Atlantoaxial (C1-C2) Fusion

Scope note. This page is the C1-C2 fusion technique guide — indications, vertebral-artery anatomy, and the Magerl-versus-Goel-Harms decision. For the atlantoaxial pathology that drives the operation (rheumatoid involvement, instability, Crowned Dens, the Ranawat classification), see atlantoaxial arthritis.

C1-C2 fusion is offered when the atlantoaxial segment is unstable, painful beyond salvage, or destroyed, and when the pathology is confined to C1-C2. Exclude cranial settling, occipital condyle deficiency and subaxial instability before committing to the segment: extension to the occiput or subaxial spine converts the plan to an occipitocervical or extended posterior construct.

Type II odontoid fracture. In the Anderson-D'Alonzo type II fracture the operative candidates are displacement greater than 5 mm, angulation greater than 11 degrees, comminution, a posterior oblique fracture line, and age greater than 50 years, beyond which nonunion risk rises steeply. An anterior odontoid screw preserves rotation but demands a favourable anterior-oblique fracture line, an intact transverse ligament, a reducible acute fracture and adequate bone density; if any criterion fails, the operation is a posterior C1-C2 fusion.

Odontoid nonunion. An established nonunion is a posterior fusion indication. The anterior screw has no role once the fracture surfaces are sclerotic.

View larger

Transverse ligament rupture. The atlantodens interval (ADI) is drawn on the lateral radiograph from the posterior cortex of the anterior C1 arch to the anterior dens cortex at their midpoints. An ADI greater than 3 mm in adults (greater than 5 mm in children) implies transverse ligament incompetence, and greater than 5 mm in an adult implies additional alar and apical failure. An intrasubstance ligament rupture (Dickman type I on MRI) does not heal and is fused; a bony avulsion of the ligament insertion (type II) may heal in a rigid orthosis with close surveillance.

View larger

Irreducible or recurrent atlantoaxial rotatory fixation. Paediatric cases failing traction and recurring after reduction are fused.

Os odontoideum. Instability on flexion-extension films, cord signal change, or myelopathy mandates fusion. The asymptomatic stable os remains controversial (see controversies).

Rheumatoid atlantoaxial instability. Fuse for progressive ADI widening, myelopathy, intractable occipitocervical pain, or a posterior atlantodens interval (PADI) less than 14 mm; PADI is the more predictive measurement because it is the space available for the cord. Operate before fixed cranial settling develops, which converts a C1-C2 fusion into an occipitocervical construct.

View larger

C1-C2 osteoarthritis. Unilateral lateral mass arthrosis causing intractable rotational occipitocervical pain, confirmed by image-guided joint block and refractory to conservative care.

Tumour or infection. Destruction of the dens, C1 lateral mass or C2 body with instability, often combined with anterior debridement and decompression.

Basilar invagination. Where the invagination is reducible, the Goel philosophy is joint distraction with intra-articular spacers and C1-C2 fixation.

The V3 segment. The vertebral artery exits the C2 transverse foramen, loops laterally then superiorly through the C1 transverse foramen, and turns medially in the groove on the superior surface of the C1 posterior arch, roughly 1.5 cm from the midline, before piercing the atlanto-occipital membrane and dura. That groove sets the exposure limit: subperiosteal dissection is safe to about 12-15 mm from the midline on the superior surface of the arch and 8 mm on the inferior surface. Dissect strictly subperiosteally and keep cautery off the superolateral arch; straying superolaterally with cautery or a periosteal elevator is the classic mechanism of exposure-phase vertebral artery injury.

Ponticulus posticus. The arcuate foramen is reported in roughly 3 to 15 percent of people depending on the series and the imaging method: an ossified bridge over the vertebral artery groove that mimics a broad posterior arch on lateral fluoroscopy. A "posterior arch" C1 screw started on the bridge drills straight into the artery. Identify it on the preoperative CT and lateral radiographs, and do not judge the entry point from arch thickness on fluoroscopy alone.

View larger

The internal carotid artery. It lies immediately anterior to the C1 lateral mass, within a few millimetres in some patients, often anteromedial and asymmetric. Bicortical C1 screws improve pull-out but risk the carotid and the hypoglossal nerve, so most surgeons stop just short of the anterior cortex, judged on the lateral fluoroscopic view and the preoperative CT angiogram.

C2 and the high-riding artery. The vertebral artery runs through the C2 transverse foramen beneath the pars and pedicle. A high-riding vertebral artery, a large, cephalad and medially looping vessel that narrows the C2 isthmus (an internal height less than 5 mm or an isthmus width less than 4 mm is a common threshold), is found on at least one side in up to 20 percent of patients and contraindicates a transarticular or pedicle screw on that side. Paramore's 94 CT studies put it at 18 percent prohibited on at least one side, 18 to 23 percent once borderline anatomy is counted, and bilateral in only 3 patients. The anomaly is usually unilateral, which is why the construct is chosen side by side.

View larger

View larger

View larger

C2 terminology. The pars interarticularis lies between the superior and inferior articular processes; the pedicle connects the posterior elements to the body and sits medial and cephalad to the artery. The pedicle screw trajectory is more medial (about 20-30 degrees) and more cephalad than the pars screw.

The C2 nerve root. It crosses the C1-C2 joint posteriorly, in the greater occipital nerve territory, embedded in a rich venous plexus directly over the C1 lateral mass screw entry. That plexus is the source of the notorious bleeding at that step.

Preoperative imaging. CT angiography is mandatory before any C1-C2 screw construct: map the artery on both sides, measure C2 isthmus height and width on sagittal and axial reconstructions, simulate the Magerl trajectory on the sagittal cut through the isthmus, and note dominance or hypoplasia, anomalous V3 loops, a ponticulus posticus and an anomalous carotid position anterior to C1. Template the screw trajectories and lengths on the CT before incision; the trajectory is chosen to fit the artery, never the reverse. Flexion-extension lateral radiographs quantify reducibility, and MRI assesses cord signal, rheumatoid pannus and transverse ligament integrity by the Dickman classification.

The wiring techniques are now adjuncts or salvage, and the choice today is between Magerl transarticular screws and the Goel-Harms screw-rod construct, made on reducibility and the vertebral artery rather than habit.

Choosing the construct. Magerl screws are drilled along a fixed trajectory across a joint that must already be reduced. The Goel-Harms construct places each screw independently and reduces at the screw heads before the rods are locked, so it serves a fixed or irreducible subluxation, and it remains feasible when a high-riding artery or a narrow isthmus prohibits a transarticular screw on one or both sides. Biomechanically it is equivalent to transarticular fixation.

When one side is hostile. A high-riding artery or a pars too small for a safe pedicle or transarticular screw takes a C2 translaminar screw on that side (or bilaterally), or the construct becomes a unilateral screw plus contralateral translaminar hybrid. Fixation is achieved without violating the artery; accept lower rotational stiffness and check laminar thickness.

View larger

View larger

Positioning and preparation. Prone on a radiolucent table with the head in a Mayfield clamp, the neck neutral or slightly flexed at C1-C2 to open the interlaminar interval but with overall alignment set to the intended fused position; reverse Trendelenburg reduces venous bleeding. Awake fibreoptic intubation and an awake positioning check are considered in myelopathic or grossly unstable patients, with neuromonitoring (SSEP and MEP) and a baseline taken before positioning in myelopathy. Lateral fluoroscopy confirms reduction after positioning and before incision, with navigation if available.

In theatre, recheck the CT angiogram, mark the planned C2 screw type for each side, and cross-match blood because the venous plexus can bleed briskly. Have bipolar diathermy, flowable haemostatic matrix and a 3.5 mm polyaxial screw system with smooth-shank C1 options open.

Exposure. A midline posterior incision from occiput to C3 through the avascular median raphe, then subperiosteal exposure of the C1 posterior arch within the 12-15 mm superior and 8 mm inferior lateral limits, followed by the C2 lamina, pars and inferior articular processes. Preserve the C2-C3 facet capsule and the semispinalis cervicis insertion on C2 where possible.

The C2 root and venous plexus. Dissect along the inferior border of the C1 arch onto the lateral mass; the root and its plexus are met here. Control bleeding with bipolar and haemostatic matrix and do not linger: prolonged dissection before screw placement worsens blood loss, whereas placing the C1 screw early tamponades the plexus.

Sacrifice or retract the root? Sectioning the root improves exposure of the joint for decortication and grafting and reliably stops the plexus bleeding, at the cost of occipital numbness and occasional neuralgia or dysaesthesia. Caudal retraction preserves the root but limits joint access. Both are defensible; know the debate.

C1 lateral mass screw. Entry is at the centre of the lateral mass at its junction with the inferior aspect of the posterior arch, with a trajectory 10-15 degrees medial and cephalad, aiming at the anterior arch tubercle on lateral fluoroscopy and stopping just short of the anterior cortex because the internal carotid lies in front. A smooth, unthreaded proximal shank leaves the screw standing proud of the plexus and the C2 root and reduces irritation. Where the arch is thick, notching the inferior arch (Tan modification) or a posterior arch screw are alternatives; beware the ponticulus posticus.

View larger

C2 screw, chosen to fit the artery. Three trajectories:

• Pedicle screw, the strongest: entry in the upper-outer quadrant of the isthmus surface, the medial pedicle border palpated with a Penfield in the canal, then 20-30 degrees medial and 20-25 degrees cephalad along the pars into the body. Contraindicated with a high-riding artery on that side.
• Pars screw: the same entry, a steeper cephalad and less medial trajectory, and a shorter screw stopping before the transverse foramen. Safer with moderate arterial risk, but less purchase.
• Translaminar screw (Wright): the artery-independent bailout. Entry at the junction of spinous process and lamina, the screws crossing within the laminae to the contralateral side, offset cephalad and caudad so that they can cross. It risks ventral laminar breach into the canal, is biomechanically weaker, and occupies the laminar graft bed.

View larger

View larger

Reduction and rod locking. Connect the rods and, where reduction is needed, lever C1 posteriorly or anteriorly against the fixed C2 screws before final tightening; confirm restoration of the atlantodental interval on lateral fluoroscopy. Never drill Magerl or pedicle trajectories with C1 unreduced.

Fusion bed, grafting and closure. Decorticate the inferior C1 arch, the C2 lamina and, for the highest fusion rates, the C1-C2 joint itself: distract, curette the cartilage and pack cancellous graft or a structural spacer. Goel joint spacers additionally distract and reduce basilar invagination. Lay corticocancellous autograft, structural or morcellised, with iliac crest remaining the gold standard, as an interlaminar strut with or without a supplemental cable, or as onlay graft. Layered closure over the muscle; a drain is optional.

View larger

View larger

Aftercare. A rigid or semi-rigid collar for 6 to 12 weeks, though rigid segmental fixation permits a soft collar for comfort in many units. Upright lateral radiographs before discharge, radiographs at 6 weeks and 3 months with flexion-extension films to confirm fusion, and CT at 6 to 12 months if fusion is in doubt or symptoms persist.

Salvage. Nonunion or screw failure is revised with the alternative C2 trajectory, joint grafting if not previously done, or extension to occipitocervical fusion if C1 purchase is lost.

Reduction first. Transarticular screws are drilled along a fixed trajectory: if C1 is subluxed on C2 the screw will miss the C1 lateral mass and may enter the vertebral artery or the hypoglossal canal. Confirm reduction on lateral fluoroscopy before drilling, reducing with positioning, Mayfield adjustment or a temporary clamp. If the joint is irreducible, abandon Magerl and use Goel-Harms with intraoperative joint reduction instead.

Trajectory. Entry is 2-3 mm superior and 2-3 mm lateral to the medial edge of the C2-C3 facet on the C2 inferior articular process. The trajectory is strictly sagittal (0-10 degrees medial) and steeply cephalad along the C2 isthmus, across the joint and into the C1 lateral mass, aiming at the anterior arch tubercle. That steepness often requires percutaneous stab incisions near the cervicothoracic junction, and a thoracic kyphosis or a barrel chest can make the trajectory unachievable, so simulate it on the preoperative sagittal CT.

The graft is not optional. Always supplement with a posterior interlaminar graft held by wiring or a clamp (a Gallie-type graft): screws alone leave the graft uncompressed and fusion rates fall. A unilateral screw plus posterior wiring is an accepted compromise when only one side has a safe arterial course.

View larger

View larger

View larger

Vertebral artery injury. Roughly 2 to 4 percent of patients with transarticular screws, lower with modern planning. Brisk arterial bleeding during drilling or tapping is met by completing the hole and placing the screw, which tamponades the vessel; the contralateral screw is not attempted, the wound is closed, and immediate angiography characterises occlusion versus pseudoaneurysm. The table below carries the rest of the prevention and management for this and every other structure at risk.

Nonunion. Under 5 percent with segmental constructs and joint grafting; inadequate joint preparation leads to it, and salvage is described with the technique above.

Loss of rotation. C1-C2 contributes roughly half of cervical axial rotation, and fusion sacrifices it permanently; counsel drivers and athletes preoperatively and document it. Adjacent segment strain at the occipito-atlantal and subaxial levels can follow over years.

View larger

Related pages: Atlantoaxial Instability is the indication this operation exists to treat and carries the ADI and space-available-for-cord thresholds that decide who is offered it; Craniocervical Junction Instability is the wider assessment, and Atlantoaxial Rotatory Subluxation is the paediatric presentation that is usually treated without fusion. Odontoid Fractures supplies most of the trauma workload here, and is where the anterior odontoid screw competes as the motion-preserving alternative. Vertebral Artery Injury is the complication every trajectory decision on this page is made to avoid. When C1 will not hold a screw or the deformity extends above it, the operation becomes Occipitocervical Fusion, and Basilar Invagination is the deformity that drives joint distraction rather than in-situ fixation. Atlantoaxial Arthritis is the degenerative indication, while Down Syndrome Orthopaedic Manifestations and Mucopolysaccharidoses are the two syndromic groups in which the instability is found on surveillance rather than after injury. Halo Vest Application and Management covers the external immobilisation these constructs were designed to make unnecessary.

• AO Spine upper cervical classification and knowledge forum publications frame odontoid fracture and atlantoaxial injury management; operative thresholds for type II fractures (displacement greater than 5 mm, angulation greater than 11 degrees, age greater than 50, comminution) are widely adopted internationally.
• Congress of Neurological Surgeons guidelines on odontoid fractures acknowledge equipoise between rigid immobilisation and surgery in selected type II fractures but note high nonunion rates with non-operative care in the elderly.
• Global practice variation: modern segmental screw-rod constructs dominate in well-resourced centres; posterior wiring with halo immobilisation remains a legitimate, effective option where fluoroscopy, navigation, or implant availability is limited — examiners expect candidates to justify technique choice by resources as well as anatomy. In parts of Asia, Goel joint-spacer techniques for basilar invagination and congenital craniovertebral anomalies are a major practice stream reflecting regional case mix.
• Epidemiology: odontoid fractures are the commonest cervical fracture in the elderly worldwide, driven by low-energy falls in ageing populations; rheumatoid atlantoaxial instability is declining in incidence in regions with early access to disease-modifying therapy, but remains prevalent where such access is limited.
• No arthroplasty registry applies; fusion outcome data derive from institutional series and multicentre AO Spine cohorts, consistently reporting fusion above 95 percent for segmental constructs.

• C2 nerve root sacrifice versus preservation: sacrifice improves joint exposure and haemostasis with mostly well-tolerated occipital numbness; preservation avoids sensory loss but limits joint preparation. Comparative series show broadly similar patient satisfaction — both positions are defensible in a viva.
• Asymptomatic stable os odontoideum: prophylactic fusion versus surveillance remains unresolved; most fuse once instability, cord signal change, or symptoms appear.
• Type II odontoid fracture in the very elderly: operative fixation versus accepting a fibrous nonunion in a collar — surgery improves union and possibly survival in fit patients, but perioperative risk is substantial in the frail; shared decision-making is the examinable answer.
• Bicortical versus unicortical C1 screws: bicortical purchase is stronger but risks the ICA; most surgeons accept unicortical fixation within a rigid segmental construct.
• Joint fusion versus posterior onlay only: direct C1-C2 joint decortication and grafting (Goel philosophy) probably maximises fusion, at the cost of more plexus bleeding and root handling.
• Navigation and robotics reduce breach rates in anomalous anatomy in cohort data, but fluoroscopy-guided freehand technique by experienced surgeons remains standard in much of the world.