Odontoid process fractures are common injuries resulting from a combination of hyperextension and axial rotation. Anderson and D’Alonzo type II fractures are the most common type of injury and occur in 65 to 74% of cases. Nonunion rates with rigid immobilization have been 12 to 54% in older patients (i.e., > 65 years of age) compared with as low as 4% in younger patients (≤ 65 years old). Although odontoid process fractures can sometimes be successfully treated with halo vest immobilization, the demonstrated success of direct screw fixation has made it the treatment of choice, especially in patients older than 50 years of age. Moreover, complication rates of up to 66%, as well as mortality rates of 42%, have been reported with the use of halo fixation in the older patients, making operative fixation a useful approach to avoid complications.
Direct screw fixation stabilizes the neck and provides the best milieu for fracture healing by closing the gap and preventing motion between the odontoid process and the body of C2 while preserving C1–2 motion, which accounts for 50% of the normal head axial rotation. It is a straightforward procedure that is well tolerated by the patient. The available instrument systems allow odontoid screw fixation to be done in a less invasive manner under precise biplanar fluoroscopic control.
5.2 Patient Selection
Patients with Anderson and D’Alonzo type II and high type III odontoid fractures are candidates for this approach. The greatest success is achieved with horizontal or posterior oblique fractures that slope downward from anterior to posterior (Grauer type IIB), but with care during patient positioning, the approach can be used in anterior oblique fractures (Grauer type IIC) as well ( ▶ Fig. 5.1). Concomitant C2 body fractures are a contraindication to this approach because the screws may fail to hold in the C2 vertebral body. Retrolisthesed odontoid fractures are not a contraindication when the system described here is used because the system provides a way to reduce them safely. Older patients may also benefit from direct screw fixation because they heal poorly with immobilization but do well with this minimally invasive approach that allows rapid return to their normal environment.
Fig. 5.1 Two common classification systems for odontoid fractures are shown with implications for clinical treatment (a–c). The Anderson and D’Alonzo classification categorizes fractures into type I (odontoid tip), type II (odontoid base), and type III (odontoid base extending into the body). The Roy-Camille classification aimed to improve delineation of type II fractures by classifying the orientation of the fracture line (d) horizontal or (e,f) oblique orientation. Similarly, the Grauer classification improved deliniation of Type II and III fractures by subclassifying fractures that were nondisplaced (type IIA), displaced transversely (type IIB), or comminuted at the odontoid base (type IIC). Grauer type IIB fractures were reported to be ideal for anterior odontoid screw placement compared with type IIC fractures.
In our experience, fractures up to 6 months old have healed as well as fresh fractures using this technique, so it may be considered for patients for whom a trial of immobilization has failed. On the other hand, chronically nonunited fractures do not fuse well (~25% union rate), so we recommend a posterior C1–2 fusion for these patients instead. Other criteria that are generally accepted for operative odontoid fixation include axis deviation of > 11 degrees, fracture dislocation of > 5 mm, a fracture gap of > 2 mm, oblique or transverse fracture, atlantoaxial instability of > 50%, movement at the C1–2 joint, and instability during dynamic fluoroscopy.
5.3 Preoperative Preparation
We typically obtain an MRI of the cervical spine to assess integrity of the transverse ligament as a prerequisite to pursuing this surgical approach. Proper patient positioning and arrangement of the fluoroscopy equipment may take more time than the actual operation but will greatly facilitate the procedure. If the patient’s neck is unstable in extension, intubation techniques that avoid neck extension, such as “blind” nasotracheal intubation, intubation with a light wand, or fiberoptically assisted intubation, should be used. In the case of anterolisthesis of the odontoid, which reduces in extension, no special technique is usually required. Routine preoperative antibiotics are administered as soon as the patient is anesthetized.
To obtain the desired screw trajectory, the patient’s neck usually needs to be maximally extended. The patient is therefore positioned on the operating table with a folded blanket beneath the shoulders to increase neck extension. If the fracture reduces in extension, a hyperextended position is chosen using lateral fluoroscopy ( ▶ Fig. 5.2); however, if the odontoid is retrolisthesed and dislocates further in extension, the head is initially elevated on folded towels to keep the neck in a neutral position. The neck will be extended later once the guide tube is placed and the body of C2 is repositioned posteriorly in alignment with the odontoid. Halter traction with 10 lbs of weight holds the head immobile. A wine-bottle cork can be used as a radiolucent bite block to keep the jaws open for the transoral view. It is manually notched for the teeth or gums.
Fig. 5.2 The patient is positioned for surgery. Note that the thorax is elevated on a folded blanket and the neck is hyperextended (the patient’s fracture reduced in extension). Positioning was achieved while monitoring spinal motion under lateral fluoroscopic control to avoid spinal canal compromise. Two fluoroscopic units are used for sequential intraoperative anteroposterior (transoral) and lateral fluoroscopic monitoring. Halter traction (10 pounds) secures the head.
Two portable C-arm fluoroscopic units are positioned to provide lateral and anteroposterior (AP) (transoral) views of the odontoid. The lateral imaging device is placed with the arc around the head of the table so the AP fluoroscope can shoot through the radiolucent table top. This can be done on a standard operating room table ( ▶ Fig. 5.2). If only one fluoroscopic unit is available, it must be rotated back and forth from the AP to the lateral position. Freedom to achieve this movement should be ensured before draping.
5.4 Operative Procedure
We describe the procedure using a widely available odontoid instrumentation system (Aesculap, Center Valley, Pennsylvania). The initial portion of the exposure should be familiar to most spine surgeons because it is identical to that used for an anterior cervical diskectomy. After routine preparation and draping, a unilateral horizontal incision is made along a natural skin crease ( ▶ Fig. 5.3) at about the level of the fifth cervical vertebra. Prior skin infiltration with a 1:200,000 epinephrine solution will help achieve hemostasis, which is then secured with bipolar cautery. The platysma muscle is elevated and divided, and the fascia of the sternocleidomastoid muscle is sharply incised along its medial border. Blunt finger dissection easily exposes the anterior surface of the spinal column at the midcervical level by opening natural planes medial to the carotid sheath and lateral to the trachea and esophagus. The fascia of the longus colli muscle is incised in the midline, and the muscle is elevated from the vertebral bodies at about the C5–6 level to allow firm fixation of the retractor blades, which is important because the cephalad retractor will pull against these blades and, unless firm fixation is achieved, may dislocate them.
Fig. 5.3 The site of the incision within a skin crease in the midcervical region approximating the C5–6 vertebral level is shown (dashed line). Inset shows the retractor in place. Note the lack of any retractor components inferiorly in the wound, which would impede achieving a proper trajectory.
The sharp, large-toothed Caspar blades (Aesculap) are inserted beneath the longus colli muscle bellies bilaterally and secured in the lateral retractor. These blades, rather than the small, fine-toothed blades, should be used to anchor the retractor system firmly. Blunt dissection with a “peanut dissector” in the retropharyngeal space quickly and easily opens a tunnel in front of the vertebral bodies up to C2. An angled retractor of the appropriate size is then inserted into this space. It is attached to the retractor blade holder, which in turn inserts into one side of the previously placed lateral retractor ( ▶ Fig. 5.3). This device allows angulation of the blade as needed. No inferior retractor is placed as it would interfere with achieving the correct trajectory for the approach to C2.
A K-wire is inserted through the incision and impacted into the inferior edge of C2 under fluoroscopic control ( ▶ Fig. 5.4 a). If a single screw is to be placed, a midline entry site is chosen. A paramedian position ~ 2 mm off the midline is used if two screws will be placed. A hollow 8-mm drill ( ▶ Fig. 5.4 b, c) is placed over the K-wire and rotated by hand to create a shallow groove in the face of C3 and the C2–3 disk and anulus to the inferior border of C2 ( ▶ Fig. 5.4 b, d). The drill guide system, which consists of inner and outer drill guide tubes that are mated together, is then placed over the K-wire ( ▶ Fig. 5.5). The spikes on the outer drill guide are walked up the face of the spine until they are over the third vertebral body. A plastic impactor cover is placed over the guide tube assembly after the K-wire is cut short, and the spikes of the outer guide tube are firmly set into the third cervical vertebra ( ▶ Fig. 5.6). This is monitored fluoroscopically, and it is essential that the guide tube remain in place using firm forward pressure for the remainder of the case until the final screw is placed. The inner drill guide is then extended in the previously placed groove to contact the inferior edge of C2 ( ▶ Fig. 5.6, inset) so it can accurately guide the drill into C2.
Fig. 5.4 (a) A K-wire is inserted into the anterior inferior edge of C1 under fluoroscopic control. A hollow-core drill is then placed over the K-wire and rotated by hand (c) to cut a shallow groove in the face of (b) C3 and (d) into the C2–3 anulus. This should not remove any bone from the inferior edge of C2 (arrow, d).