Intended for healthcare professionals

Clinical Review Lesson of the week

Immobilisation of the cervical spine in children

BMJ 2002; 324 doi: (Published 09 March 2002) Cite this as: BMJ 2002;324:591
  1. Sophie Skellett (sophie7{at}, fellow,
  2. Shane M Tibby, consultant,
  3. Andrew Durward, consultant,
  4. Ian A Murdoch, director
  1. Department of Paediatric Intensive Care, Guy's Hospital, London SE1 9RT
  1. Correspondence to: S Skellett
  • Accepted 29 August 2001

Recognition and early management of children at risk of injury to the cervical spinal cord is lacking

Trauma to the cervical spine accounts for 1.5% of admissions in children with trauma in the United States, 35% of which have an associated injury of the cervical spinal cord.1 Such an injury may occur along with bony or ligamentous damage. Ligamentous damage encompasses injury to the spinal cord without radiological evidence of abnormalities.2 Immobilisation of the cervical spine is mandatory in patients at risk to prevent the development of injury to the cervical spinal cord and its progression.24

Most children requiring hospital admission for trauma are considered at risk of injury of the cervical spinal cord. The commonest causes are road crashes, falls from greater than 4.6 metres, head injuries, and injuries elsewhere to the spine. 5 6 These patients require immobilisation of the spine, commonly achieved with a rigid collar and a supplemental device. Immobilisation should be discontinued only after exclusion of bony and ligamentous damage, which needs an adequate clinical examination.7 Important features of the examination are neck pain, midline tenderness on palpation of the neck, and any signs or symptoms of neurological injury.8 This is possible only in a conscious patient (Glasgow coma score >13) who does not have painful distracting injuries. The negative predictive value of clinical examination is almost 100%.8 Radiological evaluation plays a part, initially in identifying unstable fractures requiring urgent surgery.

As the lead centre for paediatric intensive care in the south east of England, Guy's Hospital provides a retrieval service for three paediatric intensive care units, collecting patients from over 30 district general hospitals. We were thus able to audit immobilisation of the cervical spine in children who were at risk of injury to the cervical spinal cord.


Of 870 patients identified from our database over an 18 month period (January 1999 to June 2000) we found 60 at risk of injury to the cervical spinal cord. We collected data from notes made during retrieval, notes from the tertiary centre, photocopies of paramedic or ambulance sheets, and notes from local referring hospitals. Forms were available for all patients. We used the Mann-Whitney test to compare continuous data and χ2 analysis to compare categorical data.

The median age was 7 years (interquartile range 3 to 12), with boys predominating (70%). Injuries concerned pedestrians in road crashes (26 patients), falls from greater than 4.6 metres (14), passengers in road crashes (8), blows to the head (6), bicyclists in road crashes (4), and hanging (2). Most patients were unconscious on presentation to the local hospital, precluding clinical examination. Overall, 65% (39 patients) had a Glasgow coma score of 8 or less, 15% (9) a score of 9 to 12, and 20% (12) a score of 13 or more. Of the 12 patients scoring 13 or more, nine deteriorated soon after arrival.

Immobilisation before arrival at hospital

Figure 1 shows the modes of transport and cervical spine protection from the site of injury to the local hospital. Overall, 23% of the children (14 of 60) had not been immobilised before arrival at hospital; six of these had been transported by either their parents or their general practitioner. Of the 54 children taken by ambulance, eight (15%) had not been immobilised.

Fig 1.
Fig 1.

Immobilisation of cervical spine before arrival at hospital

Immobilisation at local hospital

The local hospital did not initiate immobilisation in any patient. The 14 patients who were not immobilised, including those transported by parents or general practitioners, were younger than those who were immobilised (median age 1.5 years (0.6-3.0) versus 9 years (5.1-12.4; P<0.001)). The non-immobilised group differed according to the type of injury; a higher proportion of injuries were due to isolated blows to the head (28% v 9%) and falls (36% v 22%) and fewer to road crashes (36% v 69%) (P<0.05).

In 2 of the 60 patients no cervical spine injuries were found after clinical examination. Both were alert and oriented and had no distracting injuries. Of the remaining 58 patients, 56 required intubation and ventilation at the local hospital. Most were sedated, often also receiving neuromuscular blockade precluding clinical clearance regardless of the Glasgow coma score. The remaining two patients had distracting injuries.

When the retrieval team arrived, 30 of the 58 patients had been immobilised (fig 2). Of the 28 who had not been immobilised, 14 had arrived at the local hospital that way and the remainder had immobilisation discontinued solely on the basis of a lateral radiograph. Imaging of the patients at risk included several views of the cervical spine (lateral, oblique, and anteroposterior; 8 patients), computed tomography of the neck (3), and radiography of the lateral cervical spine only (38). Eleven patients had no imaging. Forty nine patients underwent computed tomography of the head, chest, or abdomen, but only three scans included the neck.

Fig 2.
Fig 2.

Immobilisation of cervical spine at local hospital

Immobilisation by retrieval team

The retrieval team used spinal boards with either rigid collars or blocks, sandbags, and tape to immobilise 52 of the 58 patients before transportation. Six patients were transported with no immobilisation. These patients were younger than the immobilised patients, with a median age of 1.5 years (0.7-1.9) versus 7.5 years (4.3-12.0; (P<0.001)), and all had had either a fall or a blow to the head.


Five patients (8%) died from severe head injuries within 12 hours of the injuries. Four of these had had only radiography of the lateral cervical spine at the local hospital. The remaining patient had had no imaging. Three patients (5%) had fractures of the cervical spine, one requiring urgent surgery. All had been in a road crash. Two had a poor neurological outcome from severe head injuries. All were retrieved with immobilisation. Two, however, had not been immobilised by the time we arrived; one had never been immobilised, and the other had had the immobilisation removed after inadequate radiography of the lateral cervical spine. All patients identified with injuries of the spinal cord showed abnormalities on radiographs.


Failure to immobilise patients at risk of injury to the cervical spinal cord seems commonplace. This may have devastating consequences as 75% of injuries to the cervical spinal cord are incomplete at presentation.1 The problem concerns both a deficiency in initiating immobilisation and giving the all clear after inadequate evaluation.

Healthcare workers from a variety of disciplines did not initiate immobilisation of the cervical spine in 8 of 54 (15%) patients transported to hospital by ambulance, 24% of patients (14 of 58) in the emergency departments, and 10% (6 of 58) of those transferred to a tertiary care centre. The reasons for this are unclear but may include lack of awareness of healthcare workers of those at risk, paucity of suitable immobilising devices for small children, and an inability to maintain immobilisation in young, uncooperative patients.

Children with isolated head injuries or who had had falls were more likely to be overlooked for immobilisation than those in road crashes. All of these groups are defined as at risk in paramedic training and paediatric life support courses. 3 4 9 The proportion of frontline staff who had received such training is, however, unknown.

Patients who were not immobilised were younger than those immobilised. This has important clinical implications in that the proportion of injuries to the upper cervical spine (vertebrae C1-C4) is higher in patients less than 12 years of age, and this carries a higher mortality than injuries to the lower cervical spine (23% v 4%).1 This may partially reflect a perceived lack of immobilising options, as rigid collars are not available for infants less than 6 months of age. Spinal boards, sandbags, and tapes are a recommended alternative by paramedic training and paediatric life support courses. 3 4 A combination of two or more devices is recommended in unconscious patients; however, only hard collars are recommended for agitated patients. 3 4 10 These recommendations potentially leave infants less than 6 months unprotected but is unlikely to be a major factor in our non-immobilised patients, given their median Glasgow coma score of 7 on presentation.

Radiological clearance

A high proportion of patients were cleared of spinal injury by radiological evaluation and had the immobilising device removed at the local hospital (16 of 46, 35%). A reduced level of consciousness precluded adequate clinical examination in 14 of these patients, yet immobilisation was removed on the basis of a lateral radiograph of the cervical spine only. This view misses 26% of bony cervical spine injuries. 11 12 Both multiviews and spiral computed tomography, although acceptable for ruling out bony injury, will not show spinal cord injury without evidence of abnormality on radiographs. A normal clinical examination excludes most injuries to the cervical spine.8 If this cannot be performed immobilisation should be maintained even though radiological findings are normal.7

Study limitations

The limitations of a retrospective study must be acknowledged. Data concerning immobilisation of the cervical spine of patients before arrival at hospital were collected from paramedic or ambulance sheets, which may have been incomplete, resulting in underreporting. However, all cases were cross referenced with notes from the local hospital.


A deficit in the recognition and early management of children at risk of injury to the cervical spine occurs across disciplines. This may be due to lack of appreciation of mechanisms of injury producing risk, confusion regarding the immobilisation of the cervical spine of infants, and failure to appreciate that clearance must be given only after an appropriate clinical examination regardless of any radiological investigation. Efforts to rectify this problem must include education of healthcare professionals and ambulance and emergency staff having ready access to immobilising devices that are appropriate for different ages.


Contributors: SS conceived the idea, was responsible for data collection, and wrote the first draft of the manuscript; she will act as guarantor for the paper. ST undertook the statistical analysis and edited the first and subsequent drafts. AD assisted in data collection and processing and edited the first draft. IM advised in project design and edited the first and subsequent drafts.


  • Funding None.

  • Competing interests None declared.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
View Abstract