Atlanto-Occipital Dislocation

Atlanto-occipital dislocation (AOD) is one of the most lethal spinal injuries. Overall mortality is 70-85 percent, and only 15-30 percent of patients reach hospital alive. Those who do need immediate recognition, because inappropriate traction or a delayed diagnosis converts a potentially survivable injury into a fatal one.

Mechanism. High-energy trauma. A high-energy motor vehicle collision is the most common adult mechanism and often comes with associated injuries; a fall from height delivers a significant axial load or hyperextension force; a pedestrian impact is a high-speed collision with direct craniocervical force. Children have a higher relative incidence, and in a child a lower-energy mechanism can produce AOD; the anatomical reasons are set out below.

What the survivor faces. Brainstem and upper cervical cord injury are common. Survivors may have cranial nerve palsies, quadriparesis or respiratory failure, and associated injuries include head injury, thoracic trauma and long-bone fractures. Survivors require a lifelong occipitocervical fusion and rehabilitation.

A ligamentous injury. The occipital condyles articulate with the superior facets of C1, and the bony craniocervical junction is intrinsically unstable: it relies almost entirely on ligaments. AOD is therefore fundamentally a soft-tissue injury. The atlanto-occipital joint depends on the paired alar ligaments, the apical odontoid ligament, the tectorial membrane and the joint capsules; once these restraints are disrupted the occiput can translate or distract relative to C1, and the vertebral arteries and brainstem are immediately at risk.

View larger

View larger

View larger

Why Type II is the most dangerous. Longitudinal distraction disrupts all the ligamentous tethers in sequence, from anterior to posterior, which is why it produces the greatest instability. The occiput separates from C1, and the brainstem and vertebral arteries are stretched.

Children. Ligamentous laxity and a larger head-to-body ratio make children more susceptible, and the fulcrum of injury is higher (C2-C3 in infants). AOD must be excluded in any child with an unexplained neurological deficit after trauma, even with a seemingly minor mechanism.

The Traynelis classification describes the direction in which the occiput has moved on C1. It is the most widely used system because the direction directly informs the reduction manoeuvre and the decision to avoid traction.

View larger

View larger

Occipital condyle fractures sit on the same craniocervical-junction spectrum as atlanto-occipital dislocation and appear in its differential. The Anderson-Montesano classification grades morphology and the Tuli classification grades stability; between them they determine whether the fracture is stable or unstable.

The Type III is the one to worry about. It is an alar-ligament avulsion, and because the alar ligament is a craniocervical-junction stabiliser, a displaced or bilateral Type III, or one with associated ligamentous injury, is unstable and may need halo immobilisation or occipitocervical fusion. Types I and II are usually stable and treated in a collar. Like AOD, occipital condyle fractures are frequently missed on plain films and need CT.

The Tuli classification grades instead by displacement and ligamentous integrity: Type 1 non-displaced; Type 2A displaced but stable; Type 2B displaced and unstable. It maps more directly onto management.

View larger

History. The mechanism is a high-energy motor vehicle collision, a fall or a pedestrian strike. A patient who was unconscious at the scene raises suspicion dramatically, and associated head, chest and abdominal trauma are common. The awake patient reports severe neck pain, occipital headache and cranial nerve symptoms.

Examination. Secure the airway first: a high cervical injury risks respiratory arrest. Examine the cranial nerves, especially IX-XII, and record motor and sensory levels. Look at the neck for swelling, step-off and crepitus, but do not manipulate it until it has been imaged, and look for associated scalp lacerations, facial fractures and long-bone injuries.

The unconscious patient. An unconscious polytrauma patient cannot report neck pain, so AOD must be excluded by CT in every high-energy mechanism patient who arrives intubated or with a Glasgow Coma Scale under 8. A normal lateral cervical radiograph does not rule it out.

The other craniocervical injuries competing for this diagnosis are developed in full elsewhere: atlantoaxial rotatory subluxation, Jefferson fracture, and, for the chronic and congenital instability that reaches clinic rather than the resuscitation room, craniocervical junction instability.

CT first. Non-contrast CT of the cervical spine is the first-line investigation in every high-energy trauma patient, especially the unconscious and those with neck pain: thin slices (1 mm or less) from occiput to T1 with sagittal and coronal reformats. CT with multiplanar reconstructions is mandatory because plain lateral cervical radiographs are unreliable, missing up to 50 percent of cases; the Powers ratio on a film requires precise identification of the basion and opisthion, which are often obscured. Any measurement exceeding its threshold requires immediate spine team notification, so document the exact values in the radiology report and communicate abnormal values immediately to the trauma and spine teams.

The order of measurement. Calculate the condyle-C1 interval bilaterally first, then the Powers ratio and the basion-dens interval; the basion-axial interval is worth recording but is poorly reproducible on reformatted CT. The condyle-C1 interval carries most of the diagnostic weight because it is the only one of these that measures the joint that is actually injured. The classic craniocervical lines are far less sensitive than their familiarity suggests.

The condyle-C1 interval. Measured on the reformatted coronal and sagittal images between the occipital condyle and the superior articular facet of C1. In Pang's 89 children the mean was 1.28 mm, no normal joint exceeded 2.5 mm, the two sides were markedly symmetrical and the value did not change with age, so abnormal widening or asymmetry signifies AOD. The normal values and the 100 percent sensitivity come from paediatric series; the use of the CCI in adults is a reasonable extrapolation rather than a separately validated threshold.

View larger

View larger

The Powers ratio. The distance from the basion to the posterior arch of C1, divided by the distance from the opisthion to the anterior arch of C1 (BC over OA), on a lateral radiograph or sagittal CT. A ratio greater than 1.0 indicates anterior displacement of the occiput on C1, the Type I pattern. Its weakness is mechanical: the ratio compares two distances that both lengthen in pure distraction, so it can stay under 1.0 while the head is lifted off the spine in a Type II injury, and it behaves erratically in the posterior Type III pattern.

The Harris rule of 12s. The basion measurements answer the question the Powers ratio cannot. The basion-dens interval (BDI), from the basion to the tip of the dens, should be under 12 mm; the basion-axial interval (BAI), from the basion to the posterior axial line (the rostral extension of the posterior cortex of the C2 body), should also be under 12 mm. Either value over 12 mm is abnormal.

The Wackenheim clivus-canal line. A line drawn along the posterior surface of the clivus and extended caudally should fall tangential to the posterior aspect of the tip of the dens. In AOD the occiput-to-dens relationship is disrupted, so the line no longer touches the dens normally; in basilar invagination it is the dens that crosses the line.

View larger

Plain-film numbers do not transfer unchanged to CT. Both 12 mm cut-offs were derived on plain radiographs. On multidetector CT, 95% of uninjured adults have a BDI under 8.5 mm, not 12 mm, so a CT basion-dens interval of 9 to 12 mm passes the classic rule while sitting outside the normal range; the error runs towards under-calling the injury, on exactly the study a trauma patient actually gets. The same series found the BAI difficult to reproduce on CT at all, and narrowed two other familiar craniocervical values: the atlantodental interval to under 2 mm rather than 3 mm, and the atlanto-occipital interval to 0.5-1.4 mm (PMID 17893223; 200 consecutive adult MDCT studies). The Powers ratio was the one measurement that did not shift between the two modalities.

The normal-CCI trap. A single normal measurement cannot overrule a globally abnormal craniocervical injury pattern, particularly when an associated C1-C2 dissociation changes the geometry. Read every CT alignment measure together with the ligament pattern rather than treating any one criterion as definitive.

View larger

View larger

CT angiography. If the CT is positive, CTA assesses vertebral artery integrity before any reduction or surgical planning. A vertebral artery occlusion or dissection alters the timing of surgery and the anticoagulation decisions.

MRI. An adjunct for the neurological deficit unexplained by CT, and for assessing ligamentous and cord injury: tectorial membrane disruption, cord oedema and brainstem signal change. The soft-tissue signs are more sensitive than any of the classic craniocervical lines, with tectorial membrane damage at 71%, extra-axial blood at C1-C2 at 75% and perimedullary subarachnoid blood at 63%, but each carries a false-positive rate of 25 to 40%, so they raise or lower suspicion rather than settle it (PMID 18091277). MRI must never delay collar immobilisation, the CT or spine-team notification; the reformatted CCI remains the measurement that makes or refutes the diagnosis.

View larger

The goal. Prevent further distraction or translation while obtaining definitive imaging, then secure the junction. Rigid neutral immobilisation from the scene onward is the single most important intervention.

Initial StabilisationOccipitocervical Fusion

At the scene. A rigid cervical collar applied in the neutral position, log-roll precautions with manual in-line stabilisation, airway management without neck extension if possible, and rapid transport to a trauma centre with spine capability.

In the emergency department. Maintain neutral alignment and do not remove the collar until the CT has been reviewed. CT of the cervical spine within 30 minutes of arrival; calculate all the measurements (Powers ratio, CCI, BDI, BAI) and notify the spine team immediately if any threshold is exceeded.

After CT, by type.

• Type I and III: gentle closed reduction under fluoroscopy if displaced, then halo or collar
• Type II: strict neutral position, no traction attempted
• All types: occipitocervical fusion within 24-72 hours once medically optimised

Never Apply Traction Before Imaging

Apply a rigid cervical collar or halo ring in the neutral position only, log-roll with in-line stabilisation, and obtain urgent CT from occiput to T1 with sagittal and coronal reconstructions. Traction is absolutely contraindicated until imaging has excluded a distraction-type injury, and in Type II it can be fatal: applying weight can convert a partial ligamentous disruption into complete separation, increase the longitudinal gap, stretch the brainstem and cause immediate cardiorespiratory arrest and death. If the CT demonstrates vertical separation, maintain the neutral collar or halo without any distractive force and make no attempt at closed reduction using traction; reduction, if required, is performed surgically after exposure.

Why fusion. The injury is purely ligamentous and will not heal in a collar or a halo alone, so rigid internal fixation of the craniocervical junction after reduction is unavoidable. Early fusion secures the junction and is what allows a survivor to rehabilitate. The construct options, the screw corridors and the price paid in lost occipito-cervical motion are set out in occipitocervical fusion.

Surgical Protocol

PlanningPre-operative

• Reduction: anatomic alignment with halo or manual manipulation under fluoroscopy (Types I and III)
• Confirm vertebral artery patency on CTA
• Approach: posterior midline from occiput to C3-C4
• Fixation: occipital plate with C1 lateral mass and C2 pedicle screws (or transarticular)

Operating RoomSurgical Steps

• Prone on a Jackson table with a Mayfield clamp maintaining neutral alignment
• Subperiosteal dissection of the occiput, C1 and C2
• Direct manipulation of the occiput under fluoroscopy if still displaced
• Occipital screws (midline keel), C1 lateral mass screws, C2 pedicle screws, rod connection
• Iliac crest autograft or allograft between the occiput and C1-C2

RecoveryPost-operative

• Halo or rigid collar for 8-12 weeks
• Log-roll precautions until fusion is visible on CT
• Flexion-extension radiographs at 3 months to confirm stability

Occiput-C1-C2 with screws

Indication

Standard adult AOD

Advantage

Rigid three-point fixation

Risk

Vertebral artery injury at C2

Occiput-C1-C2-C3 extension

Indication

Osteoporosis or poor C2 bone

Advantage

Additional fixation points

Risk

Longer fusion, reduced motion

Paediatric wiring techniques

Indication

Small C1 lateral mass in children

Advantage

Avoids screw placement in small bone

Risk

Less rigid, requires halo longer

Fixation Options for Occipitocervical Fusion

Construct	Indication	Advantage	Risk
Occiput-C1-C2 with screws	Standard adult AOD	Rigid three-point fixation	Vertebral artery injury at C2
Occiput-C1-C2-C3 extension	Osteoporosis or poor C2 bone	Additional fixation points	Longer fusion, reduced motion
Paediatric wiring techniques	Small C1 lateral mass in children	Avoids screw placement in small bone	Less rigid, requires halo longer

View larger

View larger

View larger

View larger

By type. No source separates survivorship by Traynelis type. In a Type I or III injury diagnosed early, the cranial nerve and cord injury is variable and long-term function is good with a successful fusion and rehabilitation. In Type II the mechanism stretches the brainstem directly, so the rate of brainstem injury is high and long-term function is poor.

Children. Relative survival is higher: in Pang's 16 AOD patients, 11 (69%) made an excellent neurological recovery. Fusion is still required, with growth to consider.

Two numbers to distrust. Type-specific survival figures in circulation ("20-30% for Type I/III, under 10% for Type II") have no source that separates survivorship by Traynelis type; quote overall survival to hospital and the mechanism instead. And early fusion is credited with "improving survival" only by selection: it is performed on patients who have already survived to theatre, and no series compares it against non-fusion in matched patients. What early stabilisation changes is the neurological outcome and the security of the junction in a survivor, not the probability of surviving. Saying it improves survival is a common and unearned extrapolation.