OAA arthrodesis can be performed using a variety of techniques. These have evolved over the years with technologic advances in implants and grafts. Initially, there was rib or iliac crest autograft. Later, there was cable and wire fixation through a loop placed through the occipital bone. Plate and screw fixation came next. Currently, it is often keel plate and rod fixation with segmental atlantoaxial and subaxial screw fixation that is the method of choice. Sublaminar wiring and hooks can also be used when screw placement is not possible.

A thorough preoperative evaluation is essential to maximize fixation and minimize complications. Flexion and extension laterals are important to quantify instability in patients with chronic pathology, such as rheumatoid subluxation. CT and MRI studies will help delineate the anatomy. In cases where occipital screws are contemplated, a CT scan is especially valuable to help plan the length of screws and to minimize penetration into the cranium. In cases of trauma where vertebral artery injury is suspected, MR or CT angiography will help to define vessel anatomy and pathology.

In certain situations where the cervical pathology may benefit from traction, placement in halo traction before imaging would help determine reducibility of the pathology. Patients may arrive in some sort of fixation, such as a halo or cervical collar, or may have chronic pathology and not be wearing any supportive devices.

Prior to any dorsal cervical procedure with instability or spinal cord compression, it is imperative that the operative team be well-versed in prone spine surgery. Lateral fluoroscopy or image guidance during the surgery is beneficial. The anesthesiologist may want to do an awake fiberoptic intubation using regional and topical anesthesia. In this case, the intubation should be done with the patient in a hard cervical collar, and neurologic function should be assessed before they are induced and turned prone. In the case of preoperative halo fixation, the patient will arrive in the operating room with the halo on. Portions of the halo may be left connected during induction, which provides more stability and less movement of the spine during the procedure. An adjunct to this would be the placement of somatosensory evoked potential monitoring with motor evoked potential (MEP) monitoring leads in all four extremities. Baseline potentials can be obtained prior to positioning, following positioning, and during the surgery to maintain safety from positioning and through the procedure.

In those patients without a halo, there are several options for fixation during the surgery. Two-point skull tongs can be placed prior to positioning. With the head on a horseshoe headrest, traction can be set to the desired weight. Alternatively, a Mayfield three-point skull tongs system can be placed before positioning and then secured to the bed afterward. Regardless of technique, it is important to make sure that a cervical collar is on and snug prior to turning the patient. The author prefers three-point fixation with the Mayfield fixed to the table. During the procedure, if a reduction or change in alignment occurs, care must be taken to make sure that the cervical spine does not “swan neck.” If this occurs, the head may need to be repositioned.

For those with a halo, the ring alone can be attached to the traction, firmly attached to the bed. For very unstable spines that have been kept in reasonable alignment in the halo, portions of the halo can be maintained during the surgery to provide stability during the procedure.

Paralytic anesthetic agents should be avoided, especially in cases where spinal cord MEPs are being performed.

Preoperative antibiotics should be given at least 1 hour prior to incision. The author prefers 2 g cefazolin preoperatively, with 1 g given every 6 hours afterward.

A midline incision is made from the EOP distally. Typically, an incision down to C4 is necessary for adequate exposure and screw placement from the occiput to C2. Subperiosteal dissection can be made sharply or with electrocautery from the occipital bone to the inferior extent of the arthrodesis. If there is tremendous instability, clamps can be used to secure and stabilize the dorsal arch of C2.

Dissection of the C1 superior arches can only proceed laterally about 10 mm in either direction to avoid injury to the vertebral arteries, which become tortuous and lie ventrally on the lateral arms of the arch at this level. As dissection on the superior aspect of C1 approaches 1 cm lateral to midline, it would be wise to not use electrocautery and instead use Penfield probes to continue dissection. The rod/loop is then brought into the field. Assuming a precontoured rod is not available, the rod can be bent to form a horseshoe. If a loop is used, it can be bent appropriately. The lateral O-C1 angle is quite significant, and the rod will have to be bent significantly in the sagittal plane. The width of the transverse arm should be about 4 cm. Ideally, this part should be placed about 3.5 cm dorsal (superior) to the foramen magnum. Each limb of the rod should extend down to at least one cervical level below the inferior extent of the arthrodesis.

After dissection with an angled instrument, the foramen magnum can be enlarged with a Kerrison punch, extending dorsally about 2 cm and laterally about 1 cm.

With the decompression complete, the wire fixation is then placed. The occiput is penetrated in two to four locations spaced equally approximately 1 cm from the foramen magnum decompression. The drill holes can be between 6 and 8 mm in diameter. Angled instruments can be use to create a safe plane over the dura to pass the wires. Wires are passed from the trephine holes through the foramen magnum opening (5). Alternatively, Wertheim and Bohlman (6) passed a loop twice through a hole drilled across at the EOP. Sublaminar wire technique can be used to fixate to C1, C2, and then distally as needed. A small laminotomy can be made laterally in C2 to keep the wires lateral at that level. Wires at C1 may need to be more medial depending on arterial anatomy. Typically, only the medial 2 cm of the arch needs to be removed to decompress the spinal cord. The lateral portions (just medial to the vertebral arteries) that remain could be wired into the construct. If more of the C1 arch has been removed for decompression, then that level will obviously not have fixation along the dorsal arch. Cables can then be tightened, starting at C2, then at the occiput. C1 arch cables cannot be tightened to the extent that the other levels can. If Songer cables are used, between 6 and 10 pounds of torque can be set on the wrench— more tension at C2 and less at C1. The occipital cables will be in the most robust bone and can be tightened more than the cervical loops (7). The surgical tools and instrumentation required for wire-rod techniques are portrayed in Table 137.2.

At this time, a thorough irrigation can be done. Next, bone and graft can be packed around the outside of the loop and over intact dorsal arches, after thorough decortication of the surrounding bone. Structural grafts can also be wired onto the bone with drill holes and wires.

Closure of the soft tissues is performed in layers. The muscle can be reapproximated with 0-Vicryl simple stitches.
The fascia can be closed with O-Vicryl figure-of-eight stitches. Subcutaneous closure can be done in any standard manner.

As long as fixation was reasonable, cervical collar immobilization for 4 to 6 months would be reasonable. If fixation is suboptimal, continued halo immobilization for at least a few weeks would be prudent (5).