Traumatic Injuries of the Craniovertebral Junction
It is useful to classify injuries of the craniovertebral junction (CVJ) as isolated ligamentous injuries, isolated bone fractures, or mixed ligamentous and bony injuries. The extent of the injuries to the bones and ligaments is important for predicting the results of treatment. In this chapter, each category of injury is considered separately. This is an important conceptual framework because ligaments are incapable of repair when disrupted.1–3 Therefore, ligamentous injuries usually require surgery to restore spinal stability. Bone fractures can usually heal as long as the bones can be reduced and immobilized satisfactorily. However, when the bones are fractured extensively and comminuted widely, or when fractures are accompanied by disrupted ligaments, then nonoperative treatments are likely to fail and surgery is required to restore permanent spinal stability. The mechanisms of injury are reviewed in Chapter 3, Biomechanics of the Craniovertebral Junction. This chapter focuses on the clinical presentation, diagnosis, treatment, and outcome of injuries to the articular and bony structures of the CVJ.
Isolated Ligamentous Injuries
Isolated ligamentous injuries include occipitoatlantal dislocations, transverse ligament disruptions, and rotatory C1-C2 dislocations. Occipitoatlantal dislocations and transverse ligament injuries are highly unstable. These injuries require surgical treatment because the ligaments are avulsed and are incapable of healing. Rotatory C1-C2 dislocations, however, are different, less severe injuries that rarely require surgery.
Occipitoatlantal Dislocations
Occipitoatlantal dislocations are usually caused by high-velocity accidents; they are highly unstable and stretch, compress, and distort the spinal cord.4–10 They tend to occur in conjunction with severe neurological injuries or cause immediate death ( Table 8.1 ). Instability causes mechanical injury by distraction or direct compression of the spinal cord, brainstem, and cranial nerves. Ischemic or vascular injury can occur if the vertebral arteries are stretched. Extensive instability requires immediate rigid spinal fixation to immobilize the ligamentous injury. Occipitoatlantal dislocations should be fixated immediately with a halo brace. Cervical collars (such as a Philadelphia collar) are contraindicated because they reproduce the distractive mechanism of injury and can cause additional severe neurological injury. Cervical traction is likewise contraindicated because it also reproduces the mechanism of injury. Even a halo brace allows significant movement to occur in this form of injury; therefore, urgent operative stabilization is advocated if neurological function can be salvaged. Treatment consists of occipitocervical fixation and fusion.
Occipitoatlantal dislocations can be difficult to diagnose using plain radiographs. Several criteria have been used ( Fig. 8.1 ). Most patients with complete spinal cord injuries from occipitoatlantal dislocations have obvious distraction of the occipital condyles from the C1 lateral masses ( Fig. 8.2 ). However, if some spinal cord function is preserved, the alignment or gap between the occipital condyles and C1 lateral masses is usually not obvious on plain radiographs because cervical muscle spasm helps to maintain residual alignment. Despite the subtle radiographic clues, patients with incomplete neurological injuries are still highly unstable and can deteriorate neurologically. Therefore, it is important to obtain a proper diagnosis and to immobilize the head satisfactorily (i.e., halo brace) until definitive internal stabilization can be performed. Most patients with occipitoatlantal dislocation have a severe head injury that can also obscure the clinical diagnosis.
Plain radiographic diagnostic clues of occipitoatlantal dislocation include severe swelling of the prevertebral soft tissues, widening of the dens-basion distance, and a gap between the occipital condyles and upper surfaces of the C1 lateral masses. Landmarks from the skull base and upper cervical spine can be measured ( Fig. 8.1 ), but they are not sensitive or specific enough to detect all occipitoatlantal dislocations ( Table 8.2 ).
A variety of plain radiographic measurement techniques may be used to detect dislocations of the CVJ ( Fig. 8.1 ). These methods assess the relationships between the skull base and cervical spine on lateral radiographs. Wackenheim′s clival line, the dens-basion distance, the Dublin method, the X-line method, and Power′s method can be applied.7,11–19
Wackenheim′s line extends caudally along the posterior surface of the clivus.19 This line should be tangential to the posterior tip of the dens. If the occiput is displaced anteriorly, the line will intersect the dens. If the occiput is distracted or displaced posteriorly, the line will be separated from the tip of the dens. Traditionally, this technique has been used to assess basilar invagination, but it can provide a general assessment for dislocation.11,12,19
Power′s ratio assesses the relationship of two lines: the distance between the basion (B) and the posterior arch of the atlas (C) and between the opisthion (O) and the anterior arch of the atlas (A).17 In normal individuals, BC/OA averages 0.77. A ratio ≥1.0 is a fairly reliable diagnostic indicator of an anterior dislocation. This technique cannot be applied to children or individuals with congenital craniovertebral anomalies, and it can have false-negatives with longitudinal and posterior dislocations. Lee′s X-line method is similar to Power′s ratio but uses landmarks of C2 (instead of C) in relation to the basion and opisthion as references.7
The Wholey dens-basion method measures the interval between the basion and the tip of the dens in a neutral position.9,12,13,18 In adults, this distance averages 9 mm but varies considerably. Any motion between these landmarks on dynamic radiographs is abnormal.7 A distance >15 mm in adults or 12 mm in children is abnormal.7
Dublin′s method measures the distance from the posterior cortex of the ramus of the mandible to the anterior portion of C1 and C2.14 These measurements must be obtained on 72-cm radiographs with the patient′s mouth closed. This is the least reliable method of diagnosis. Normal distances to C1 range from 2 to 5 mm; normal distances to C2 range from 9 to 12 mm. This method is invalid if a mandible fracture is present, and it is unreliable with posterior dislocations.7
Plain radiographic measurement techniques are non-specific and insensitive for diagnosing dislocations for several reasons. True lateral films are needed, and it can be difficult to identify reliably the appropriate landmarks (e.g., opisthion, basion). The mastoid processes and mastoid air cells often obscure the visualization of the occipitoatlantal articular surfaces.7 At best, plain radiographic measurements detect 50 to 75% of dislocations.7 Wackenheim′s clival line and the dens-basion distance are the most sensitive measures for detection of dislocations on plain radiographs. However, only 71% of the cases were detected using the best plain radiographic measurement technique ( Table 8.2 ).
Plain radiographic methods of assessing the alignment of the CVJ have additional limitations. Each method is applicable only to specific subtypes of dislocations. None of these methods reliably detect rotational dislocations or minimally displaced subluxations. These techniques are invalid if atlas or axis fractures are present, or if the clivus, atlas, or axis is malformed.7,10–13,16,19,20
All suspected occipitoatlantal dislocation injuries should be evaluated rapidly to confirm the diagnosis. Repeat lateral cervical radiographs often display a change in alignment or distraction of the occipital condyles, especially if a cervical collar, which causes distraction ( Fig. 8.3 ), has been applied. Thin-section computed tomography (CT) with three-dimensional (3D) reconstruction can be very helpful for demonstrating a dislocated, rotated occipital condyle. Magnetic resonance imaging (MRI) is less useful because it does not clearly depict the osseous anatomy. However, it can confirm the extensive ligament and soft tissue injury in the region and assess the integrity of the spinal cord and brainstem.
Treatment of occipitoatlantal dislocations is based on the extraordinary instability of the ligament avulsions, the risk of delayed neurological injury, and the inability of ligament disruptions to spontaneously heal satisfactorily.
Attempted realignment of dislocations may cause injury and should be instituted cautiously and only under radiographic or fluoroscopic guidance. Axial loading or gentle compression of the head may reduce distractions. Some authors advocate axial traction with low weights to attempt realignment of dislocations.17,18,21–23 However, these maneuvers are dangerous. Cervical traction or cervical collars reproduce the distractive mechanism of injury, can precipitate additional neurological injury, and are contraindicated. Surgery provides a means to obtain a controlled realignment and to achieve permanent stabilization.
A halo brace alone is inadequate to maintain permanent alignment of the CVJ after an occipitoatlantal dislocation. Acute internal fixation is also needed. However, the halo brace provides a temporary, supplemental means of external stabilization until internal fixation and fusion are attained.
An aggressive operative treatment strategy is justified if patients have potentially salvageable neurological function. Normal patients or patients with incomplete neurological injuries should be treated urgently because they have a high risk of loss of neurological function due to the extensive instability. Rotational, translational, and distractive injuries are equally unstable. The extreme acute instability and the inadequacy of nonoperative therapy for ligamentous injuries justify the need for early internal fixation. A posterior occipitocervical arthrodesis should be performed for internal fixation to preserve function and to permit the maximal recovery of neurological function.
Transverse Ligament Injuries
Disruption of the transverse atlantal ligament results in anterior C1-C2 subluxation.24,25 Anterior C1-C2 subluxation, however, can also occur without a disrupted transverse ligament, as with os odontoideum or odontoid fractures. A disrupted transverse ligament is manifested by a widened atlantodental interval (ADI) on lateral cervical radiographs when the neck is flexed. When the head and neck are in a neutral or extended position, the ADI may appear normal. If the ADI exceeds 3 mm in an adult (or 5 mm in children), a transverse ligament disruption should be suspected. MRI with gradient echo sequences can be used to image directly the ligament and to assess its anatomical integrity.26 Disruption of the ligament appears as high-signal intensity within the ligament, loss of anatomical continuity, and blood at the insertion site of the ligament.
Injuries involving the transverse atlantal ligament can be classified into two distinct categories; each subtype has a separate prognosis and requires different treatments.25 Type I injuries are disruptions of the substance of the transverse atlantal ligament. Type II injuries are fractures or avulsions that detach the bony tubercle for insertion of the transverse ligament on the C1 lateral mass ( Fig. 8.4 ). These two types of injuries can be differentiated using a combination of MRI to assess the soft tissue pathology (i.e., the anatomy of the ligament) in conjunction with thin-section CT to assess the osseous pathology ( Figs. 8.5 and 8.6 ). Although plain radiographs are useful to screen for a potential abnormality, plain radiographic indices are unreliable for predicting the status of the transverse atlantal ligament because they do not directly demonstrate its anatomy ( Fig. 8.7 ).
Type I injuries are incapable of healing with an orthosis because the ligamentous substance is incapable of repair. These injuries should be treated with early surgery to fixate C1-C2 internally. Type II injuries detach the tubercle and render the transverse ligament physiologically incompetent even though the ligamentous substance is not torn.
Type II injuries have a 74% chance of healing satisfactorily when treated with a rigid cervical orthosis (halo brace) ( Fig. 8.8 ). Surgery is reserved for type II injuries that have nonunion with persistent instability after 3 to 4 months of immobilization. Type II injuries have a 26% rate of failure of immobilization; therefore, close monitoring is needed to detect patients who will require delayed operative intervention.