The upper cervical spine is a unique portion of the spine that allows great mobility. The distinctive architecture, including the articulation of the dens with the anterior ring of C1, provides much of the rotation of the cervical spine. The atlantoaxial motion segment provides 50% of cervical rotation and additionally provides approximately 10% of cervical flexion-extension. This high degree of mobility predisposes the C1-C2 level to instability, particularly in the event of ligamentous or bony pathologic changes.
Congenital malformations, trauma, ligamentous laxity, and inflammatory processes such as rheumatoid arthritis are all potential causes of atlantoaxial instability. Various techniques have been developed to fuse the C1-C2 segment to avoid the potential consequences of instability, which include pain, deformity, neurologic decline, and even death. Gallie developed a posterior wiring technique in 1939 that used a structural bone graft for support. Numerous other techniques have also been developed since that time, including Brooks and Jenkins double-looped wiring method. Both methods are good at reducing flexion, but the Gallie technique reduces axial rotation by only 67%, whereas the Brooks-Jenkins method reduces axial rotation by 91%. All of these methods are technically less challenging than some of the more modern techniques, but the results are not as stable. The wiring techniques are also not feasible when the structural integrity of the posterior ring of C1 is compromised. Because of the mobility at these segments, these techniques generally require use of external supplemental immobilization, such as a halo vest.
Cumbersome rigid external immobilization can be avoided with the use of different techniques that provide more rigid internal fixation. This can lead to fewer complications from halo-type immobilization, which is especially critical in the elderly patient. Two of the most prominent modern techniques include posterior C1-C2 transarticular screw fixation, as described by Magerl and Seemann, and posterior C1-C2 fusion with polyaxial screws and rods described by Harms and Melcher. Each of these techniques has advantages and disadvantages, which are the focus of this chapter.
An 81-year-old nonsmoking man with a history of chronic lymphocytic leukemia, atrial fibrillation, degenerative joint disease, and dizziness came to the emergency department after falling from a ladder while trimming a tree of unknown height. The patient fell backward directly onto his neck, back, and arm. Upon arrival to the emergency department he complained of neck, back, and right wrist pain. He reported no loss of consciousness.
Exam: The physical examination revealed a Glasgow Coma Scale score of 15 and normal findings on neurologic assessment. The patient had abrasions on his face and midline tenderness of the entire cervical spine.
Imaging: Computed tomography (CT) of the cervical spine with coronal and sagittal reformatted images showed a type II posterior oblique odontoid fracture with 2-mm posterior displacement of the odontoid ( Figure 12-1 ) ( Tips from the Masters 12-1 ). There was posterior subluxation of the left lateral mass of C1 associated with a mildly displaced intraarticular fracture of the lateral mass ( Figure 12-2 ). There were also nondisplaced fractures of the right C7 lamina and extensive degenerative changes with rotatory scoliosis. The patient’s other injuries included displaced fractures of the left first and second ribs, a fracture of the left maxillary antrum, and a right distal radius fracture.
Anatomic variation is common in the upper cervical spine. Contrast-enhanced CT is advisable in every case.
After obtaining appropriate informed consent, the surgeon proceeded with closed reduction and internal fixation of the type II odontoid fracture with a single 4.0 × 38 mm cannulated screw through a right-sided anterior approach. A satisfactory reduction was obtained at the time of surgery based on intraoperative findings; however, a CT scan done 3 days later showed early signs of loss of solid fixation of the screw ( Figure 12-3 ). The patient’s postoperative course was complicated by delirium and swallowing difficulties.
All placement of screws in C1 and C2 is a physician-directed (off-label) application and requires disclosure and consent.
The patient’s clinical status was monitored carefully, and radiographs taken at 2 weeks postoperatively demonstrated that the odontoid screw continued to lose fixation and the patient developed posterior subluxation of the odontoid fragment ( Figure 12-4 ). At this point, approximately 2 weeks after the index operation, the patient was referred to a cervical spine surgeon.
Initially, nonoperative care using a collar could have been considered based on the minimal displacement of the fracture. The recommendation is to consider that option only in elderly patients with fractures that have minimal displacement and angulation, and that show no change in alignment or fracture gap between radiographs taken in the supine and upright positions. Such radiographs were not made before placement of the odontoid screw, but based on the marked displacement that occurred postoperatively and the fracture of the lateral mass of C1 with posterior displacement of the joint, the spine was obviously unstable, and nonoperative care would also have failed.
Immobilization in a halo is an option for treatment of minimally displaced odontoid fractures in younger age groups. The complication rate associated with halo immobilization is much higher in the elderly, however, so most surgeons have abandoned their use in that population. The risk of using a halo would have been exceptionally high in this case considering the patient’s rib fractures as well as his age.
C1-C2 fusion with a wiring construct alone also is not a good option in this case, since postoperative immobilization in a halo is often used to achieve an acceptable fusion rate.
Although odontoid screw fixation is a reasonable option for type II odontoid fractures (with the appropriate fracture angulation and without comminution), it was not a good choice for the index operation. Studies have shown lower union rates (70% to 88%) and higher complication rates in the elderly with this procedure. In addition, the fracture of the lateral mass of C1 and rotatory C1-C2 subluxation in this case are signs of instability and foretold a poor result. Some have recommended using two screws in the odontoid, but adding a second screw provides no significant change in the load to failure and would have been unlikely to prevent the loss of fixation.
Revising the screw as an isolated procedure was not an option at the time of referral for the same reasons that it was not a good choice initially. In addition, the anterior aspect of C2 had then failed, and a screw would not have been able to gain adequate purchase in C2. The screw might need to be removed if it migrates or causes esophageal irritation; otherwise it may be left in place.
There has been a trend in recent years to augment posterior C1-C2 fusions using either transarticular screws as described by Magerl and Seemann or C1 lateral mass screws connected via bilateral rods to C2 pars interarticularis screws (Harms technique). This is true across all age groups and across the full spectrum of pathologic conditions, including odontoid fractures in the elderly. The advantages and disadvantages of these techniques were considered when deciding which procedure to use to salvage this case and are reviewed below.
Transarticular Screw Fixation
The technique of augmenting a C1-C2 fusion with transarticular screws was described by Magerl and Seemann in 1987. It involves placement of bilateral screws, inserted from a posterior approach, through the pars interarticularis of C2 and across the C1-2 joints. The screws achieve excellent fixation because they engage three or four cortices, depending on whether the tip of the screw is bicortical or unicortical in the lateral mass of C1. The increased rigidity conferred by the screws eliminates the need for halo immobilization postoperatively and leads to a higher fusion rate than with wiring constructs. Furthermore, the screws provide a means for achieving fusion in patients with a deficient posterior arch of the atlas. The traditional wiring techniques are not applicable in that setting, since they rely on the wires encircling the arch to provide stability.
Although the placement of transarticular screws has distinct advantages, the procedure is a technically demanding one that requires substantial preoperative planning to minimize risk to nearby neurovascular structures. Because of variation in patient anatomy, a detailed understanding of the patient’s neurovascular and bony anatomy is essential. The course of the vertebral artery through the lateral mass of the axis has been found to be asymmetric in approximately one half of patients. A high-riding transverse foramen is located in the path of a transarticular screw in approximately one fifth of patients. In addition, the isthmus of C2 in approximately 10% of patients is less than 5 mm in height and width, which makes fixation with a 3.5-mm screw risky, especially if placed under fluoroscopic guidance alone.
The internal carotid artery (ICA) is also at risk during transarticular screw fixation, especially when bicortical fixation is needed. The center of the lateral mass of C1 is the optimal exit point for transarticular screws. Unfortunately, the ICA may be located within 1 mm of that site ( Figures 12-5 and 12-6 ). A transarticular screw with bicortical purchase puts the ICA at moderate risk on at least one side in approximately half of cases and at high risk in 12% of case. Therefore, careful planning with contrast-enhanced CT is advised before transarticular screw fixation. If preoperative imaging reveals that the ICA is within 2 mm of the anterior cortex of C1 and medial to the transverse foramen, a unicortical screw or alternative technique should be used. This practice is also supported by biomechanical data, which show no statistically significant difference in pullout strength of unicortical versus bicortical transarticular screws in this setting, if the bone quality is good and the screws are well positioned.
Another limitation of the technique is that spine malalignment must be reduced before the screws are placed. Incomplete C1-C2 reduction may lead to increased risk of vertebral artery injury if C1 is subluxed anteriorly or to failure of fixation in the lateral mass of C1 if the atlas is subluxed posteriorly.
Finally, because of the acute angle required for insertion of these screws, it may be difficult or impossible to achieve the correct trajectory in obese patients or patients with increased thoracic kyphosis.
Posterior C1-C2 Fusion with Polyaxial Screw and Rod Fixation
The technique of posterior C1-C2 fusion with polyaxial screw and rod fixation was originally developed in 2001 by Harms and Melcher based on the pioneering work of Goel and Laheri with C1-C2 posterior plates. The technique has many of the advantages of the transarticular screw technique, including its stability and lack of reliance on the arch of the atlas. Biomechanically, transarticular screws and polyaxial screw and rod constructs have been found to be equivalent. The placement of the lateral mass screws in C1 and the pars-pedicle screws in C2 provides solid fixation that spans the entire sagittal dimension of the bones.
The screw and rod construct has several advantages over transarticular screws. The C1 and C2 screws can be placed independently, which allows the spine to be reduced after the screws have been placed. This permits more flexibility in screw positioning, which may make the technique safer and feasible in a larger percentage of cases. It also allows more control over the final alignment of the spine and does not require placing the patient into an awkward position on the operating table. In addition, the C2 pars-pedicle screws of the screw and rod construct are angled more medially than transarticular screws, which provides a greater margin of safety relative to the vertebral arteries. Although the C2 screw is often described as a pedicle screw, the implant is generally placed into the pars interarticularis, because the C2 pedicles may be too small to safely accept a screw and the exaggerated angle required for a pedicle screw is often difficult to achieve. In cases in which neither a pars nor a pedicle screw can be inserted safely, the screw and rod construct may be connected to laminar screws in C2. These modular options also give the screw-rod construct the advantage that the screws can be easily connected to fusion levels above and below C1-C2.
Although the screw and rod construct theoretically has a lower risk of injuring the vertebral artery than transarticular screws, the ideal exit points of the screws in the C1 lateral mass are the same and therefore carry a similar risk to the ICA and the hypoglossal nerve. Both techniques require special preoperative imaging and rigorous planning as noted previously ( Tips from the Masters 12-3 and 12-4 ). The primary disadvantages of screw and rod constructs over transarticular screws are the increased cost and complexity of the former.
Preoperative planning and proper patient positioning are the keys to success.
Rehearse the steps with your operating room team before placing a C1 screw so your eyes can stay focused on the wound.
Transarticular Screw Technique
Preoperative advanced imaging is required with contrast-enhanced CT to further understand the neurovascular anatomy as well as the bony dimensions of the atlantoaxial complex. In certain cases, especially when a high-riding transverse foramen is present or the ICA is close to the anterior portion of the atlas, transarticular screws with or without bicortical fixation may not be possible.
Once adequate preoperative planning has been completed, the patient is brought to the operating room and intubated. Multimodality neurologic monitoring can be instituted with the patient in the supine position to obtain baseline readings. The patient is then placed prone on a Jackson table with the head held in place with a pinion and Mayfield attachment. Reverse Trendelenburg position can be maximized using the Jackson table by inserting the head of the table in the highest selectable position and the foot of the table in the lowest selectable position. This will help reduce venous pressure and decrease bleeding during the procedure. Fluoroscopy is used to reduce the deformity and place C1 and C2 in the proper orientation for the instrumentation, and to demonstrate that adequate images can be obtained in the chosen position. Rotation of the head should be neutral. A second set of neurologic monitoring tests can be performed once reduction is complete, and a postposition wake-up test can be conducted if deemed necessary.
The posterior scalp and neck are clipped or shaved, and are then prepared and draped in the usual sterile fashion from the occiput proximally to the midthoracic spine distally. Once this has been accomplished, a midline skin incision is made over the C1-C2 level and skin and subcutaneous tissue are retracted to reveal the cervical fascia. The cervical fascia is incised in line with the skin incision. A careful subperiosteal dissection of the C1 arch and C2 lamina are accomplished, and self-retaining retractors are inserted. If possible, the insertion of the semispinalis cervicis should be preserved on C2. If it must be sacrificed, ideally it can be removed with a small piece of bone from the spinous process C2 to provide anchor for repair at the end of the procedure. The arch of the atlas is exposed, with care taken to protect the vertebral artery on its superior surface. The starting point for the screws is located just lateral to the junction of the lamina with the articular mass, close to the lower edge of the caudal articular process of C2 so that the isthmus is crossed close to the posterior surface. A caliper and the preoperative CT scan can assist in measuring the distance between the two screws. This should be compared with the medial border of the pars interarticularis so that violation of the canal is avoided. The starting point can be created using a 2-mm high-speed bur to avoid slipping off the axis while using a Kirschner wire (K-wire) or drill bit.
Once the starting point is located, the trajectory needs to be determined. This is done using fluoroscopy in a lateral view or image guidance. Use of the percutaneous technique, described by McGuire and Harkey, makes it unnecessary to expose the entire region from C1 to the entrance point of the screws. Once the trajectory is determined, small vertical stab incisions are made in the upper thoracic spine (around T2) to allow for appropriate placement of the trocar along the correct trajectory. These stab incisions are generally the same width apart as the entrance holes burred in C2, since the trajectories of the transarticular screws are generally parallel to the midsagittal plane. The trocar is then passed through the stab incisions and docked in the entrance holes in C2. On the lateral image, the trocar should be aiming at the anterior tubercle of C1. The ideal target point for the drill, relative to the anterior tubercle, is just below the midpoint in the 20% to 40% region, where 100% represents the cranial border of the arch and 0% the caudal border. A cannula is passed over the trocar and the trocar is removed. A 2.5-mm drill bit is then inserted and under fluoroscopy is passed into the C1 lateral mass. The hole is then tapped, and a 3.5-mm screw is inserted under fluoroscopic guidance. The same process is repeated on the contralateral side. The posterior elements are decorticated and a bone graft with posterior wiring can be placed based on surgeon preference. If the posterior arch of C1 is deficient, then the C1-2 joints are decorticated and packed with bone graft. If a wiring construct can be used, it is not necessary to directly fuse the C1-2 joints.
The semispinalis cervicis is repaired if it was detached during exposure. The wound is irrigated and closed in a layered fashion over a drain. A sterile dressing and collar are applied. The patient’s neurologic status is checked postoperatively.
Technique of C1-C2 Fusion Using a Screw and Rod Construct
The setup for C1-C2 fusion using a screw and rod construct technique is the same as for placement of transarticular screws, except that it is not mandatory to reduce the spine before exposure and the head and neck may be placed in physiologic alignment. Refer to the previous section for further details. Once the patient is prepared and draped, an incision is made over the midline of C1 and C2. Dissection is carried down to the C2 spinous process. Again, care should be taken to preserve the attachments of the semispinalis cervicis muscle to the C2 spinous process if possible. The superior aspect of the C2 lamina should be identified and subperiosteally dissected to identify entry points for the C2 pars screws. Dissection of the posterior arch of the atlas is carried out laterally to identify the lateral masses of C1. The vertebral artery should be avoided as described previously. Bleeding often occurs during dissection around the epidural venous plexus near the C1-2 joint. This can be controlled with thrombin-soaked absorbable gelatin sponges and judicious use of bipolar electrocautery ( Tips from the Masters 12-5 ). The C2 pars interarticularis or pedicle screw entry points are identified by locating the medial border of the pars. The entry point is marked with a 2-mm bur. Generally the direction of the bit is approximately 25 degrees medial and cephalad. The entry point should be in the cranial and medial quadrant of the isthmus surface of C2.