Education And Debate

Lesson of the Week: Don't twist my child's head off: iatrogenic cervical dislocation

BMJ 1995; 311 doi: (Published 04 November 1995) Cite this as: BMJ 1995;311:1212
  1. A T H Casey, spinal research fellowa,
  2. M O'Brien, international spinal research fellowa,
  3. V Kumar, neurosurgical registrarb,
  4. R D Hayward, consultant neurosurgeonb,
  5. H A Crockard, consultant neurosurgeona
  1. aThe National Hospital for Neurology and Neurosurgery, London WC1N 3BG
  2. bThe Hospital for Sick Children, London WC1N 3JH
  1. Correspondence to: Mr Crockard.
  • Accepted 24 March 1995

Excessive rotation of two children's necks while they were undergoing cervical lymph node biopsy probably resulted in their dislocation; delayed diagnosis necessitated complex spinal surgery to correct the deformity. “Wry” neck, a synonym for atlantoaxial rotary subluxation, is a relatively common condition in young children often associated with tonsillitis. It usually resolves spontaneously,1 because the joint surfaces remain in contact. A rotation injury is another possible cause, often following an apparently minor trauma such as heading a football. Iatrogenic torticollis is rarely reported. A complete dislocation of the atlantoaxial joint will not resolve spontaneously and may have important long term effects on the child's facial development. We have treated two otherwise normal children whose necks were dislocated under general anaesthesia while they were undergoing cervical lymph node biopsy. The injury was not appreciated at once and definitive treatment was inevitably delayed.

  • “Wry neck” in children may be due to atlantoaxial rotatory subluxation and should be promptly evaluated by detailed cervical spine radiography

Case reports CASE 1

A 10 year old girl underwent lymph node biopsy from the posterior triangle of her neck. After surgery she awoke with a torticollis, a painful neck, and occipital dysaesthesiae. These symptoms ultimately subsided on conservative treatment but the torticollis persisted. Radiography at that stage showed atlantoaxial rotary dislocation. This failed to reduce following manipulation under anaesthesia and a period of traction performed three months later. At that stage, an upper cervical laminectomy was inappropriately performed in the mistaken belief that spinal cord compression played an important part in the pathological cause. There was no change in the deformity or neurological status. She was subsequently referred to this neurosurgical unit 15 months after the original injury, at which stage minor facial asymmetry was apparent. The atlantoaxial disclocation was explored and reduced successfully via an open operative (the “extreme lateral”) approach.2


A 6 year old girl underwent lymph node excision biopsy from the right side of her neck for persistent lymphadenopathy of unknown cause. On return from theatre she was reported as having a torticollis (fig 1) and neck pain, which was attributed initially to focal pain at the site of the incision. When this did not improve over the next few months further investigations, including electromyography (which showed that this was not a spasmodic torticollis), and, later, computed tomography with 2D and 3D reconstructions, were performed. These showed an atlantoaxial rotary dislocation which, not surprisingly, failed to reduce by closed methods eight months after the lymph node excisional biopsy. One year after the original injury she was referred to our neurosurgical unit for further evaluation and treatment. The atlantoaxial complex was again exposed by the “extreme lateral” approach and reduction was achieved, with postoperative immobilisation achieved by the use of a halo jacket for 10 weeks.


Case 2. Preoperative picture showing typical cock-robin position. Photograph reproduced by permission of the patient's parent


Around 70% of all paediatric spinal injuries affect the upper cervical spine.3 This is due to a combination of factors including ligamentous laxity, poorly developed cervical musculature in association with a large head, horizontally oriented facet joints, and wedge shaped cervical vertebrae. If excessive axial torque is applied to this region and the resultant angular rotation exceeds 65°, then complete bilateral C1-C2 facet dislocation with ligamentous rupture may occur,4 5 as experienced by our patients during positioning of the head under anaesthesia. The most likely mechanism is that the child was supine on the operating table and, to gain access, the surgical and anaesthetic team rotated and held the head to one side to expose the posterior triangle of the neck. Atlantoaxial dislocation resulted in a torticollis in these two cases, but potential compromise of the vertebral artery or the spinal cord can occur in this type of injury.

The torticollis makes interpreting the lateral cervical spine radiographs difficult, with the mandible often obscuring the atlas and axis. Computed tomography with multiplanar reconstructions are therefore invaluable aids to the diagnosis and surgical planning (fig 2).


Axial computed tomogram showing rotary atlantoaxial dislocation. C1 and C2 indicated by arrows

If dealt with at an early stage this injury can usually be managed by conservative means consisting of traction, manipulation, and subsequent immobilisation by a collar. When treatment is delayed, however, reduction is unlikely to be successful, and complex spinal surgery, with the attendant risks of spinal cord or vertebral artery injury, will be required.

What are the lessons to be learnt from these two cases? Firstly, extreme care must be taken in the positioning of the anaesthetised and paralysed child where the normal protection from cervical musculature is lost: extremes of neck rotation in children are dangerous. Furthermore, a wry neck following surgery should be promptly evaluated by cervical spine radiographs including an open mouth view supplemented by computed tomography through the relevant area.

Secondly, as in all dislocations, successful reduction is a time related phenomenon, and prompt treatment is vital. This point is well accepted in orthopaedic circles for dislocations of the elbow and injuries to the wrist and hand. However, the hard lessons learnt there do not appear to have carried themselves across to the atlantoaxial joint. Moreover, if reduction fails one should consider the possibility that there may be soft tissue interposition impeding successful reduction, or cross union may have occurred. This was the case in both our patients, where at operation, owing to the delays in treatment, osseous union had occurred between the odontoid process, anterior arch of the atlas, and the medial aspect of the lateral mass of the axis (case 1) and, in case 2, where there was ligamentous tissue preventing reduction.

Finally, long delays in treatment may result in asymmetry of facial growth, as was seen in our first case,1 5 and long term concerns must exist about the late development of spinal cord damage from an inadequate spinal canal at the level of injury.


The surgeon and anaesthetist must ensure that overvigorous positioning of the child's head and neck does not occur or the atlantoaxial joint may dislocate. When it does, prompt treatment is mandatory as delays in management will compromise the chances of successful reduction to correct the deformity by closed methods and increase the necessity of complex spinal surgery with all its attendant risks.


View Abstract