lead to a VA injury, especially when an incorrect trajectory is selected (Fig. 22.5). Even when all anatomical landmarks are identified correctly and b

310

22  Surgical failures

Fig. 22.3  First screw broken during final tightening requiring other two screw introduction. (a) Peroperative fluoroscopical image of broken thread-shank transitional area of 3.5 mm stainless steel screw. (b) Introduction of the other two screws

a

b

a

b

c

d

Fig. 22.4  Patient referred to our hospital from another institution coming for regular check without any complaints. (a) Transoral radiogram showing too long odontoid single screw used for fixation of type II fracture. (b) CT in sagittal ­reconstruction depicting

the enjambment of the screw over the odontoid process apex (also pseudarthrosis was revealed). (c) Transoral picture obtained odontoid pseudarthrosis after screw removal. (d) Posterior transarticular fixation according to Magerl

311

References Fig. 22.5  Erroneous trajectory of second transarticular screw tapping leading to VA injury in RA patient with atlas settling. (a) Correct purchase of the first screw. (b) Incorrect too low C2 trajectory of the tap for second screw. (c) Point of VA injury (pulsating arterial blood came after tap removal). (d) Plugging with a short pars screw

a

b

c

d

an appropriate entry point is selected, attention has to be paid to the trajectory of any screw and prompt an early identification of vertebral foraminal breach. Ignoring the radiographic signs of incorrect trajectory then results in tapping of the VA, a point of no return in terms of avoiding the injury. In conclusion, many potential complications exist during reconstructions of the CVJ and UCS. Preparation and planning of intended procedure as well as bailout options are the best solution to complication avoidance. If a complication does occur, prompt identification is essential to avoid long-term sequelae. When appropriate, all tools available should be used to protect the patient from inadvertent injuries in the operating room (monitoring, fluoroscopy, and microscope). Complications do occur, even in the hands of an experienced surgeon. This is not an excuse for difficult procedures to be done by inexperienced centers. Obviously, if one does not perform CVJ reconstructions, they will never encounter a complication thereof. However, surgical morbidity needs to be minimized with experience, continued learning, and audit of one’s own

results. We strongly believe that centralization of care for patients with UCS and CVJ pathologies is the answer to reduction of complication rate and guaranteed continued education of new generations of CVJ surgeons.

References 1. Cao, Z.L., Ying, Q.S., Liu, J.F., et al.: The reason and prevention of upper cervical reoperations. Zhonghua Wai Ke Za Zhi 41, 567–569 (2003) 2. Finn, M.A., Apfelbaum, R.I.: Atlantoaxial transarticular screw fixation: update on technique and outcomes in 269 patients. Neurosurgery 66, A184–A192 (2010) 3. Ito, K., Ishizaka, S., Sasaki, T., et  al.: Safe and minimally invasive laminoplastic laminotomy using an ultrasonic bone curette for spinal surgery: technical note. Surg Neurol 72, 470–475 (2009). discussion 475 4. Rihn, J.A., Winegar, C.D., Donaldson 3rd, W.F., et  al.: Recurrent atlantoaxial instability due to fracture of the posterior C1 ring: a late finding following posterior C1-C2 fusion using the Halifax clamp. J Surg Orthop Adv 18, 45–50 (2009) 5. Rudzki, J.R., Lenke, L.G., Blanke, K., et al.: Pseudarthrosis of a thirty-nine-year-old dens fracture causing myelopathy. A case report. J Bone Joint Surg Am 86-A, 2509–2513 (2004)

312 6. Suchomel, P., Stulik, J., Klezl, Z., et al.: Transarticular fixation of C1-C2: a multicenter retrospective study. Acta Chir Orthop Traumatol Cech 71, 6–12 (2004) 7. Wanibuchi, M., Fukushima, T., Zenga, F., et  al.: Simple identification of the third segment of the extracranial vertebral artery by extreme lateral inferior transcondylar-transtu-

22  Surgical failures bercular exposure (ELITE). Acta Neurochir (Wien) 151, 1499–1503 (2009) 8. Yamazaki, M., Okawa, A., Hashimoto, M., et al.: Abnormal course of the vertebral artery at the craniovertebral junction in patients with Down syndrome visualized by three-dimensional CT angiography. Neuroradiology 50, 485–490 (2008)

Index

A AAD. See Atlantoaxial dislocation AAOA. See Atlantoaxial osteoarthritis AARF. See Atlantoaxial rotatory fixation ALL. See Anterior longitudinal ligament Aneurymal bone cysts (ABCs) autologous bone grafts, 255 description, 253 diagnosis “ballooning out”, bone cortex, 254 coronal plane and sagittal reconstruction, 256 spinal canal contents, 256 progression, prevention, 254–255 recurrence rate, 255 spinal canal contents, 256 treatment en bloc resection, 254 radiation therapy, 254 Anterior longitudinal ligament (ALL) atlanto-occipital membrane, 10 degenerative disc disease, 189 Anterior spinal artery (ASA), 13 Anterior structural reconstruction techniques bridge fixation principle, 60 C0-C1-C2 segment, 60 CVJ, 59 rigid immobilization, 61 AOD. See Atlantooccipital dislocation Atlantoaxial dislocation (AAD) diagnosis, 216 distractive force, 217 etiology and epidemiology, 216 radiology, 216 treatment dislocations, 217 neurological compromise, 216 translational injuries, 217 Atlantoaxial osteoarthritis (AAOA) AA instability, 302 C1–C2 transarticular fixation, 303 description, 299, 305 marked osteophytosis, 300 primary conservative treatment, 303 principal symptom, 299–300 Atlantoaxial rotatory fixation (AARF)

defined, 30 dynamic imaging, 30 Atlantooccipital dislocation (AOD) clinical symptoms, 139 CVJ craniometrics, 287 diagnosis, 142–143 etiology, 139 intubation and prone positioning, 142 morphological classification, 142 MRI, 143 radiology, 139–141 treatment strategy, 141 vision and fluoroscopic guidance, 142 Atlas anchoring structure anterior C1 lateral mass screw anterior/posterior approach, 71–72 neurovascular structures, 72 posterior arch, 72–73 posterior lateral massa screw, 73 anterior arch, 5 C1–2 combination injuries, 157 classification, fractures devoid muscular and soft tissue, 152 intra-articular, 153 posterior arch, 151 radiological literature, 152 sagittal split, 154 therapeutic consequences, 151–152 transoral radiogram, 152 transverse ligament disruptions, 153 types, 151–153 clinical symptoms, 154 decision process, 157 description, 151 diagnosis, fractures atlantoaxial joint, 154 LTA, 154–155 TAL deficiency, 154–155 transoral radiographs, 154 etiology, fractures axial head compression, 153 “bursting” mechanism, 154 odontoid process, 154 transverse process fractures, 154

P. Suchomel and O. Choutka, Reconstruction of Upper Cervical Spine and Craniovertebral Junction, DOI: 10.1007/978-3-642-13158-5, © Springer-Verlag Berlin Heidelberg 2011

313

314 lateral masses, 5–6 lower articular surface, 6 posterior arch, 5 superior articular surface, 6 transverse foramen, 6–7 treatment, fractures AA posterior fixation, 159–160 active surgical approach, 156 bony fusion, 158 direct compressive osteosynthesis, 159, 161 gradual traction release, 157 gun shot, 159–160 Harms fixator, 159, 161 healing, bony, 155 neural structures, 158–159 non sagittal fracture, 158 potential risks, 156 screw methods, 156 wedge-shaped lateral masses, 157 Axis, anchoring structure C2 vertebra, 80 laminar C2 screws anatomical background, 93 atlantoaxial stabilization techniques, 92 cortex and cancellous intralaminar bone, 94 drawback, 100 free-hand technique, 93 lag screw tightening, 99–100 monocortical isthmic screws, 93 surgical technique, 93–94 transisthmic screw technique, 94 Wright’s method, 93 long pars interarticularis screw anatomical structures, 91 anterior tubercle, 86 atlantoaxial dislocation, 91 bailout technique, 89 bony canal, 90 C0–1 joint, 90 3D modeling, 90 drawbacks, 92 electrophysiological monitoring electrodes, 90 image-guided placement, 87 imaging techniques, 88–89 intra-articular bone fusion, 91 intraforaminal cortical breaches, 87 isthmic bone bridge, 86 joint capsule, atlantoaxial, 88 lateral fluoroscopy, 91 Magerl’s technique, 88 nomenclature, anatomical, 86 preclude screw placement, 85 preoperative CT planning, 87 space available, 89 spinal canal contents, 89 transverse foramen perforation, 86 VA-bone occupancy ratio, 88 venous bleeding, 92 odontoid process screw anatomical background, 94–95

Index Apfelbaum retractor, 98 double odontoid screw purchase, 99 fractures, 94 lateral C2 view, 98 oblique surgical canal spreading, 98–99 preoperative fluoroscopical testing, 97 rectangular fluoroscopes, 96–97 surgical technique, 95–96 pedicle screw axis ring fractures, 85 compressive osteosynthesis, 80 cortex penetration, 84–85 3D modeling, 81 entry and exit points, 84 free hand technique, 82 lateral radiogram, 82–83 neural anatomy and pathology, 83 placement algorithm, 81 standard technique, 81–82 structures, neural, 83–84 subaxial cervical spine, 80 virtual planning, 83 short C2 pars interarticularis screw advantage, 92 rod connection, 92 use, 92 B Basion-dental interval (BDI) and BAI, 28 Dublin method, 28–29 radiology, 139–140 “Rule of Twelve”, 28 Basion-posterior axial line interval (BAI) MRI, 141 normal values in adults and children, 28 parameters, 140 Benign primary bone tumors ABCs (see Aneurymal bone cysts) classification, grading and staging, 250 clinical symptoms, 251 fibrous dysplasia, 257 GCTs (see Giant cell tumors) Gorham disease, 257 hemangiomas, 257 LCH, 257 lesions, 250 osteoid osteomas and osteoblastomas C2 lamina, 252 difference, 251–252 intralesional excision, 252 male predominance, 251 radiation therapy, 253 radical resection, 253 symptomatic patient, 253 total spondylectomy, 254 vertebra involvement and structure, 252 “vertebra plana” appearance, 252 radiology, 251 treatment, 251

315

Index WBB surgical staging and Enneking status, 250–251 vertebral tumor, 250 Bergman’s ossicle. See Ossiculum terminale Biomechanical remarks atlantoaxial complex (C1–C2), 17 atlantoaxial joint anterior horizontal displacement, 20 rotation-limiting ability, 20 transverse ligament, 20–21 atlantooccipital joints (C0–C1), 17–18 CVJ and UCS axial load distribution clinical and morphological instability, 20–21 occipitoatlantal joint AO hypermobility, 20 basilar invagination, 20 BDI and BAI, 20 occipitoatlantoaxial complex (C0-C1-C2), 17 Burst fractures, axis body comminution, 204 external immobilization, 203 surgical fixation and fusion, 203, 205 unconscious patient, 204 C Cervicomedullary angle (CMA) brainstem and spinal cord compression, 27 description, 287 Chondrosarcoma description and diagnosis, 267 treatment, 267–268 Chordoma anterior reconstruction, 265 description, 259 diagnosis anterior and intraspinal, 260 axial CT scan, 259 plain films, 259 follow-up, 260, 267 middle column reconstruction and occipitocervical fusion, 260, 266 navigation system, 261 treatment chemotherapy, 259–260 long-term cure, 259 surgical radical resection, 260 wide-margin resection, 260 Clivus basilar portion, 71 canal angle, 286 growth and correct formation, 5 sphenoid bone, 71 wedge-shaped, 71 CMA. See Cervicomedullary angle Combined atlas-axis fractures external bracing, 211 hangman’s with odontoid, 210 hard collar, 212 miscellaneous C2 and AA instability, 213

neurological deficit, 210 odontoid type III, hangman and Jefferson, 212 rotatory atlanto-axial subluxation, 213 Computer tomographic angiograms (CTA), 74 Condylar hypoplasia, 287 Condylus tertius, 287 Congenital and developmental abnormalities atlantooccipital assimilation partial atlas, 289 simultaneous synostosis, 288 atlas anomalies arch defects, 289 split atlas, 289 axis anomalies, 289 basilar impression, invagination acquired deformity, 292 groups, 293 IAAD, 294 posterior and anterior decompression, 293 reconstruction, 293 basioccipital hypoplasia bulging, 288 clivus shortening, 287 fused odontoid process, 288 clinical appearance, 286 condylar hypoplasia atlas assimilation, 287 condylus tertius, 287 “cranial settling”, 294 etiology distorted CVJ development, 285 unfinished bone growth, postnatal, 285 extended anterior decompression, 295 occiput anomalies, 287 odontoid hypoplasia and aplasia, 290 Os odontoideum anterior and posterior instability, 290–291 definition, 290 deterioration, 291 dystopic type, 290, 292 Magerl and Goel-Harms technique, 292 orthotopic type, 290, 291 surgery, 291 ossiculum terminale, persistent, 289 radiology cervicomedullary angle (CMA), 287 “the basal line”, 286 Wackenheim’s clivus line, 286–287 transoromaxillar approach, anterior, 294, 295 Coronal axis body fractures anterior graft and plate fusion, 201 partial involvement, posterior wall, 200 stable, 199, 200 transverse, 199, 200 vertical, 200 Craniovertebral junction (CVJ) anterior column reconstruction, 60 muscles, 11 atypical axial load distribution, 19

316 biomechanical properties, 55 clinical and morphological instability, 19–20 complex bony abnormalities, 23 components, 28 computer guidance, 130 craniometric parameters, 24 deformities, RA, 242 distorted development, 285 dorsal approaches, 3 dynamic imaging, 30 Halo ring, 141 ICA, 13 instrumentation, 134 intradural tumors (see Intradural tumors) iso-C navigational techniques, 129 ligaments, 9, 142–143 MRI evaluation, 142 osteoarthritis, 299 posterior techniques, 59–61 primary abnormalities (see Congenital and developmental abnormalities) radiographic evaluation, 33 reconstructions, 307, 311 soft tissue and bony dynamics, 24 structures, 30 transoral decompression, 304 trauma evaluation, 142 tumor resections, 308 unique morphology, 55 VA, 12 vascular evaluation, 31 vertebral artery compression, 32 visualization, 24 vital structures, 130 CVJ. See Craniovertebral junction D Degenerative disorders clinical symptoms degenerative pannus, 300 pain, 299–300 synovial cyst, 300–301 etiology, 299 Goel-Harms method, 304 history cervical arthritis, 299 transarticular fusion, 299 occipital headache, 304 radiology coronal/parasagittal reconstructions, 301, 302 degenerative intradental cyst, 303 “geodes”, 303 obliteration, AA joint, 301 vertebromedullary relationship, 302 treatment conservative, 303 surgical, 303–304 Dynamic reference array (DRA) anterior approaches, 130 C2 surface registration, 127

Index dorsal spine bony anatomic landmarks, 127 navigational system, 47–48 E Enneking staging malignant spine tumors, primary, 258 primary benign spine tumors, 250 EOP. See External occipital protuberance Eosinophilic granulomas. See Langerhans cell histiocytosis Ewing sarcoma (ES), 268 Extended transoral approaches anterior arch, 49 Crockard’s mouth distractor, 47–48 electrophysiology, 47 fluoroscopical visibility, 47 infiltrated mucosa, 48 maxillotomy, 50 microsurgical transoral odontoidectomy, 48 odontoid pseudoarthrosis, 49 preoperative fluoroscopical testing, 47 radical extirpation, 50 stabilization procedure, 47 transmandibular, 46–47 transmaxillar, 46 uvula, 48 External occipital protuberance (EOP) bone thickness, 4, 66 screw placement, 67 subperiostal dissection, 67 Extradural UCS tumors primary bone benign, 250–257 chondrosarcoma, 267–268 chordoma, 259–267 ES, 268 malignant, 257–259 OS, 268 solitary plasmocytoma, 268–269 radiological remarks gadolinium, 248 narrowed vertebral artery, 249 oropharynx displacement, 249 plain radiographs, 247 stabilization procedures, 247, 248 secondary bone classification, grading and scoring, 272 diagnosis, 269–272 therapeutics, 272–273 spine, 275 surgical oncologic terms margin, 249 resection, 248, 250 therapeutics, 248 treatment teams, 247 F Foramen magnum (FM) clivus, 285 narrowing, 285

317

Index neural compromise, 235 odontoid distance, 287 palate and posterior, 286 Fractures, ring of axis. See Hangman type fractures G “Geodes”, 303 Giant cell tumors (GCTs), 256–257 Goel-Harms technique bone graft, 108 C1–2 posterior fixation, RA, 241 fusion rate, 108 joint distraction and placement, 108 vs. Magerl’s technique, 109 occipitocervical fusion, 108 Os odontoideum, 292 Gorham disease, 257 H Hangman type fractures borderline instability, 192 classification Effendi, 181–182 Francis, 182 Levine and Edwards, 182 Roy-Camille, 182 definition, 183 direct osteosynthesis, 189, 191 dislocation, extension, 189 etiology and epidemiology, 182 extension and flexion, 193 fracture line visibility, 189, 191 hanging punishments, 179 isolated, 189 Levine type, 180–181 “long drop”, 179–180 MRI, 188 non-displaced, C2 abruption, 190 radiology C2 ring to verberal body, 183, 184 distant fracture, arch, 183, 184 dynamic MRI, 186 Levine type II, 185–186 line invasion, 184 “slipping”, 185 superior facet joint involvement, 184–185 symmetric transisthmic pattern, 184 symptoms and signs, 182–183 terminological alternatives, 179 treatment anterior surgery, 187–188 anterolateral approach, 188 direct pars fixation, 188 halo immobilization, 186–187 halo-vest, 187 posterior and combined approach, 188 traction and bracing, 186 verbal shortcuts, evaluation, 181 High anterolateral approach oblique visibility, 43

submandibular gland, 42 surgical technique, 42 I IAAD. See Irreducible atlantoaxial dislocations Internal carotid artery (ICA) lumen, 13, 73 risk, 73 standard angiography, 83 UCS and CVJ, 13 Intradural tumors ependymoma and astrocytoma, 277–278 intramedullary cavernous hemangioma, 279 meningiomas, 276 removal, 277 statodynamic system, 275–276 UCS region, 275 Intraoperative electrophysiological monitoring (IOM) electrodes, 39 fiberoptic guidance, 45 UCS surgery, 47 Irreducible atlantoaxial dislocations (IAAD), 294 L Langerhans cell histiocytosis (LCH), 257 Lateral approaches C1-C2 transarticular fixation, 41–42 posterolateral, 40–41 Ligamental tubercle avulsion (LTA), 154–155 Ligaments and joints atlantoaxial lateral, 10 atlantodental joint, 10 atlanto-occipital, 10 UCS and CVJ ligamentous connections, 9–10 M Magerl’s technique atlantoaxial transarticular screws, 105 biomechanical stability, 104 Goel-Harms procedure, 106 graft-related problems, 106–107 Olerud modification, 107 pediatric population, 105 wire fusion, 104–105 Malignant primary bone tumors classification, grading and staging, 258 diagnosis clinical complaints, 257 mild osteolysis, 258 needle/incisional biopsy, 258 osteolytic lesion, 257, 258 treatment decision-making, 258–259 radical surgery, 259 Minimally invasive approaches endoscopic techniques, 50 image-guided techniques, 51 odontoid fixation, 51 transcervical route, 51 Miscellaneous C2 fractures

318 C1-2 combination, 206 clinical symptoms, 198 incidence and classification polytrauma patients, 197 sagittal, 197 skeletal traction, 198 radiology anterior C2-3 dislocation, 198 oblique fracture, 199 plain radiographs, 198 sagittal MRI, 199 treatment burst fractures, axis body, 203–205 conservative, 198 coronal axis body fractures, 199–201 non-hangman injuries, 205 sagittal axis body fractures, 201–202 superior facet area, 205–207 tear drop fractures, 205, 206 transverse axis body fractures, 202–203 transverse foramen, 206 Monosegmental fusion constructs anterior, fixation methods C1-2, 113–114 C2-1, 112–113 lateral angled skin incision, 114–115 drawbacks, 115 odontoid fractures, 114 spinal canal penetration, 115 posterior, fixation methods atlanto-occipital instabilities, 101 C0-1, 101 C1-2, 102–112 Multiple fractures, axis combinations, 209 odontoid screw and C2-3 ACDF, 209, 211 “single” fractures, 209 teardrop and hangman’s combination, 209–210 Multisegmental constructs, CVJ and UCS anterior, 117 decision process, 115 occipitocervical anchoring elements, 116–117 biomechanical argument, 117 fibular strut graft, 115 modern modular systems, 117 plate and screw technique, 116 semi-rigid techniques, 116 subaxial lateral masses, 115 wire and frame application, 116 stabilization procedures, 115 suboccipital, 117 N Neural anatomy cervical spine nerves, 14 spinal cord anterior entry zones, 13

Index ascending and descending fibers, 13 craniocervical junction, 13 Non specific inflammation clinical symptoms, 228 differential diagnosis, 228, 231 external bracing, 231 incidence orofacial infection, 227 pyogenic osteomyelitis, 227 posterior fixation, 232 pyogenic inflammation, 232 radiology AA subluxation, 228 coronal plane reconstructions, 229 flexion and extension, 230 liquid abscess, 230 osseous destruction, 228 retropharyngeal pus spread, 229 spinal cord compression, 229, 230 treatment delayed diagnosis, 231 long-term antibiotics, 231 tuberculosis, UCS, 232 UCS needle aspiration, 231 O Occipital bone anchoring structure clivus, 71 condyles, 68–71 squama, 66–68 clivus, 5 condyles, 4–5 squama, 3–4 Occipital condyle fracture (OCF) bilateral condyle fractures, 148–149 classification, 145 clinical symptoms, 146 coronal reconstruction, 147 diagnosis, UCS, 145 etiology and epidemiology, 145–146 occipital condyle, 148 radiology broad condyle avulsion, 147–148 CT reconstruction, 146 diagnosis, 146 fracture morphology, 146–147 simple activity restriction, 147 treatment strategy, 147 types and subtypes, 145 Occipital condyles anterior transarticular axial-atlantocondylar screw, 70–71 cranial image guidance, 68 multipoint construct, 67–68 posterior transarticular atlantocondylar screw, 68–69 posterior transcondylar screw, 69–70 Occipital squama bicortical screw placement, 67 cervical polyaxial screws, 66 cranial anchor, 66 lateral placement, screws, 66

319

Index midline thickness, 67–68 radiological evaluation, 67 surgical technique, 67 “Occipital vertebrae”, description, 285 OCF. See Occipital condyle fracture Odontoid fracture (OF) AA instability, 175 anterior compressive osteosynthesis, 172–173 C3 coronal split, 175 classification Anderson D’Alonso, 166 direction based, 165 location based, 165 type IIT, 166 clinical symptoms, 167 disk prolapse, 170, 171 double-screw fixation, 172, 174 endosteal bone formation, 174–175 etiology and epidemiology, 166 fibrous union, 175 halo-ring traction, 173–174 halo-vest fixation, type II, 172 K-wires, 173, 174 MRI evaluation, 170 potential instability, 170, 171 radiology algorithm, 167 lateral tear, 168 non dislocated, 167 posteriorly displaced, type II fracture, 167 type II fracture, 168 transverse ligament damage, 170, 172 treatment anterior screw osteosynthesis, 170 Gallie technique, 169 halo-vest fixation, 169 horizontal fracture, 169 one/two screw construct, 170 posterior atlantoaxial fixation, 170 type I and III fracture, 168 Odontoid pseudarthrosis. See also Posttraumatic deformity spinal cord compression, 221 stable, 225 OF. See Odontoid fracture Os odontoideum anterior and posterior instability, 290–291 definition, 290 deterioration, 291 dystopic type, 290, 292 Magerl and Goel-Harms technique, 292 orthotopic type, 290, 291 surgery, 291 Ossiculum terminale, 289 Osteogenic sarcoma (OS), 268 P Posterior C1-2 fixation methods crosslaminar screw and rod fixation advantages, 110

atlantoaxial fixation techniques, 110 disadvantages, 110 Wright’s technique, 110–111 Gallie method, 102 Goel-Harms technique, 108–109 halifax atlantoaxial interlaminar clamps subaxial cervical spine trauma, 103–104 intralaminar screws lateral mass fixation, 111 Magerl’s technique, 105, 107 Mixter and Osgood silk loop, 102 surgical immobilization, 103 wedge compression technique, 103 wiring techniques, 103 Posterior longitudinal ligament (PLL), 10 Posterior meningeal artery (PMA), 12 Posterior midline approach, 40–41 Posterior paramedian approach, 40 “Posterior rachischisis”, 289 Posterior spinal artery (PSA), 13 Posttraumatic deformity anterior double-screw osteosynthesis, 225 clinical symptoms injury type, 219 vascular compromise, 221 etiology Goel-Harms fixation, 220 instability, 219 Marphan’s syndrome, 220 hard collar fixation, 223 irreducible odontoid malunion, 225 neurological deficit, 223 odontoid pseudarthrosis fixed, stable and unstable types, 222 type II odontoid fracture, 222 radiology AA instability, 221 odontoid process, imaging, 221 spinal cord compression, 221 surgery, 222–223 transarticular screw and graft fusion, 224–225 treatment, 221–222 “Pott’s disease”, 232 R RA. See Rheumatoid arthritis Radiographic data analysis atlanto-axial parameters instability, 29 posterior atlantodental interval (PADI), 29 atlanto-occipital joints, 24 basal/clival parameters CT/MRI images, 24 plain lateral radiograph, 24–25 craniocervical parameters basilar invagination, 26 BDI, 28 clivus and the odontoid process, 26 CMA, 27 CVJ and UCS alignment, 28–29 odontoid protrudes, 27

320 platybasia/basilar impression, 26–27 vertical atlantoaxial index, 28 craniometric parameters, 25 CVJ and UCS, 24 “Real Time” Image Guided Surgery (rIGS) direct spinal CT guidance, 130 3D isofluoroscopy, 130 intraoperative MRI (iMRI), 130 intraoperative scan, 129 vascular/neural structures, 129–130 Reconstruction techniques, basic principles construct design anterior column, 59–61 CVJ, plate and screw, 56–59 metallic implants, 56 surgical spinal stabilization, 56 defect/instability/decompression, 55 fracture healing/bone fusion biological principles, 61 lag screw, 62 morphogenic enhancers, 61–62 posterior fusion techniques, 61 tissue differentiation, 61 Rheumatoid arthritis (RA) cervical spine disease, 240 clinical symptoms neural compression, 236–237 Ranawat classification, neurologic deficit, 236 definition, 235 etiology and UCS pathophysiology cranial settling, 235 genetic susceptibility, 235 “stepladder deformity”, 235, 236 high-riding VA, 241 history and incidence cervical spine, 235 mortality rate, 236 serum RF, 236 irreducible kyphotic deformity, 241, 242 odontoid migration/basilar invagination, 242 quadriparesis, 243 radiology AA dislocation, 237, 238 dynamic MRI, 239 flexion/extension, 237 high riding VA, 238 MRI, 238–239 vertical subluxation, 237 reducible AA dislocations, 241 rheumatoid pannus, 242 traction-irreducible cranial settling, 243–244 treatment strategy AADI and PADI length, 239 OC fusion, 240 plating systems, 240 reducible lesion, 240 surgical, 239 rIGS. See “Real Time” Image Guided Surgery S Sagittal axis body fractures axial loading, 201

Index C1–C2 fusion, 202 unilateral/oblique, 201, 202 union, 202 Secondary bone tumors classification, grading and scoring, 272 diagnosis biopsy, 272 pain, 269 generalized melanoma, 273 gynecologic, 274 palliative occipitocervical fusion, 271 patient’s survival time, 273 radical resection, 273 renal carcinoma, 272 spine metastatic involvement, 269 therapeutics, 272 Solitary plasmocytoma chemo and radiotherapy, combined, 268 high anterolateral approach and vertebral puncture, 269, 270 needle bone cement vertebroplasty, 269 Special radiology CVJ, 24 data analysis, parameters atlanto-axial, 29–30 basal/clival, 24–26 craniocervical, 26–29 dynamic imaging AARF, 30 atlantoaxial pathologies, 30 CVJ and UCS abnormalities, 30 MRI, 31 upper cervical instabilities, 31 neoplastic conditions, 33–34 prevertebral soft tissue swelling, 23 tomography, defined, 23 traumatic cases, 33 vascular imaging dynamic angiography, 32 invasive angiography, 31–32 posterior circulation strokes, 32 vertebrobasilar insufficiency, 32–33 “Stepladder deformity”, 235, 236 Surgical anatomy bony structures atlas, 5–7 axis, 7–9 occipital bone, 3–5 CVJ and UCS internal carotid artery (ICA), 13 muscles, 10–11 vertebral artery (VA), 12–13 ligaments and joints atlantoaxial lateral, 10 atlantodental, 10 atlanto-occipital, 10 odontoid apex, 10 UCS structures, 9 neural anatomy cervical spine nerves, 14 spinal cord, 13 Surgical approaches

321

Index C1–C2 transarticular fixation antero-lateral aspect, 41 Barbour’s technique, 41 drilling, 41–42 retro-SCM approach, 41 high anterolateral carotid artery, 42 digastric muscle, 42 drawbacks, 43 exposure and oblique approach, 43 subaxial cervical spine, 43 surgical technique, 42 posterior midline, 39–40 paramedian, 40 posterolateral, 40–41 transoral extended, 46–49 minimally invasive, 50–51 traspharyngeal, 43–46 Surgical failures atlantoaxial fixation and fusion, 307 CVJ reconstructions, 311 direct decompression anatomical variants, 308 tools, 308 hardware insertion brittle and hard bone, 309 broken screws, 309, 310 erroneous trajectory, 311 odontoid screw fixation, 309, 310 improper positioning, 308 indirect decompression dislocation, reduction and fixation, 309 realignment errors, 308 venous bleeding and air embolism, 307 T Tear drop fractures, 205, 206 Total intravenous anesthesia (TIVA), 141 Transoral approach extended, 46–50 minimally invasive, 50–51 traspharyngeal, 43–46 Transoral traspharyngeal approach anesthetics and adrenalin, 45 atlantal ligament, 44 C1–2 dislocations and inflammatory process, 43 lateral fluoroscopy, 46 odontoid process, 45 posterior fixation techniques, 44 prophylactic antibiotics, 45 rheumatoid pannus, 44 Transverse axis body fractures, 202–203 Transverse ligament (TAL) coincidental atlas fracture, 159–160 deficiency, 154–155 functional integrity, 152 MRI assessment, 153 traumatic intrasubstance disintegration, 156 Tumors extradural UCS

primary bone, 250–269 radiological remarks, 247–248 secondary bone, 269–274 spine, 275 surgical oncologic terms, 248–250 therapeutics, 248 intradural ependymoma and astrocytoma, 277–278 inoperable infiltrative cystic spinal cord, 279 intramedullary cavernous hemangioma, 279 meningiomas, 276 removal, 277 statodynamic system, 275–276 UCS region, 275 U UCS and CVJ, reconstruction techniques anchoring structure atlas, 71–79 axis, 79–101 occipital bone, 66–71 monosegmental fusion constructs anterior, 112–114 lateral, 114–115 posterior, 101–112 multisegmental constructs anterior, 117 occipitocervical, 115–117 suboccipital, 117 Upper cervical spine (UCS) AA subluxation, 240 abnormalities, 30 anterior decompressive procedures, 61 asymptomatic hyperextension, 47 atypical axial load distribution, 19 clinical and morphological instability, 19–20 combination injury, 212 computer guidance, 130 cranial settling, 240 delicate structures, 127 diagnosis, 145 extradural tumors (see Extradural UCS tumors) failure, formation, 285 fracture combinations, 210 instrumentation, 134 laminae, 24 ligamentous connections, 9 mechanical instability, 55 minimally invasive approaches, 50–51 motion characteristics, 18 multidirectional instability, 44 muscles, 11–12 needle aspiration approaches, 231 OCFs, 147 posterior approach, 44 posttraumatic deformity (see Posttraumatic deformity) pyogenic inflammation, 232 reconstructions, 307, 311 screw breakage, surgery, 309 “snaking”, 187 synovial joints, 10 transoral procedure, RA, 240

322 tuberculosis, 232 vascular anatomy, 12–13 visualization, 24 V Vertebral artery (VA) ASA, 13 AV redundancy, 12 C1-2 screw, 309 high-riding, RA, 237, 238, 241 PMA, 12 PSA, 13 segments, 12 tapping, 311 Virtual image-guided surgery (vIGS) atlantoaxial fixation, 134 C1-2

Index fixation, 132–133 posterior transarticular screw, 127 dorsal spine bony anatomic landmarks, 127 drill position, 128 fracture dislocations, 130–131 hangman’s fracture, 131 intact transverse ligament, 130–131 intraoperative imaging CT scanner, 129 2D fluoroscopy, 129 3D fluoroscopy, 129 MRI, 129 lag screw compressing, 132 lateral fluoroscopy, 134 navigational plan, 133–134 preoperative imaging, 128–129 vertebra and instruments, 127