The Evolution of Posterior Cervical and Occipitocervical Fusion and Instrumentation

John R. Vender, MD; Andy J. Rekito, MS; Steven J. Harrison, M.; Dennis E. Mcdonnell, MD.


Neurosurg Focus. 2004;16(1) 

In This Article

Atlantoaxial Fusion

Fusion and instrumentation strategies that focus only on the atlantoaxial junction have followed a similar evolution. Although these constructs do not face the same requirements when the occiput (and therefore the occipitocervical angle) is not a factor, significant mechanical strain must still be resisted. In extreme cases of atlantoaxial instability, particularly if the posterior arch of C-1, the lamina of C-2, or the lateral masses or facets of C1–2 are significantly abnormal, fusion and instrumentation constructs are extended rostrally to include the suboccipital bone or caudally to include the subaxial cervical spine.

Gallie Fusion. The Gallie type of fusion,[25] initially described in 1939, simply involves the placement of a notched bone graft between the dorsal portion of the arch of C-1 and the posterior spinous process and medial laminar arches of C-2 (Fig. 3). This graft is secured by sublaminar wires, which pass beneath C-1 and C-2. The primary drawback of this technique is that it represents a single, midline point of fixation that is susceptible to rotational forces.[25] In addition, concern exists when more than one spinal segment is spanned using sublaminar wires because of the increased risk of encroachment on neural elements within the spinal canal during the actual passage of wires as well as the risk of chronic encroachment of the wires over time. This is particularly significant in cases in which anterior translation of the atlas results in a decrease in spinal canal dimensions and a relative spinal canal stenosis. Therefore, although easy to perform, this method is not very stable and requires a rigid orthosis, preferably a halo, for postoperative immobilization.[49]

Drawing showing a Gallie type of fusion with two variations. Upper: A single loop of wire around the posterior arch securing the onlay bone graft to the posterior arch of C-1 and the superior laminar surface of C-2. Lower: The double-loop technique with a wire around the arch of C-1 again securing the onlay graft in place. Both examples eliminate the need for sublaminar wires at C-2.

Brooks Fusion. The Brooks type of fusion,[4] described in 1978, overcomes the rotational deficiencies of Gallie fusion by incorporating bilateral, interlaminar bone grafts (Fig. 4 upper). Again bilateral sublaminar wires, which have been passed under both C-1 and C-2, are used to secure the individual grafts. Notches can be placed in the bone graft to secure the position of the wire. By tightening the wire, compression can be applied to the graft.[4] Modifications that have improved the results attained using this method include oversized autogenous iliac crest bone grafts and the use of double strands of wire for each individual graft.[49] The Brooks fusion method shares the disadvantage associated with the Gallie procedure of requiring multisegment sublaminar wires. Heavy-gauge singlestrand wires pose a significant danger of causing injury to the spinal cord during their passage. The evolution of braided cables has made the passage of these wires somewhat safer.[26,36,73]

Upper: Drawing showing a Brooks type of fusion with bilateral interlaminar bone grafts secured with C-1 and C-2 sublaminar wires. The bilateral fixation afforded by this technique better resists rotational forces. Lower: Halifax clamp technique in which bilateral interposition grafts are secured by the Halifax clamp construct.

Sonntag's Modified Gallie Fusion. In this modification the sublaminar wires under C-2 are eliminated and, instead, the spinous process of C-2 serves as a fixation point. Intraoperatively, the C1–2 interlaminar space is widened. The inferior surface of the C-2 spinous process is notched. A loop of wire is then passed sublaminarly under the C-1 process in the caudorostral direction. An autograft or allograft strut is fashioned in a similar way to the Gallie procedure. The wire is then passed over the graft, which can be notched to accept and secure the wire. The loop is passed over the arch of C-2 and is secured to the undersurface of the spinous process that rests in the notch previously created. This secures the bone graft in the interlaminar space that is contained within the wire loop. The free ends of the wire are then brought over the C-2 laminar surface and secured either by twisting them or by being secured with a crimp device after an adequate tightness had been achieved. This connection secures the formerly free ends of the wire against the inferior surface of the C-2 spinous process below the looped portion of the wire. Again, the use of braided multistrand wire facilitates the passage of the wires because of its decreased wire rigidity; this not only makes wire passage easier, with a lower rate of morbidity, but allows greater tightening and results in better fixation and a stronger construct.[49]

Locksley Intersegmental Tie-Bar Technique. The Locksley intersegmental tie-bar technique can also be used to fix the atlantoaxial articulation (Fig. 5). The grafts are secured with sublaminar wires that are twisted in a figure- eight fashion. The Locksley method differs from other onlay graft–wiring techniques because of the addition of a posterior stabilization plate, which is secured by wires to the posterior spinous process. In this fashion the Locksley method applies three-point fixation with immediate rigidity and resistance to all axes of movement. The Locksley method can be applied to any cervical segment and can also span the cervicothoracic junction. The autogenous rib graft can be cut cross-wise with a full-sized rib used bilaterally. For longer constructs, to avoid harvesting two ribs, the rib graft can be split in the midline and each half of the rib can be used separately. The rib, by its natural contour, is an ideal graft selection. An iliac autograft can also be used.[80]

Drawing showing the Locksley method in which sublaminar wires are used with bilateral onlay rib grafts, spinous process wires, and a posterior spinous process plate. Unlike the other sublaminar wiring techniques, the addition of the posterior spinous wires and plate creates a three-point fixation.

Stabilization of the atlantoaxial region can be achieved with the addition of acrylic resins that are polymerized in situ, which provides immediate stability. These resins must be used cautiously and sparingly, however, because they do not promote fusion. Application of methylmethacrylate can impart immediate stability to the C1–2 interspace. Nevertheless, methylmethacrylate will not bond directly to bone and therefore, the acrylic construct must be "anchored" to the bone by applying either wires or pins. Several key technical points must be observed for this fusion to be successful. If wires are to be used, there must be adequate space around the wire so that the wire can be contained completely within the body of methylmethacrylate. Other fixation devices that have been described include Kirschner wires passed through the spinous processes or screws placed in the articular pillars. For inclusion of the arch of C-1, screws are placed obliquely into the long axis of the arch, with the head protruding at least 4 mm. An appropriate amount of acrylic must be applied to provide an adequate structural strength to resist the forces involved. The area must be keep free from blood and CSF until the acrylic has hardened.[17,49] Great care must be taken while applying this method to assure that the exothermic reaction that is created when the acrylic is curing does not subject the spinal cord or nerve roots to elevated temperatures. In experienced hands this can be managed safely.

Claw-type constructs such as Halifax clamps or Knot's rods have also been applied to the C1–2 segment. The Halifax clamp was initially described by Tucker[78] in 1975 (Fig. 4 lower). Halifax clamps have also been used successfully in lower cervical spinal segments, as documented by a long-term follow-up review.[12,18,34] Using these techniques, immediate fixation is achieved and the risk associated with sublaminar wires is avoided. Bone grafts are again placed in the interlaminar space bilaterally, similar to a Brooks-type fusion. These bone grafts are secured under the interlaminar clamps.[51] In most cases these devices are easy to apply, although difficulty can occur if the laminar hooks do not fit readily in place. Loosening of the screws and dislodgment of the clamps have been described. Newer clamps have an increased number of threads to reduce the risk of screw backout and loosening. Other solutions to this problem include the placement of acrylic over the screws and threads or destruction of the threads with a burr or clamp after the devices have been tightened. Overtightening has also been described and has resulted in increased angulation of the odontoid, causing ventral encroachment and possible compression of the spinal cord. The risk of this complication can be minimized by using bone grafts that have been carefully sized to prevent overcompression.[49] Statham, et al.,[75] reported complications in 14 of 45 patients who underwent atlantoaxial arthrodesis. In 10 patients screw loosening occurred and in four patients one of the clamps disengaged. Nine patients (20%) required a repeat operation. The unique anatomical features of the atlantoaxial motion segment were cited as an explanation for this widely observed failure rate at the level of C1–2. The posterior arch of C-1 is more rounded, in contrast to the sharp edge of the subatlantal laminae. This creates a mechanically suboptimal interface between the arch and the clamp at C-1. In addition, it is difficult to achieve vertical alignment of the clamps because of the large spinous process of C-2. This results in angulation of the construct in the sagittal plane, which decreases the surface area of contact between the clamp and the arch of C-1. This may further increase the susceptibility of the constructs to disengagement during rotation. Loosening of the screws can compound this complication.[75] Huang and Chen[35] in 1996 reported on a series of 32 patients who underwent atlantoaxial fixation and arthrodesis with Halifax clamps. Contrary to other series, these authors had a 100% successful fusion rate. In their patients, however, a halo orthosis was applied to all patients for 3 months.

A major advance in posterior C1–2 fixation was initially described by Magerl (Fig. 6). Fixation of the C1–2 lateral masses with transarticular screws provides immediate stabilization. The method does not require a complete or competent arch of C-1 for success. In most cases an interspinous Gallie bone graft is placed between the posterior arch of C-1 and the posterior spinous process and superior laminar surface of C-2. Careful preoperative planning is required, with thin-cut CT scans and CT reconstructions being used to confirm that adequate bone exists to allow passage of the screws without injury to the VAs. After the screws have been passed, immediate fixation is attained. An interspinous Gallie bone graft is often used and may or may not be supplemented with wires. In this procedure the occiput and the posterior arches of C-1, C-2, and C-3 are exposed subperiosteally. Passage of the wire, drill, and screw require a fairly "flat" angle and, frequently, secondary incisions need to be created in the upper thoracic region to obtain the required angle, thus avoiding the need to perform a much longer cervical incision. By exposing the superior edge of the C-2 lamina and the isthmus bilaterally, the medial surface of the C-2 pedicle can be palpated using a probe. The posterior surface of the C1–2 joint and capsule are then exposed laterally to the sulcus arteriosus of the C-1 arch. Fluoroscopic control is used to monitor the orientation of the drill. The drill entrance point is located on the inferior facet of C-2 2 mm lateral from the medial edge and 3 mm above the caudal edge. With the aid of the fluoroscope, the trajectory is adjusted to exit C-2 at the posterior edge of the superior articular surface, near the entrance of the C-2 pedicle. The lateral mass articulation is crossed so that the subarticular surface is entered at its posterior to middle third. The ideal trajectory orients the drill toward the anterior arch of C-1. The drill orientation is maintained in the sagittal plane or angled 10° medial to the sagittal plane. One preferred method consists of using two drills so that the initial drill can be left in place to stabilize the C1–2 relationship while the opposite side is accessed and the screw is placed. Instrumentation can then be placed on the initial side. If the VA is damaged, pulsatile bleeding will be seen and the procedure cannot be completed on the opposite side. The primary deficiency of this procedure occurs when an inadequate purchase is attained on C-1 or C-2 by the drill or screw. This can be minimized by careful patient selection, strict adherence to the procedural protocol, and a thorough understanding of the patient's unique 3D atlantoaxial anatomy.[49]

Drawing showing a Magerl-type transarticular screw fixation. Screws are seen in their trajectory on both the anteroposterior (upper) and lateral (lower) views. On the lateral view the interspinous bone graft is demonstrated. With the addition of wires or hooks this would provide a three-point fixation.

Olerud and Olerud[54] in 2001 described a new fixation device for C-1. Transarticular screws are passed through a plate adapted with an arm containing a claw that is fixed to the posterior arch of C-1. In this fashion the transarticular screw technique, which alone provides two-point fixation, is converted into a three-point fixation construct. In the cases described by these authors midline posterior wiring was not required.

In 2002 Matsumoto, et al.,[48] described the use of titanium mesh cages in place of structural autograft in cases in which transarticular screws are used. In this procedure, the cages are filled with autograft and placed between the posterior arch of C-1 and the superior laminar surface of C-2 bilaterally. The inferior surface of the posterior arch of C-1 and the upper portion of the C-2 lamina are prepared to accept the cage by using a diamond-burr drill. The cages range in height from 8 to 10 mm. These cages can be tightly secured with a titanium cable placed sublaminarly at C-1 and C-2. In the series described by Matsumoto, et al., a double cable was used; it was cut into two separate pieces after passage. Additional cancellous autograft bone chips were placed across the posterior arch of C-1 and the lamina of C-2 to supplement the fixation further. This technique provided a rigid fixation. Four of five patients underwent simultaneous placement of transarticular screws and posterior cage and wire fixation. In one patient the cage was inserted and posterior wiring alone was performed because there was difficulty placing the transarticular screws. Traction was applied to facilitate cage placement and wire passage. No rigid orthosis was required and only a soft collar was used if necessary for postoperative wound-related pain. In all patients the fusion was successful.

Harms and Melcher[33] described a novel method of atlantoaxial stabilization in 2001. Polyaxial head screws are inserted into the lateral mass of C-1 and the pedicle of C-2. These screws are secured by rods bilaterally. Using this technique C1–2 fixation can be attained in patients who are not candidates for transarticular screws because of fixed subluxation of C-1 on C-2 as well as in patients in whom the VA is in an aberrant location, which would preclude the passage of screws. In these patients 3.5-mm polyaxial screws are inserted into the lateral masses of C-1 and through the pars articularis into the pedicles of C-2. If necessary, a fluoroscopically guided reduction of the C1–C2 articulation can then be performed prior to the insertion of the 3-cm rods. A cancellous onlay bone graft can be used as a supplement. No structural grafting or wiring is required. In a series of 37 patients no neural or vascular injury was reported. In all patients a solid fusion was attained. In some cases, if only temporary fixation was required, the implants could be removed.[33] Resnick and Benzel[64] in 2002 described a similar method in which screws are placed in the C-1 and C-2 pedicles bilaterally. Fixation was achieved by using a fixed–moment arm cantilever beam system. In their case report a single patient was presented in whom a transarticular screw fixation was not feasible.

Among wire techniques, with the exception of the Gallie procedure, adequate stability can be achieved.[28] Richter and associates[65] described a comparison of six methods including the Gallie technique alone, transarticular screws with Gallie fixation, transarticular screws alone, transarticular screws and an atlas claw, isthmus screws in the axis and an atlas claw, and lateral mass screws in the atlas and isthmus screws in the axis connected with rods. Their conclusions support previous observations that three-point fixation is superior to one- or two-point fixation. In this particular selection of techniques, the transarticular screws and the atlantal claw provided a rigid internal fixation that was not dependent on bone graft or sublaminar wiring. In cases in which transarticular screws were not feasible because of the patient's anatomy, isthmus screws with a claw or lateral mass screws with a midline fixation would be the next best alternative. Naderi and coworkers[52] evaluated four combinations of cables, grafts, and screw fixation at C1–2. Transarticular screws again were more effective in preventing lateral bending and axial rotation than posterior cable graft constructs alone. Cable graft constructs do prevent flexion and extension better than stand-alone screw fixation techniques. Again, increasing the number of fixation points often will significantly decrease the rotation and translational movement, confirming that it is mechanically advantageous to include as many fixation points as possible. Reilly, et al.,[63] confirmed as well that transarticular screws are superior to wire techniques. Dickman, et al.,[15] evaluated four cable fixation methods. All wire techniques were ineffective with one Gallie fusion technique proving to be no better than an unfixed spine. The Brooks and interspinous methods were more effective than the Gallie methods; however, to attain adequate control of movement at C1–2 by using these posterior fixation techniques (two Gallie types, one Brooks, and interspinous wires), additional fixation is required in the form of a collar, a halo orthosis, or rigid internal fixation with transarticular screws.