21 Revision Surgery of the Posterior Occipitocervical Junction

10.1055/b-0034-81398

21 Revision Surgery of the Posterior Occipitocervical Junction

Kim, Youjeong, Lehman Jr., Ronald A., Kuhns, Craig A., Riew, K. Daniel

Occipitocervical arthrodesis can be done successfully, especially with the recent advent of internal fixation devices available for treatment of instability. Occasionally, the spine surgeon will be faced with having to revise one of these constructs. As with all surgeries requiring a revision surgery, the first steps are recognizing the complication and identifying the underlying reason for failure of the previous surgery.

Background

The occipitocervical junction is a complicated interaction between the cranium and the upper cervical spine. It comprises the atlanto-occipital and atlantoaxial joints and is the most mobile section of the cervical spine, stabilized only by capsuloligamentous attachments. As a result, it poses significant challenges for fusion options, and attempts to stabilize this region require a thorough understanding of the anatomy. This becomes even more imperative when assessing possible reasons for failure and performing revision surgery.

The occiput has several bony landmarks that are advantageous to identify. The external occipital protuberance (EOP) is a dense ridge of bone that marks the thickest portion of the cranium. It measures 11.5 to 15.0 mm thick in men, 10.0 to 12.0 mm in women. The superior nuchal line is a similar thickened ridge that runs laterally from the EOP. The inferior nuchal line runs parallel but inferior to the superior nuchal line. The transverse sinus lies at the superior nuchal line, and the confluence of the transverse sinus is at or above the EOP; caution must be taken to avoid placing screws that may disturb the intracranial contents there. The occiput also takes an acute curve anteriorly from the superior nuchal line to the for-amen magnum. The occiput interacts with the cervical spine through two articular condyles on either side of the foramen magnum. In conjunction with the ligamentous structures, the occipitoatlantal joints allow 13° of flexion and extension, 8° of lateral bending, and no rotation.1 The vertebral artery takes a sharp medial turn at the level of the C1 ring and courses 1.5 cm lateral to the midline over the posterior ring of C1. With lateral dissection, the vertebral artery is vulnerable to injury.

The atlas has broad articular processes that interface with the occipital condyles superiorly and the axis inferiorly. Another articular surface on the posterior aspect of the anterior arch articulates with the odontoid process of the axis as a synovial joint. The atlantoaxial articulation allows 47° of rotation and ∼10° of flexion and extension. The atlas has no body and is unique in its bony anatomy. The mean thickness of the posterior ring is 8 mm, and the cortical bone is thin.2 The transverse ligament holds the odontoid process against the anterior arch of the atlas and is the principal stabilizing structure for the atlantoaxial articulation.

When dissecting the posterior aspect of the occiput, atlas, and axis, injuries to associated neurovascular structures should be avoided. The greater occipital nerve is the medial branch of the dorsal ramus of the second cervical nerve that is mainly cutaneous to the back part of the scalp and in some patients if injured may cause bothersome neuralgia.

Common causes for occipitocervical instability requiring arthrodesis are trauma, inflammatory diseases, including rheumatoid arthritis, degenerative osteoarthritis, neoplasm, and congenital abnormality. Patients with atlantoaxial instability due to failed C1–C2 fusion and those with deficient or fractured arches of C1 and C2 may also require occipitocervical fusion.

Even without pathological changes, a significant amount of motion in the sagittal and rotatory planes occurs in this region; therefore, constructs used to provide stability must be able to resist these forces and withstand the lever arm that is created by the angle of the suboccipital bone and the cervical spine. Furthermore, these constructs must be able to accommodate adjustments in positioning the patient intraoperatively into an appropriately neutral occipitocervical position prior to final tightening and preparation for fusion. Consideration should also be given to postoperative care and the need for immobilization, depending on the rigidity of the construct.

Historical Techniques for Occipitocervical Fixation

Historically, internal fixation was inadequate to fix the spine rigidly, and external immobilization was often necessary to provide stability. With current instrumentation, rigid external immobilization is usually not necessary. Although many of the techniques described below are no longer used in the United States, we include a brief description of them because, when performing revision surgery in this region, readers may encounter them. Occasionally, one of these techniques may be necessary during revision surgery, if all of the available fixation points have been used or destroyed with previous operations.

Onlay Bone Graft without Internal Fixation

Onlay bone grafting was the first technique described for occipitocervical fusion. Stability until a bony fusion was achieved was with external fixation—cranial tongs with bed rest, a Minerva jacket, and/or a postoperative halo vest. Foerster first described this technique in 1927.3 A fibular graft was inserted between the occiput and C7 to stabilize a progressive atlantoaxial dislocation after an odontoid fracture. Graft choice and methods to fashion the graft have varied. In general, autograft has been favored, but allograft has also been used. Techniques have varied from wedges to struts spanning the occiput to C2 or from the occiput to various levels in the subaxial spine. different types of grafts have been used, including cancellous and corticocancellous strips or wedges (unicortical, bicortical, or tricortical), as well as cortical struts, such as the fibula, humerus, and femur. Elia et al. reported their results with onlay grafting without internal fixation.4 They used the autologous iliac crest and reported an 89% fusion rate. Although a successful fusion can be obtained, most surgeons are skeptical that such a high fusion rate can be obtained without internal fixation. An important criticism of onlay grafting alone is the need for external traction/immobilization, which can be difficult for many patients to tolerate. Another limitation of this technique is in malalignment reduction or maintenance of reduction. Postoperative traction is also concerning in patients with ligamentous instability because distractive forces may be contraindicated. Some surgeons continue to employ this technique in children because of the tremendous osteogenic potential in this patient population, as well as the difficulty with internal fixation in pediatric patients with inherently small osseous structures. Most surgeons would still recommend at least wiring in these patients to assist with stability until a fusion is achieved.

Onlay Bone Graft with Wiring

The first modern reference to internal fixation in the cervical spine is from 1891 in a report by Hadra, who performed cervical spinous process wiring.5 Wiring techniques of C1 and C2 developed and usually involved sub-laminar placement at C1 and either sublaminar or spinous process wiring of C2. The first published study in which wire was used in fixation of bone graft to the occipital cervical junction was by Cone and Turner in 1937.6 different occipital wiring techniques have been tried. Passage of wire through the foramen magnum and out through burr holes in the occiput can be complicated by the very adherent dura to the bony margin of the foramen magnum. To avoid this, the foramen magnum is often widened posterocentrally to ease the passage of wires out of the adjacent burr holes. Wertheim and Bohlman popularized a technique that avoided burr holes in the occiput by using a trough at or near the EOP and tunneling wire beneath the outer table.7 This technique circumvented the problems associated with passing the wire through burr holes and potential cerebrospinal fluid (CSF) leaks. A second wire was placed sublaminar at C1, and a third was passed through a drill hole in the spinous process of C2. Autologous corticocancellous strips of the ilium were fixed with this triple-wire technique. In Wertheim and Bohlman’s study, 13 patients underwent an occipitocervical fusion using this technique, and all 13 were noted to have a solid fusion at 2 to 7 years’ follow-up.7 Despite such good reports, wiring does not provide rigid fixation; therefore, patients still require postoperative treatment with a halo for supplementary external fixation, and many surgeons believe that an improved fusion rate is obtained with more rigid internal fixation ( Fig. 21.1 ).

Rod and Wire Technique

In 1986, Ransford et al. introduced the first rigid internal fixation of the occipitocervical junction, which used a looped Luque steel rod to bridge the occiput to the cervical spine.8 The rod was carefully contoured to contact the occiput and cervical lamina for increased rigidity and was fixed with occipital and sublaminar wires. Bone graft assisted in fusion. Other surgeons have made modifications to this technique, including the use of a threaded Steinmann pin and specially designed threaded rods, which arguably improve wire fixation by not allowing the wires to slip on the rod as easily. The rod and wire technique provides more rigid fixation and allows surgeons to reduce and hold reductions at occiput–C1 and C1–C2. Despite the relatively rigid fixation of this construct, however, many surgeons still recommend supplementing the internal fixation with a postoperative halo. Another limitation to the rod and wire technique is seen in patients in whom a posterior decompression has been performed, as well as those who have posterior element fractures, because the typical sublaminar or spinous process wiring would be compromised ( Fig. 21.2 ). Surgeons have reported on the use of facet wiring in these cases, but this fixation has not been commonly used. There is also some concern over the use of wire fixation in patients with osteoporosis and rheumatoid arthritis, as there is a significant risk of screw cut-out with posterior element wiring alone. Moskovich et al. reported on 150 patients with rheumatoid arthritis who presented with vertical instability and multilevel involvement. The patients were treated with a contoured occipitocervical loop affixed with sublaminar wires and no bone grafting. The authors found that there was no clinical difference between those who were fused and those with bone grafting.9 They went on to assume that the clinical improvement might be attributable to stabilization of the joint rather than to osseous fusion. Apostolides et al. reported a 97% success rate for achieving a stable occipitocervical construct in 39 patients.10 They concluded that rigid segmental fixation of the craniovertebral junction using a wide-diameter, contoured, threaded Steinmann pin and supplemental autograft creates excellent fusion with minimal complications.

**Fig. 21.1a–d** **a, b** Lateral and anteroposterior radiographs of a patient who had undergone an occipitocervical arthrodesis with wires in the remote past. Although this often resulted in successful fusions, it required postoperative halo immobilization. Upon presentation to us, there was an obvious pseudarthrosis. In addition, there was subaxial spondylosis and radiculopathy. **c, d** Postoperative anteroposterior and lateral radiographs using modern instrumentation. The revision was performed with a skull plate that was then attached to a screw-rod system to fixate the cervical spine. This provides a rigid construct that obviates the need for a halo and increases the likelihood of a solid fusion.

Rod and Hook Technique

More recently, Faure described an occipital fixation technique in which hooks are placed back to back in the same burr hole, with one hook positioned up and the other down.11 This created an occipital clamp construct. The technique may be most useful in patients with significant osteoporosis, in revision occipital instrumentation cases, or when there is significant thinning of the occipital bone. Another technique that may be recommended in this difficult patient population was described by Pait et al., who used an inside-out inverted bolt in the occiput, with the head of the bolt against the inner table of the occiput. This provided a broad surface area for fixation in the occiput and could be useful in patients with significant osteoporosis.12 One problem with this system, however, is that revisions can sometimes be challenging. Because the burr holes for the bolts are so large, much of the surface area for revision fixation and bone grafting is compromised ( Fig. 21.3 ).