Secondary Logo

Atlantoaxial Rotatory Subluxation in Children

Neal, Kevin M. MD; Mohamed, Ahmed S. MD

JAAOS - Journal of the American Academy of Orthopaedic Surgeons: June 2015 - Volume 23 - Issue 6 - p 382–392
doi: 10.5435/JAAOS-D-14-00115
Review Article
Free
SDC

Atlantoaxial rotatory subluxation is a rare condition in which patients present with the acute onset of torticollis. Atlantoaxial rotatory subluxation represents a spectrum of disease from muscle spasm to a fixed mechanical block to reduction of the atlantoaxial complex. If left untreated, some cases may resolve spontaneously; however, other cases may result in the development of secondary changes in the bony anatomy of the atlantoaxial joint, leading to persistent deformity. Diagnosis of the condition is largely clinical but can be aided by various imaging modalities, including radiographs, dynamic CT scanning, three-dimensional CT reconstructions, or MRI. Consideration should always be given to infection or other inflammatory disease as an underlying, precipitating cause. Treatments include observation, the use a cervical collar and analgesics, halter or skeletal traction, and posterior fusion of C1-C2. The most important factor for success of conservative treatment is the time from the onset of symptoms to recognition and the initiation of treatment.

The human cervical spine naturally allows a significant amount of motion. Up to 60% of cervical rotation occurs at the atlantoaxial joint (C1-2).1 Children, or less commonly adults, may present with an apparent fixed rotation of the atlantoaxial joint, manifesting as an acute, often painful torticollis. The exact incidence is unknown, but it appears to be relatively rare. Because of this rarity, the nomenclature, etiology, natural history, and preferred treatments have been based predominantly on case reports and case series. The condition has been referred to by various names, including acute acquired torticollis,2 atlantoaxial rotatory subluxation,3 atlantoaxial rotatory fixation,4 atlantoaxial rotatory dislocation,5 or similar variations. When the condition is attributable to infectious or postoperative retropharyngeal inflammation, it is referred to as Grisel syndrome.6 Despite the impression that the various names refer to different etiologies or severities, it is likely that they are all a part of a spectrum of disease referring to the same condition. For simplicity, we prefer the term atlantoaxial rotatory subluxation (AARS). AARS is a rotation of the atlantoaxial complex that is held in a fixed position, the result of either a muscle spasm or a mechanical block to reduction.

Back to Top | Article Outline

Anatomy of the Atlantoaxial Joint

The first and second cervical vertebrae represent a unique complex in the spine that allows for an increased amount of rotation compared with the rest of the spinal column. The atlas (C1) does not have a vertebral body; it articulates superiorly with the occiput and is formed from three cartilaginous ossification centers.7 The C1 ring surrounds the dens of the axis (C2). Similarly, the axis develops from five ossification centers: two for the neural arches, a basal central ossification center, a dentate center, and an apical center.8 The key ligament that stabilizes the atlantoaxial complex is the transverse ligament, which represents the thickest, strongest portion of the cruciform ligament. The transverse ligament runs posterior to the odontoid process, between the lateral tubercles of C1, locking the odontoid against the anterior arch of C1. The ligament has a smooth fibrocartilaginous surface that allows easy gliding of the odontoid during rotation. A synovial capsule is located anteriorly between the odontoid and the transverse ligament, with the tectorial membrane, epidural fat, and dura mater covering the spinal cord posteriorly9 (Figure 1). Synovial folds exist in some atlantoaxial joints, and their presence may be increased in children.10 This abundance of soft-tissue structures adjacent to the mobile C1-2 articulations may provide material that can become swollen or incarcerated during atlantoaxial motion, leading to the fixed subluxation of AARS.

Figure 1

Figure 1

The normal cervical spine is capable of approximately 180° of rotation. When neck rotation is initiated, C1 moves first, and C2 remains stationary until C1 has rotated 23°. After 23°, C2 begins to rotate, but C1 continues to move at a greater rate, such that the angular relationship between C1 and C2 continues to increase. After C1 has reached about 65° from midline, the ligaments between C1 and C2 become taut, producing a “yoking effect,” and C1 and C2 rotate together until maximum head turn. This difference in rates of motion between the axis and the atlas produces a natural subluxation of the C1-2 facets during normal rotation.1 The key finding that differentiates AARS from normal cervical motion is that the subluxation of the C1-2 facets is prevented from returning to normal, either by muscle spasm or by a mechanical block to reduction.

Back to Top | Article Outline

Causes

The onset of AARS has been attributed to various pathologies that fall into the broad categories of trauma or inflammation in the head and neck region. Traumatic etiologies include injuries that are severe enough to cause fractures in the head and neck region, including C1 fractures,11 odontoid fractures,12 lower cervical spine fractures,13 humerus fractures,3 and clavicle fractures.11 However, AARS may occur after only minor trauma, such as short falls or bumping the head.14 Surgical procedures may produce minor trauma from manipulation of the cervical spine during anesthetic induction or during the surgery itself. Cases have been reported following surgeries performed on the head or neck,2,13,15 as well as procedures performed in locations more remote from the atlantoaxial complex.16 Even minimal trauma, such as from application of an orthotic device or removal of a body cast, has been implicated as an instigating factor.4

Inflammation preceding AARS appears to be common. Presumably any condition that results in inflammation in the head or neck may be implicated. Infections, whether bacterial or viral, are especially common. Case reports document AARS following upper respiratory tract infections,2-4,13,14,17-20 retropharyngeal abscesses,17,21,22 pharyngitis,21 cervical adenitis,15,17 sinusitis,17 tonsillitis,11,17 mastoiditis,17 and otitis media.13,20 AARS has also been associated with Sandifer syndrome,17 which is characterized by spasmodic posturing of the neck, back, or upper extremities associated with gastroesophageal reflux, esophagitis, or a hiatal hernia.

AARS has been described in patients with autoimmune disorders, including juvenile idiopathic arthritis,3,23 ulcerative colitis,15 and HLA-B27-positive seronegative spondyloarthropathy.24

In 1930, Grisel6 described AARS related to inflammation, resulting in the term Grisel syndrome. Head or neck inflammation may lead to AARS by spread of the inflammation to the joints and ligaments of the atlantoaxial complex via the connection between the lymphatic and the venous systems.25 Inflammation around the atlantoaxial complex, including the synovium, may cause involuntary muscle spasm, leading to the characteristic torticollis. It is possible that ligamentous laxity is also induced by the inflammation.

In the 1930s, Jones26 suggested that inflammation could decalcify the bony attachments of the transverse ligament, leading to increased stretch or rupture. The theory of ligamentous laxity is strengthened by reports of AARS occurring after surgical procedures or after minor trauma in patients with Marfan syndrome.16

Less commonly reported causes of AARS include congenital abnormalities. Lin et al27 reported one case in which odontoid hypoplasia predisposed a patient to AARS, although inflammation was the actual inciting event. In another case, a patient became quadriparetic following thoracic and lumbar spine fusion. The patient was subsequently noted to have Klippel-Feil syndrome, with congenital atlanto-occipital and C2-C3 fusions, resulting in a predisposition to AARS from manipulation of the neck for anesthesia and positioning for the spine procedure.28 Other reports have also suggested C1-occiput fusions as contributing factors.11,23 Although the presence of congenital abnormalities may increase the risk for AARS, some inflammation or minor trauma is still the most likely inciting event.

Despite the inciting event, the underlying pathologic mechanism is discomfort in the atlantoaxial complex that leads to muscle spasm. Presumably, the spasm is an unconscious mechanism that relieves pain or pressure from edema created in the atlantoaxial complex, either from trauma, inflammation, or entrapped soft-tissue structures.

Back to Top | Article Outline

Classifications

Classifications of AARS have been based on the duration of symptoms and imaging parameters. No single definition exists to determine when AARS transitions from an acute condition to a chronic condition. Ishii et al14 classified patients as acute when symptoms were present for <8 weeks; patients were classified as chronic when symptoms were present for >3 months. Pang and Li13 classified patients as acute when symptoms had been present for <1 month, as subacute from 1 to 3 months, and as chronic for >3 months. Likewise, because the presence of torticollis is based on physical examination, no single imaging modality is recommended to diagnose and classify AARS. Most imaging-based classifications have been based on CT scans, and although CT scanning is not mandatory to establish the diagnosis, these classifications may provide insight into the underlying pathoanatomy of the condition.

Fielding and Hawkins4 classified AARS into four types based on axial CT images of C1 and C2. Type I is a unilateral facet subluxation with an intact transverse ligament; no displacement occurs between the anterior arch of C1 and the dens. The dens acts as a pivot, with one C1-2 facet subluxating anteriorly and the other facet subluxating posteriorly (Figure 2). Type I is the most common type noted in the authors’ case series. Although their images show potential spinal canal compromise for patients with type I AARS, the authors feel that it was the most benign type, without deficiency in the transverse ligament. Therefore, the authors recommend that patients be treated more or less expectantly. Type II AARS is a unilateral facet subluxation with anterior displacement measuring 3 to 5 mm between the arch of C1 and the dens; one facet subluxates and the other facet remains located and acts as the pivot point. The authors note more potential for canal compromise in type II AARS because of the deficiency of the transverse ligament. Type III AARS is bilateral anterior facet displacement; the interval between the C1 arch and the dens measures >5 mm. Type IV is the most unusual type noted in their series; it consists of a deformity with the atlas displaced posteriorly, which must be associated with a deficiency of the dens. In the only type IV case, the patient had rheumatoid arthritis and an absent dens. Presumably, trauma producing a fracture of the odontoid process could have a similar appearance.13 The authors consider types III and IV forms of C1-2 dislocation, and therefore potentially catastrophic. It seems reasonable to categorize Fielding and Hawkins types III and IV as distinct entities from AARS and consider them to be the major cervical spine dislocations that they represent. It is likely that most cases of AARS seen in children would correspond to Fielding and Hawkins type I (Figure 2).

Figure 2

Figure 2

McGuire et al29 recommended a classification based on dynamic CT scanning. Patients were classified as stage 0 when dynamic CT scanning was normal. Patients were classified as stage 1 when there was <15° of difference between C1 and C2 rotation, but C1 crossed the midline. In stage 2, C1-C2 movement is fixed. Importantly, the authors noted that the severity of AARS increased with the duration from the onset of symptoms to presentation, with an average of 6.7 days for stage 0, 8.6 days for stage 1, and 20 days for stage 2.

Pang and Li1,13 established a new classification in which they plotted C1-2 motion curves generated from dynamic CT scanning of AARS patients; this information was then compared with normal control subjects. CT scans were obtained with the head in the presenting position, with the nose pointing straight forward (ie, neutral), and with the head turned as far to the opposite side as comfort would allow. In group 1, the maximally corrected C1-2 angle decreased <20% from the presenting position. This indicated nearly coupled C1-2 motion or locking of the atlantoaxial complex. In group 2, there was reduction of the C1-2 angle >20% from the presenting position. However, there was insufficient motion for C1 to cross C2 and rotate in the opposite direction. In group 3, C1 rotated to the opposite side, but yoked C2 along with it, such that C2 rotated an abnormal amount to the corrected side. Group 4 showed C1-2 motion within normal parameters. Group 5 was a diagnostic gray zone in which there was only mild abnormal coupling of C1-2 motion; the authors referred to this as stickiness.

All groups from Pang and Li’s classification would be considered Fielding and Hawkins type I because no mention was made of any increasing distance between the arch of C1 and the dens. Pang and Li1,13 derived their series from patients who presented after minor trauma or surgery; cases that were precipitated by an inflammatory process were excluded. No osseous abnormalities were noted in the images, but the authors speculated that group 1 patients may have had dislocated and locked facets, cartilage tears that blocked reduction, or buttonholing of part of C1 through a capsular rent. The authors suggested that groups 2 and 3 represented less severe soft-tissue impingement, such as capsular invaginations or synovial folds, and that groups 4 and 5 were the result of muscle spasm alone. In a subsequent paper, Pang30 noted that patients with symptoms for >3 months that were in group 1 were more likely to have recurrent slippage, undergo more prolonged treatment, and require more aggressive therapies, such as halo traction or surgery. The patients were also more likely to experience permanent loss of neck rotation. The graphic representation from Pang and Li13 aids in the understanding of normal C1-2 motion and depicts how the motion becomes more coupled as the severity of the AARS progresses (Figure 3).

Figure 3

Figure 3

Ishii et al14 used three-dimensional CT scans to classify AARS based on the lateral inclination of the atlas on the axis and the presence of C2 facet deformity (Figure 4). In patients who presented within 8 weeks of the onset of torticollis, no facet deformities were noted; these patients were classified as grade I. In patients presenting >3 months from symptom onset, those with moderate facet deformity with <20° of C1 inclination were classified as grade II, and those with severe facet deformity with >20° of C1 inclination were classified as grade III. Ishii et al14 were able to document that the progressive dysplastic changes were reversible with appropriate treatment.

Figure 4

Figure 4

Back to Top | Article Outline

Presentation

The diagnosis of AARS is based on the presence of a new, fixed torticollis in a patient with no preexisting history of congenital muscular torticollis, significant trauma, or congenital abnormality. The patient’s head is held in the classically described cock robin position with the head tilted to one side and with the chin rotated to the side opposite the facet subluxation4 (Figure 5). The patient may have pain in the neck and jaw and resists touch by the examiner. In some patients with AARS, sternocleidomastoid spasm occurs on the side of the chin; Phillips and Hensinger3 attributed this finding as a means to reduce the facet subluxation. This sign is in contrast to congenital muscular torticollis, in which the sternocleidomastoid spasm is opposite the chin deviation.

Figure 5

Figure 5

Back to Top | Article Outline

Diagnostic Testing

Because many cases of AARS are caused by inflammation of the head and neck, consideration should be given to testing for infection or inflammatory disease. Laboratory studies for generalized inflammation, including infection, include a complete blood count, erythrocyte sedimentation rate, and C-reactive protein level.31

Because many patients with AARS present without any intrinsic bony abnormality,2 and subluxation noted on imaging often remains in the normal range,13 AARS is primarily a clinical diagnosis. However, several imaging modalities may be useful to aid in classification, to rule out trauma and inflammatory conditions, and to aid in making treatment recommendations.

Radiographs play a role as the initial imaging modality, mainly to rule out obvious fracture or congenital abnormality.28 The open-mouth odontoid view, in particular, may show displacement of the odontoid process or overlapping of the lateral masses32 (Figure 6). However, head positioning and limited range of motion associated with torticollis may make radiographs of the cervical spine inaccurate and difficult to interpret.33 Head tilt may obscure the normal anatomy and landmarks of the vertebrae on the lateral radiograph. The odontoid view is best obtained with the head in the neutral position; however, this may not be possible in these patients.

Figure 6

Figure 6

In the past, CT has been the most widely used imaging modality and was integral to several classifications.4,13,14,29 However, each classification used a different CT scanning method and a different treatment algorithm. Fielding and Hawkins4 used a single static CT scan with the head resting in the position of presentation. McGuire et al29 and Pang and Li13 used dynamic CT scanning and plotted the difference in the C1-2 angles to gauge the amount of stickiness, or resistance to normal motion. Ishii et al14 added three-dimensional reconstructions to judge both lateral mass subluxation and acquired deformity of the C2 lateral mass in chronic cases. In all of these studies, the more minor types of AARS were more likely to occur in patients with acute presentations.

The value of diagnostic CT scanning may be questioned in light of more recent evidence. Alanay et al34 noted both poor reliability and poor reproducibility of dynamic CT scanning and recommended against its routine use, especially in patients presenting acutely. Hicazi et al2 studied a group of patients who presented within 4 days of symptom onset. The authors found no significant motion abnormalities on dynamic CT scanning and called into question the utility of the study in the acute setting. Several authors have advocated using multiple follow-up CT scans to judge the effectiveness of their treatments,13,18 but because CT imparts some risk from ionizing radiation, clinicians should carefully consider the type and frequency of CT scans needed to guide diagnosis and treatment.35 When considering the use of CT for acute patients, keep in mind that patients with acute torticollis are more likely to have fewer observable abnormalities on CT scans and respond favorably to nonsurgical management; additionally, more aggressive treatments can be instituted if nonsurgical management fails. Thus, CT for acute patients may not outweigh the risks associated with radiation exposure.

MRI can be accomplished without any risk of ionizing radiation, and this modality can detect edema in the C1-2 complex. T2 fat-suppressed sequences and STIR sequences are able to detect edema in the joint capsules, posterior ligamentous system, alar ligaments, and the transverse ligament. MRI may also be useful if some other underlying condition, such as a tumor or infection, is suspected. MRI findings have been included in treatment algorithms for acute AARS, but it is unclear how MRI might contribute to more chronic or recalcitrant cases of AARS.5 From a practical standpoint, advanced imaging, including MRI, may not make a difference regarding a decision to initiate treatment of AARS.

Back to Top | Article Outline

Treatment

Some acute cases of AARS have resolved with medication alone.32 More commonly, some form of passive cervical manipulation is instituted. Options to realign the neck into a reduced position include cervical collars, halter or skeletal traction, halo immobilization, and surgery. The duration of symptoms before presentation is the most significant predictor of the type of treatment that will be required.3,13-15,20,23 In the absence of fracture or neurologic compromise, nonsurgical management is a viable option before resorting to surgery. No formal definition exists to determine when AARS is considered acute, but many series note differences in the type of treatment required when patients present within 1 month from the onset of symptoms.3,13,20,23

Been et al32 reported a series of four patients who presented to their institution within 3 weeks from the onset of torticollis. CT scans, with muscle spasm eliminated by anesthesia, were normal. The patients’ symptoms all resolved within 4 weeks with analgesics only. Pang and Li13 noted that patients in the diagnostic gray zone were an average of 13.5 days from the onset of their symptoms. Patients in the least severe category of AARS were an average of 46 days from the onset of symptoms, and patients in the most severe category were an average of 113 days from the onset of symptoms. Halter traction was routinely used for all of the patients diagnosed with AARS. No patients with symptoms for <1 month required halo traction or surgery. All patients in the diagnostic gray zone were treated with analgesics alone. Because most acute cases of AARS represent the least severe types, and are the most amenable to nonsurgical management, it seems acceptable to initiate such treatment without exposing patients to ionizing radiation from CT.35

The most consistently employed acute measures for AARS are analgesics and a cervical collar for comfort. If the clinical torticollis resolves quickly, no other treatment is required. However, the definition of “quickly” remains controversial. Various protocols recommend use of an orthosis until the clinical resolution of torticollis, or use for some predetermined time regardless of resolution. Recommended time in the orthosis ranges from 1 week to 6 months.2,5,13,20,22,33 An orthosis alone will not be successful for all patients.2,13,20,23 Landi et al5 recommended an MRI protocol to determine the length of time in the orthosis. Patients had resolution within 1 week, but the collars were kept intact until repeat MRI showed resolution of hyperintensity in the transverse ligament and alar ligaments. None of the patients had complete resolution of the hyperintensity at 1 month, 6 patients had resolution at 3 months, and 4 patients had resolution at 6 months. All patients were asymptomatic after 1 year.

Chronic cases of AARS and acute cases that fail to respond to nonsurgical management require more aggressive intervention. Patients with more severe AARS have an increased duration of symptoms.29 Patients presenting 1 month or more after the onset of symptoms have the most resistance to reduction; this resistance increases the further patients are from the onset of symptoms.13 All methods of treatment of chronic AARS have the goal of achieving a stable resolution of the torticollis.

Phillips and Hensinger3 documented successful reduction of AARS of more than 1 month’s duration in six of seven patients treated with halter traction. Recurrence occurred in four of the six patients, and two patients eventually required arthrodesis. Subach et al20 treated patients with AARS with halter traction, followed by immobilization in an orthosis; traction was repeated if recurrence was noted. Ten of 15 patients had a successful reduction with halter traction, but five patients eventually required arthrodesis. Beier et al23 used halter traction on eight patients; definitive treatment was successful in four patients. Park et al19 reported on the use of halter traction to treat a patient who presented >3 months after the onset of symptoms. The patient required halter traction for 6 weeks, and then bracing and a collar for a total of 6 months.

Skeletal traction, applied either with a halo or tongs, has been advocated by some clinicians as a primary treatment to achieve reduction of AARS. Protocols for cranial traction vary by author, but in general, 3 to 5 pounds of weight are applied; adjunctive analgesics and muscle relaxants can be used to encourage the reduction. The amount of weight may be increased by 2 to 3 pounds each day, up to 33% to 55% of body weight, keeping in mind that routine neurologic evaluations, including checking for cranial nerve palsies, are critical. Similar to traction for scoliosis, wheelchair and walker devices can be fabricated to allow more freedom of movement while in traction (Figure 7). No formal end point has been established to decide when to discontinue the traction if it is not successful, but most authors recommend proceeding with other methods after 2 weeks. If adequate reduction can be established with cranial traction, the reduction should then be held by using immobilization for several months, with most authors recommending a halo vest. Ishii et al14 had no recurrences after treating seven patients with AARS with skeletal traction followed by an orthosis. However, in patients with symptoms of >8 weeks' duration, patients either did not achieve reduction with skeletal traction or had a recurrence following reduction. All patients eventually required arthrodesis. Pang and Li13 used skull traction for severe chronic cases or for recurrences. The authors recommended halo immobilization following successful reductions; arthrodesis was recommended if reduction could not be achieved after 2.5 weeks.

Figure 7

Figure 7

Halo vest devices have been used to reduce AARS and to maintain reduction after other treatment methods3,11,13 (Figure 8, A). Ishii et al18,36 successfully managed patients with chronic AARS using closed manipulation and halo vest devices. The authors determined the time course for treatment based on remodeling of the C2 facet on three-dimensional CT scans. They demonstrated that dysplasia was reversed after halo immobilization for an average of 2.8 months (Figure 8, B through E).

Figure 8

Figure 8

C1-2 arthrodesis is usually required for cases recalcitrant to nonsurgical management. Rates for eventual fusion for chronic cases vary from 30% to 100%.3,13,14,20,23 Several different methods have been described to perform C1-2 arthrodesis. Posterior wiring methods have been described by Gallie37 and McGraw and Rusch.38 Brooks and Jenkins39 described wiring with an autogenous iliac crest graft between C1 and C2. Lateral mass screw fixation of C1 and pedicle screw fixation of C2 was described by Harms and Melcher40 (Figure 9). Tauchi et al15 used the Harms and Melcher technique to perform arthrodesis for chronic AARS in a series of six consecutive patients, with good results. Successful fusion was documented in all six patients using radiographs and CT, although one patient achieved fusion with the head in a slightly tilted position.

Figure 9

Figure 9

Back to Top | Article Outline

Authors’ Preferred Algorithm

At initial presentation, the physician should perform a history and physical examination and obtain radiographs of the cervical spine. The presence of AARS should be based mainly on the history and physical examination finding of fixed torticollis.

We recommend against the use of CT or MRI unless there is a strong clinical suspicion for infection, fracture, or congenital abnormality based on the history, physical examination, or initial radiographs.

Once the diagnosis of AARS is established, the orthopaedist should proceed according to the algorithm seen in Figure 10.

Figure 10

Figure 10

Back to Top | Article Outline

Summary

AARS is a rare condition in which patients present with the new onset of torticollis. Initial evaluation should be focused on eliminating inflammatory conditions or trauma as underlying causes. For acute cases, nonsurgical management can be employed to treat the torticollis, beginning with the use of a cervical collar and analgesics (Figure 10). For chronic cases, or acute cases that fail cervical collar treatment, halter traction, analgesics, and muscle relaxers are appropriate. If halter traction is successful, maintenance of the reduction should be accomplished by using a cervical collar. If halter traction is not successful, skeletal traction should be considered. If reduction with tongs is successful, the reduction should be maintained with a halo device. If reduction is unsuccessful, or if the torticollis recurs, reduction and posterior fusion of C1-C2 is appropriate. Whereas various imaging modalities, including radiographs, dynamic CT scanning, three-dimensional CT reconstructions, and MRI, have been used to aid in making a diagnosis and to assist with determining the length of treatment required, there is no universally agreed-upon treatment algorithm, and the amount of ionizing radiation delivered to the head and neck region should be minimized, especially in children.

Back to Top | Article Outline

References

Evidence-based Medicine: Levels of evidence are described in the table of contents. In this article, reference 13 is a level II study. References 34 and 35 are level III studies. References 2-6, 11, 12, 14-24, 26-30, 32, 33, 36, and 38-40 are level IV studies. Reference 37 is a level V expert opinion.

References printed in bold type are those published within the past 5 years.

1. , : Atlantoaxial rotatory fixation: Part 1—Biomechanics of normal rotation at the atlantoaxial joint in children. Neurosurgery 2004;55(3):614–625
2. , , , , : Atlantoaxial rotatory fixation-subluxation revisited: A computed tomographic analysis of acute torticollis in pediatric patients. Spine (Phila Pa 1976) 2002;27(24):2771–2775
3. , : The management of rotatory atlanto-axial subluxation in children. J Bone Joint Surg Am 1989;71(5):664–668
4. , : Atlanto-axial rotatory fixation. (Fixed rotatory subluxation of the atlanto-axial joint). J Bone Joint Surg Am 1977;59(1):37–44
5. , , , , : Atlantoaxial rotatory dislocation (AARD) in pediatric age: MRI study on conservative treatment with Philadelphia collar—experience of nine consecutive cases. Eur Spine J 2012;21(suppl 1):S94–S99
6. : Enucléation de l’atlas et torticollis naso-pharyngien. Presse Med 1930;38:50–54.
7. , , , : Developmental morphology and ossification patterns of the C1 vertebra. J Bone Joint Surg Am 2013;95(17):e1241–e1247
8. , : Developmental anatomy of the atlas and axis in childhood by computed tomography. J Neurosurg Pediatr 2011;8(3):235–243
9. , , , : Ligaments of the craniocervical junction. J Neurosurg Spine 2011;14(6):697–709
10. , , , , : Adiponectin enhances IL-6 production in human synovial fibroblast via an AdipoR1 receptor, AMPK, p38, and NF-kappa B pathway. J Immunol 2007;179(8):5483–5492
11. , , , , : Successful conservative treatment for neglected rotatory atlantoaxial dislocation. J Pediatr Orthop 2013;33(4):389–392
12. , , , , : Report of an unusual upper cervical spine injury: Traumatic atlantoaxial rotatory subluxation with vertical odontoid fracture in a child. Neurol India 2013;61(2):197–199
13. , : Atlantoaxial rotatory fixation: Part 2—New diagnostic paradigm and a new classification based on motion analysis using computed tomographic imaging. Neurosurgery 2005;57(5):941–953
14. , , , , , : Pathognomonic radiological signs for predicting prognosis in patients with chronic atlantoaxial rotatory fixation. J Neurosurg Spine 2006;5(5):385–391
15. , , , : Surgical treatment for chronic atlantoaxial rotatory fixation in children. J Pediatr Orthop B 2013;22(5):404–408
16. , , : Atlantoaxial rotatory subluxation in patients with Marfan syndrome: A report of three cases. Spine (Phila Pa 1976) 2000;25(4):524–526
17. , : Inflammatory torticollis in children. Arch Otolaryngol Head Neck Surg 1990;116(3):310–313
18. , , , : Remodeling of C2 facet deformity prevents recurrent subluxation in patients with chronic atlantoaxial rotatory fixation: A novel strategy for treatment of chronic atlantoaxial rotatory fixation. Spine (Phila Pa 1976) 2011;36(4):E256–E262
19. , , , : Successful reduction for a pediatric chronic atlantoaxial rotatory fixation (Grisel syndrome) with long-term halter traction: Case report. Spine (Phila Pa 1976) 2005;30(15):E444–E449
20. , , , : Current management of pediatric atlantoaxial rotatory subluxation. Spine (Phila Pa 1976) 1998;23(20):2174–2179
21. , , , : Grisel’s syndrome: A case report and review of the literature. Int J Pediatr Otorhinolaryngol 2005;69(12):1689–1692
22. , , , , : Medically treated paravertebral Brucella abscess presenting with acute torticollis: Case report. Surg Neurol 2007;67(2):207–210
    23. , , , , , : Rotatory subluxation: Experience from the Hospital for Sick Children. J Neurosurg Pediatr 2012;9(2):144–148
    24. , : Isolated atlantoaxial rotatory fixation in a child with seronegative spondyloarthropathy presenting with torticollis. J Rheumatol 1998;25(1):169–172
    25. , , : The pharyngovertebral veins: An anatomical rationale for Grisel’s syndrome. J Bone Joint Surg Am 1984;66(4):568–574
    26. : Spontaneous hyperaemic dislocation of the atlas. Proc R Soc Med 1932;25:586–590.
    27. , , : Atlantoaxial rotatory instability secondary to odontoid hypoplasia as a cause of acute torticollis in children: Report of one case. Zhonghua Min Guo Xiao Er Ke Yi Xue Hui Za Zhi 1995;36(6):448–451
    28. , , , : Atlantoaxial rotatory fixation in the setting of associated congenital malformations: A modified classification system. Spine (Phila Pa 1976) 2010;35(4):E119–E127
    29. , , , , : Torticollis in children: Can dynamic computed tomography help determine severity and treatment. J Pediatr Orthop 2002;22(6):766–770
    30. : Atlantoaxial rotatory fixation. Neurosurgery 2010;63(suppl 3):161–183
    31. : Rheumatologic conditions in children who may present to the orthopaedic surgeon. J Am Acad Orthop Surg 2011;19(3):163–169
    32. , , : Suspected atlantoaxial rotatory fixation-subluxation: The value of multidetector computed tomography scanning under general anesthesia. Spine (Phila Pa 1976) 2007;32(5):E163–E167
    33. , , : Atlantoaxial rotary subluxation after minor trauma. Pediatr Emerg Care 2008;24(12):852–856
    34. , , , : Reliability and necessity of dynamic computerized tomography in diagnosis of atlantoaxial rotatory subluxation. J Pediatr Orthop 2002;22(6):763–765
    35. , , , : The use of computed tomography in pediatrics and the associated radiation exposure and estimated cancer risk. JAMA Pediatr 2013;167(8):700–707
    36. , , , , : Management of chronic atlantoaxial rotatory fixation. Spine (Phila Pa 1976) 2012;37(5):E278–E285
      37. : Fractures and dislocations of the cervical spine. Am J Surg 1939;46:495–499.
      38. , : Atlanto-axial arthrodesis. J Bone Joint Surg Br 1973;55(3):482–489
      39. , : Atlanto-axial arthrodesis by the wedge compression method. J Bone Joint Surg Am 1978;60(3):279–284
      40. , : Posterior C1-C2 fusion with polyaxial screw and rod fixation. Spine (Phila Pa 1976) 2001;26(22):2467–2471
      © 2015 by Lippincott Williams & Wilkins, Inc.