Odontoid Fractures

Odontoid fractures have a bimodal age distribution. Young adults break the dens in high-energy trauma; the elderly break it in a low-energy fall onto osteoporotic bone, and in the elderly the odontoid is the most common cervical fracture pattern. Every treatment decision in that older group comes back to the same balance: the risk of the operation against the morbidity of a nonunion.

Mechanism. In the young the energy comes from a motor vehicle accident, diving or sport; in the elderly, from a fall. The direction of the force predicts the direction of displacement: hyperflexion displaces the dens anteriorly, hyperextension displaces it posteriorly, and a rotational component may be present.

Associations. Look beyond the dens:

• Head injury is common with a high-energy mechanism
• Other cervical spine fractures in 10-20%
• Vertebral artery injury is rare, but assess for it
• Thoracolumbar fractures: check the whole spine

The axis. C2 is unlike any other vertebra. The odontoid process (dens) projects superiorly into the ring of C1, articulates with the anterior arch of the atlas in front, and is held there from behind by the transverse ligament. 50% of cervical rotation occurs at C1-C2.

The blood supply. The tip of the dens is fed by the apical arcade and the posterior descending arteries; the base is fed by ascending arteries from the C2 body; the body itself has the vertebral body vessels. The base of the dens, the level of a Type II fracture, is the watershed between the two supplies, so a fracture there lies in the bone with the poorest supply. That is the anatomical reason Type II fractures fail to unite so often, and why a Type I fracture through the tip and a Type III fracture through the cancellous body both heal well.

View larger

The transverse ligament (TAL) is the primary stabiliser of C1-C2. It holds the dens against the anterior arch of C1, and when it ruptures the result is atlantoaxial instability. Its integrity is read indirectly from the atlantodental interval on the lateral radiograph or CT, and directly on MRI; the numbers are in Investigations.

The alar ligaments run from the tip of the dens to the occipital condyles and limit axial rotation. A Type I fracture is an avulsion at their origin, which is why a Type I fracture is a prompt to check for occipitocervical instability.

The rule of thirds. At the level of C1 the spinal canal is divided into three: one third odontoid, one third spinal cord, one third free space filled with CSF. That free third is a physiological reserve, and it is why neurological injury is relatively rare in odontoid fractures despite what can be significant displacement: there is room for the cord to escape.

Anderson and D'Alonzo (1974) is the standard classification and it works by location: tip, base or body. Location predicts blood supply and blood supply predicts union, which is why the type frames the treatment. Grauer's modification (2005) subclassifies Type II by the orientation of the fracture line, which is what decides whether an anterior screw can work.

View larger

View larger

View larger

View larger

View larger

History. The mechanism sorts itself by age: a motor vehicle accident in the young, a fall in the elderly. Ask about neck pain, which is posterior or suboccipital; about head injury, which is common after a high-energy mechanism; and about previous neck problems, particularly degenerative disease. Neurological symptoms are rare, because the cord has space.

Examination. Keep the immobilisation on until the spine is cleared. Palpate for tenderness over the C2 spinous process, examine the cord and the roots fully, and complete the trauma survey for other injuries. The vascular examination is rarely abnormal, but it is part of the assessment.

Neurology. Neurological injury is uncommon, under 10%, despite what can be significant displacement: the canal at C1-C2 is large (the rule of thirds) and gradual displacement allows the cord to adapt. When a deficit is present, suspect cord injury; urgent decompression may be needed.

Red flags. Features that require immediate attention:

• Neurological deficit (rare but urgent)
• Rapidly progressive symptoms
• Respiratory compromise (high cord injury)
• Vascular symptoms (vertebral artery)
• Polytrauma with haemodynamic instability

CT is the gold standard for diagnosis and the source of the bony detail for classification and surgical planning. It shows the fracture type, the displacement, the angulation and any comminution, and the sagittal reconstructions are essential because they show the orientation of the fracture line, on which the Grauer subtype and the case for an anterior screw depend. Always get the CT first: MRI alone may miss a subtle fracture.

MRI is for the transverse ligament and the rest of the soft tissue. It is required when the ADI is increased or instability is suspected, it shows cord oedema when there is a neurological concern, and it is the next test when the CT findings do not explain the symptoms.

Flexion-extension radiographs assess atlantoaxial instability, and only for a fracture that has healed or whose stability has been confirmed. Never in the acute setting with an unstable fracture.

Measurements. The numbers read off the lateral film or the CT:

The ADI is the distance from the anterior arch of C1 to the dens. Under 3mm is normal in an adult and under 5mm in a child; over 3mm in an adult suggests the transverse ligament has ruptured.

View larger

The radiographic appearance of the dens can be deceptive. Several entities mimic or coexist with an acute odontoid fracture and must be actively distinguished, because the management differs substantially.

Os odontoideum is worth knowing as a condition, not just a line in the table. It is a rounded, well-corticated ossicle separated from a hypoplastic dens by a wide gap above the level of the transverse ligament, with a characteristically hypertrophied anterior arch of C1. Its origin is debated, congenital (failed fusion of the dens ossification centres) against post-traumatic (an unrecognised childhood dens fracture that healed as a rounded nonunion), and it is associated with Down syndrome, Klippel-Feil, Morquio and the skeletal dysplasias.

Two morphologies. Orthotopic, with the ossicle in the normal dens position and moving with the C1 arch; and dystopic, with the ossicle fused to the clivus or occiput, which is the more unstable of the two.

The clinical problem is atlantoaxial instability. Because the dens-equivalent is not anchored to C2, the transverse ligament has lost its post, so flexion-extension views and the space available for the cord (the posterior atlantodental interval) matter more than the ADI. An asymptomatic, stable, incidental os odontoideum may be observed with activity precautions; instability, neurological symptoms or signs, or a reduced space available for the cord mandate posterior C1-C2 fusion.

Algorithm

View larger

The decision turns on the fracture type, the patient's age and the stability of the atlantoaxial joint. Type I and Type III fractures are treated in a collar and usually unite. Type II is the controversy, because the nonunion rate with a collar is 30-40% and higher again in the elderly, so a Type II is risk-stratified: the DOGS features, comminution (IIA), late presentation and a failed collar all push it towards surgery. Whatever the type, a neurological deficit, C1-C2 instability or a transverse ligament rupture goes to urgent assessment and surgical stabilisation.

The age threshold. The DOGS factors and the summary use 65. The halo-failure series (Lennarson) and the AANS guidance both use 50, and the evidence base records that Lennarson's 50 was the dichotomy chosen rather than a tested cut-point, and that biological risk is continuous.

Non-operativeOperativeAnterior ScrewPosterior Fusion

Type I. Rigid collar for 6-8 weeks, with union in 95% or more. Because this is an alar ligament avulsion, check for occipitocervical instability before calling it benign.

Type II, selected patients. A rigid collar (Philadelphia, Miami J) or a halo, best suited to the young patient with minimal displacement (under 5mm) and a transverse line. Union is 50-60% and lower in the elderly (the union table under Outcomes gives the figures by age), so follow closely with CT at 6 and 12 weeks.

Type III. Usually non-operative unless the displacement is significant: a rigid collar for 10-12 weeks, longer because the fracture is larger, with union in 85% or more. A halo is rarely needed.

Immobilization

Collar

Moderate

Halo

Best

Comfort

Collar

Better

Halo

Poor

Complications

Collar

Fewer

Halo

Pin infections, loosening

Elderly tolerance

Collar

Better

Halo

High complication rate

Union rate (Type II)

Collar

50-60%

Halo

60-70%

Collar vs Halo

Factor	Collar	Halo
Immobilization	Moderate	Best
Comfort	Better	Poor
Complications	Fewer	Pin infections, loosening
Elderly tolerance	Better	High complication rate
Union rate (Type II)	50-60%	60-70%

Halo in the elderly - quote the right two numbers

The figures are worse than the commonly repeated "40% complications", and they are two separate numbers. In Tashjian's series (J Trauma 2006;60(1):199-203, PMID 16456456) of 78 patients over 65 with a type II or III odontoid fracture (mean age 80.7), comparing those managed with a halo vest against those managed without:

• In-hospital MORTALITY 42% with a halo versus 20% without (p = 0.03)
• MAJOR COMPLICATIONS 66% with a halo versus 36% without (p = 0.003)

There was no difference in injury severity or baseline medical condition between the groups. The mechanisms are respiratory compromise and aspiration, pin loosening and infection, falls with the added head mass and altered balance, and poor tolerance generally.

Three caveats to state alongside them. Allocation was not randomised, so some confounding by indication is inevitable - frailty, dementia, respiratory reserve and swallowing were not measured, and those are exactly what a clinician weighs when choosing a collar. The endpoint is in-hospital death, not union, so this is not evidence that a collar heals the fracture better - only that the halo costs more lives while it is on. And the comparator is mixed: the 40 "without halo" patients were 27 in a collar and 13 operated, while 4 of the 38 halo patients wore the halo after surgery - so this is halo versus collar-or-surgery, not the head-to-head against a collar that the recommendation implies. The direction of the finding is not in doubt; its precision is.

Practical position: in the patient over 75 with a type II fracture, choose a rigid collar or surgical stabilisation; the halo is now widely regarded as a relative contraindication in this group. Full detail in halo vest application and management.

Indications. Any of the following:

• Type II with high-risk features (the DOGS factors)
• Type IIA (comminuted)
• Failed collar treatment
• TAL rupture (C1-C2 instability)
• Neurological deficit
• Unable to tolerate a collar or halo

Mnemonic

DOGSType II Surgical Factors

D

Displacement

Greater than 5mm displacement

O

Old age

Age over 65 years

G

Gap/angulation

Angulation over 10 degrees

S

Smoking/comminution

Smoker or Type IIA pattern

Hook:If the patient has DOGS factors, consider surgery!

Choosing the operation. The anterior screw fixes the fracture and leaves the joint alone; every posterior technique fuses C1 to C2 and gives up the rotation. The screw therefore belongs to the acute, well-oriented Type II in good bone, and the posterior fusion to the elderly and osteoporotic, the comminuted IIA, the nonunion, the ruptured ligament, the reverse-oblique line and the late presentation.

Anterior Screw

Technique

Odontoid screw fixation

Indication

Acute Type II (no contraindications)

Advantage

Preserves C1-C2 rotation

Harms (C1-C2)

Technique

C1 lat mass + C2 pedicle/pars screws

Indication

Type II elderly, nonunion, IIA

Advantage

High fusion rate

Magerl (C1-C2)

Technique

Transarticular screws

Indication

C1-C2 instability

Advantage

Rigid fixation

C1-C2 Wiring

Technique

Brooks/Gallie

Indication

If screws not possible

Advantage

Simple technique

Surgical Approaches

Approach	Technique	Indication	Advantage
Anterior Screw	Odontoid screw fixation	Acute Type II (no contraindications)	Preserves C1-C2 rotation
Harms (C1-C2)	C1 lat mass + C2 pedicle/pars screws	Type II elderly, nonunion, IIA	High fusion rate
Magerl (C1-C2)	Transarticular screws	C1-C2 instability	Rigid fixation
C1-C2 Wiring	Brooks/Gallie	If screws not possible	Simple technique

The ideal case. An acute Type II with a transverse line or a Grauer IIB line, an intact transverse ligament, and a young patient. The screw crosses the fracture into the apex of the dens and compresses it, and because nothing is fused, C1-C2 rotation is preserved.

Mnemonic

BARRELAnterior Screw Contraindications

B

Barrel chest

Obese or kyphotic - cannot get trajectory

A

Angulation

Fracture angulated posteriorly

R

Reverse oblique

Posterior-superior to anterior-inferior line

R

Ruptured TAL

Transverse ligament incompetent

E

Elderly osteoporotic

Poor screw purchase

L

Late presentation

Chronic nonunion (over 6 weeks)

Hook:BARREL chest or reverse oblique = no anterior screw!

Late presentation. The mnemonic draws its line at 6 weeks. Apfelbaum's series fused 88% of fractures 6 months old or less and 25% of those 18 months or more; the interval between was not studied.

Technique in outline. An anterior Smith-Robinson approach under C-arm guidance in AP and lateral, and a screw across the fracture into the apex of the dens; the steps are in Surgical Technique. A single screw, properly placed, is sufficient and technically easier; Apfelbaum's finding that union was independent of the number of screws is under Classification.

View larger

Harms, the current gold standard. C1 lateral mass screws and C2 pedicle (or pars) screws, connected with a rod, with bone graft to achieve the fusion. It sacrifices C1-C2 rotation, which is half of all cervical rotation, and that is the price of its reliability.

Magerl transarticular screws pass through the C1-C2 facet joint. The fixation is very rigid, the vertebral artery risk is higher, and the anatomy has to be favourable, which the CT tells you.

Wiring (Brooks, Gallie, Sonntag) is for the patient in whom screw fixation is not possible. It is less rigid and may need a halo afterwards.

Harms

Fusion Rate

95-98%

VA Risk

Low

Key Factor

Requires C2 pedicle width over 4mm

Magerl

Fusion Rate

95-100%

VA Risk

Higher

Key Factor

Need safe trajectory on CT

Wiring

Fusion Rate

80-85%

VA Risk

Minimal

Key Factor

Less rigid, may need Halo

Posterior Technique Comparison

Technique	Fusion Rate	VA Risk	Key Factor
Harms	95-98%	Low	Requires C2 pedicle width over 4mm
Magerl	95-100%	Higher	Need safe trajectory on CT
Wiring	80-85%	Minimal	Less rigid, may need Halo

View larger

View larger

Type II nonunion. A symptomatic nonunion is confirmed on CT, more than 6 months out with no bridging bone, and then assessed: flexion-extension views for instability, and MRI for the transverse ligament. Posterior C1-C2 fusion by the Harms technique is the gold standard. An anterior screw rarely works, because the fracture site is fibrous tissue, and if the C1 lateral mass is compromised, consider extending the fusion to the occiput.

View larger

The first six weeks. Every surgical patient goes into a collar, with wound care, DVT prophylaxis and analgesia in the first fortnight. Radiographs at 2 and 6 weeks; gentle range of movement when comfortable, and physiotherapy for the surrounding muscles.

The 6-week CT decides the rest. If fusion or healing is progressing, the collar continues to the duration in the table; if it is not, the plan is reassessed. When the CT shows union or fusion, wean the collar, confirm stability with flexion-extension radiographs, and progress physiotherapy to range of movement and strengthening.

Union. The figures by type and treatment, which are the ones the treatment decision is built on:

Mortality. An elderly Type II fracture treated in a collar carries a 15-25% one-year mortality, and nonunion contributes to it; function is often limited with a fibrous union. Surgery brings its own perioperative risk, but early mobilisation reduces complications, some studies report lower mortality in the elderly, and long-term function is better if fusion is achieved.

The elderly Type II dilemma. Nonunion with a collar is frequent, the halo carries the Tashjian mortality and complication figures given under Management, and surgery carries perioperative risk. Recent evidence favours primary posterior fusion in the fit elderly patient for better union and earlier mobilisation. Be precise about survival: the AOSpine GOF cohort found a 30-day survival advantage with surgery, but its long-term mortality signal was a trend with a confidence interval crossing 1, and the evidence base below is explicit that it should not be quoted as a survival benefit.

Side-by-Side Society Guidance

Registry and Resource-Setting Notes

• No dedicated dens-fracture implant registry exists; outcome data derive from prospective cohorts such as the AOSpine North America Geriatric Odontoid Fracture (GOF) study rather than arthroplasty-style registries.
• High-resource settings: ready CT/MRI access, navigation or fluoroscopy for screw placement, and early posterior C1-C2 fixation in fit elderly.
• Limited-resource settings: reliance on plain radiographs and rigid collars, restricted MRI for ligament assessment, and a lower threshold for nonoperative management - making accurate stability assessment and falls prevention even more important.

Odontoid fracture management is one of the most genuinely contested areas in spinal trauma. A strong viva candidate can articulate where the evidence is soft and argue both sides.