Intended for healthcare professionals


Discontinuation of cervical spine immobilisation in unconscious patients with trauma in intensive care units–telephone survey of practice in South and West region

BMJ 1997; 314 doi: (Published 07 June 1997) Cite this as: BMJ 1997;314:1652
  1. K J Gupta, specialist registrara,
  2. M Clancy, consultantb
  1. a Department of Anaesthesia, Frenchay Hospital, Bristol BS16 1LE
  2. b Department of Accident and Emergency, Bristol Royal Infirmary, Bristol BS2 8HW
  1. Correspondence to: Dr K J Gupta Department of Anaesthesia, Royal United Hospital, Combe Park, Bath BA1 3NG
  • Accepted 14 March 1997


Objective: To study how the cervical spine is assessed before discontinuation of cervical spine immobilisation in unconscious trauma patients in intensive care units.

Design: Telephone interview of consultants responsible for adult intensive care units.

Setting: All 25 intensive care units in the South and West region that admit victims of major trauma.

Main outcome measures: The clinical and radiological basis on which the decision is made to stop cervical spine immobilisation in unconscious patients with trauma.

Results: In 19 units cervical spine immobilisation was stopped in unconscious patients on the basis of radiology alone, and six units combined radiology with clinical examination after the patient had regained consciousness. Sixteen units relied on a normal lateral radiological view of the cervical spine alone, five required a normal lateral and anteroposterior view, and four required a normal lateral, anteroposterior, and open mouth peg view.

Conclusions: There are inconsistencies in the clinical and radiological approach to assessing the cervical spine in unconscious patients with trauma before the removal of immobilisation precautions. There is an overreliance on the lateral cervical spine view alone, which has been shown to be insensitive in this setting.

Key messages

  • Immobilisation of the cervical spine is essential in victims of major trauma until the cervical spine is fully assessed for injury

  • In the NHS region surveyed intensive care units show little consistency in how the cervical spine is assessed before the removal of immobilisation precautions

  • Most practitioners would be prepared to accept radiography of the lateral cervical spine as the sole radiological assessment, although it has been shown to be an insensitive screening tool

  • Computed tomography targeted at poorly visualised or suspicious areas in three plain radiographs (lateral, anteroposterior, and open mouth peg views) is a sensitive technique for the radiological detection of injury to the bony cervical spine


The reported incidence of cervical spine injuries in victims of major trauma varies from 2.3% to 12%.1 2 3 Victims of trauma therefore have their cervical spine immobilised at the earliest possible opportunity. Those patients with a severe head injury (Glasgow coma scale <8) are usually rapidly anaesthetised and intubated as part of their emergency care and are subsequently admitted to an intensive care unit. The cervical spine will normally remain immobilised during these procedures until injury to the neck can be ruled out.

In the advanced trauma life support course for physicians the American College of Surgeons' committee on trauma states that “Patients…should be presumed to have an unstable cervical spine injury and the neck should be immobilized until all aspects of the cervical spine have been adequately studied and an injury excluded.” It also states that discontinuation of spinal immobilisation “usually occurs when no roentgenographic abnormality has been documented, and no symptoms or signs relating to the spine or cord exist.”4 While this has become the accepted standard in conscious patients it is less easily applied if the patient is unconscious. The assessment of the cervical spine before removal of immobilisation precautions in unconscious trauma patients therefore remains controversial.


KJG conducted a telephone survey of all of the general intensive care units dealing with major adult trauma in a single NHS region (South and West) over a period of two weeks. These units were identified from The Directory of Emergency and Special Care Units.5 Units were excluded only if they were in hospitals without an accident and emergency department from which to admit victims of trauma. The consultant in charge of the intensive care unit on the day of the call was interviewed, or the clinical director of the unit if the consultant was not available. If neither were available then the call was made on another day when another consultant was covering the unit.

The participating consultants were given the same case scenario of an unconscious and intubated trauma victim with a head injury being admitted to their intensive care unit. They were told that the patient had no specific clinical evidence of a cervical spine injury but that the patient's cervical spine was immobilised with a semirigid cervical collar.

They were asked what clinical and radiological investigations were normally required on their unit before such immobilisation was discontinued. Each consultant was then allowed to answer without further prompting or interruption. If specific information required for the survey was not obtained in the first answer a prepared structured question pertaining to that specific aspect was used.

The specific information sought during the interview was, firstly, whether the cervical spine immobilisation precautions were normally discontinued before or after the patient had regained consciousness; secondly, which radiological views were required as a minimum, assuming each view was anatomically complete and of good quality; thirdly, the specialty and grade of those normally interpreting the film; fourthly, whether the hospital or unit had an established policy or guidelines for this problem and, if so, whether the responses were based on this; fifthly, if the respondents had in their practice encountered morbidity from the prolonged use of a semirigid cervical collar; and, finally, if the respondents had encountered any patients with a missed cervical spine injury detected only after removal of immobilisation precautions.


All 25 suitable intensive care units responded to the telephone survey. In each unit the information was provided by a consultant with a regular commitment to intensive care. Sixteen of the units were in hospitals whose accident and emergency departments saw over 30 000 patients a year, seven in hospitals seeing over 50 000 patients a year, and two in hospitals seeing more than 70 000 accident and emergency patients a year. Three of the hospitals contained neurosurgical centres, and one was a centre for spinal injuries.

Ten consultants responded according to guidelines relating to this problem already established in their hospital. These guidelines had been set in three cases by intensivists, three by neurosurgeons or orthopaedic surgeons, one by radiologists, and three by a multidisciplinary team. Fifteen units had no established guidelines or protocol in place.

The practice on 19 units was to remove all cervical spine precautions while the patient was still unconscious. Six would maintain cervical spine immobilisation until patients had regained consciousness. Table 1) shows the minimum radiological requirements accepted by units with these different criteria.

Table 1

Minimum radiological requirements before removal of cervical spine immobilisation precautions according to consciousness of patients. Values are numbers of hospital units

View this table:

There was a large overlap in the specialty of the doctor commonly asked to interpret the radiographs. Five units regularly used a radiologist, and seven units regularly used either an orthopaedic surgeon or neurosurgeon. Ten units would use either specialist depending on which was available. On three of the units the intensive care consultants would regularly interpret the films. All the units used grades of senior registrar or above in the relevant specialty.

Eighteen of the consultants had, in their practice, encountered morbidity associated with the prolonged use of semirigid cervical collars on intensive care patients: 10 related to the occurrence of occipital or mental pressure sores, or both; four had experienced considerable difficulties during extubation or reintubation of patients in collars; and two volunteered that nursing staff had complained of general difficulties, including awkward mouth care and increased nursing requirements during “log roll” turns of patients.

Four of the consultants were aware of cases of missed injury of the cervical spine, discovered after removal of immobilisation precautions. Information on the outcome in these patients was not available.



Telephone surveys are more effective than postal surveys in achieving complete participation as posted forms may not be received or may be mislaid or ignored. We were thus able to survey practice in all the hospitals in our region that may receive victims of major trauma. A criticism of telephone interviews is that the interaction they afford allows the interviewer inadvertently to bias the respondents. We minimised this by allowing each consultant to give a full and uninterrupted answer after an initial open ended question relating to a case scenario. Only if specific information required for the survey was not obtained from the first answer did interaction take place. This was in the form of specific structured questions.

In any telephone survey, assessment of the reliability of responses and their applicability to actual practice may be difficult. It is important to interview the person responsible for making the relevant decision.6 We therefore interviewed the consultants responsible for the unit on the day of the interview. Respondents may also be influenced by personal opinion based on a perceived “ideal” practice. This is difficult to control, but the consultants were asked to describe the normal practice in their unit in an attempt to minimise the influence of personal bias, particularly in those units without established guidelines. It is also possible that consultants from different hospitals could influence each other's answers through discussion of this problem at local or regional meetings. We therefore carried out the survey over a brief period of time, hoping to achieve a “snap shot” of practice across the region.

Findings in the context of current knowledge

Sixteen of the 25 (64%) units were prepared to rely on a lateral plain radiograph as their sole cervical spine radiology. Twelve of these units would be prepared to remove cervical spine immobilisation on the basis of this view while the patient was still unconscious. Several studies have clearly shown that the cross table lateral view in isolation will miss about 15% of patients with a cervical spine fracture or dislocation, even if the films are anatomically complete, of good quality, and are read by an expert.2 7 8 Between 37% and 63% of these missed injuries will be potentially unstable.7 8 The problem of lateral views is further compounded when the injuries are missed because of inexpert interpretation or incomplete films, or both.1 2 3 8 The commonest areas of missed injuries on the lateral view are the C1/2 and C7/T1 regions.8 9 Between 7% and 25% of lateral views will not show the top of T1, despite the use of arm traction.3 10 11

Five of the 25 centres reported using three plain views to visualise the cervical spine: the lateral, anteroposterior, and open mouth odontoid peg view. Several studies have shown the combination of these views to have a sensitivity of 90-99% for the detection of cervical spine injuries.2 7 8 9 11 12 Again, expert interpretation of the films is necessary to achieve these rates. In our survey 22 units were specifically asking a radiologist, a senior orthopaedic surgeon, or a neurosurgeon to interpret the radiographs.

None of the consultants said that their units were using computed tomography as a first line investigation, although several were using it as a second line investigation if plain radiography was inadequate. Computed tomography can miss horizontally orientated fractures,12 13 14 although this risk can be reduced by the use of reconstructed sagittal images. The sensitivity of computed tomography when used alone varies from 80% to 92%.11 13 14 If it is used in combination with three plain view films by being targeted to poorly visualised or suspicious areas on these films, however, the sensitivity for the detection of cervical spine injury approaches 100%.9 11 14

Six units would wait until the patients had regained consciousness before removing immobilisation precautions. We were unable to find any published evidence in unconscious patients to support or refute this practice. It seems sensible to wait for the added reassurance of normal results from a clinical examination in these patients. Several papers, however, have highlighted the poor reliability of such normal results in patients with spine injury who have an altered consciousness because of head injury, sedative drugs, or alcohol, or who have other distracting injuries.8 15 16 Also the effectiveness of semirigid cervical collars in providing immobilisation when used alone is questionable19 20 and, as implied by the anecdotal reports in this survey, longer term use of semirigid cervical collars can be associated with considerable morbidity.17 18

What is the incidence of the problem?

Four consultants knew of a patient in whom a cervical spine injury was diagnosed only after removal of immobilisation precautions. These reports are anecdotal, but we would not expect an intensive care unit to reproduce figures on a problem it would encounter comparatively infrequently. Anecdotes are often the only source of information on uncommon events and therefore these results are still relevant. Some reports claim that injuries of the cervical spine are missed or their diagnosis is delayed in 4.6% to 17% of victims of trauma with spine injury.1 2 21 Delay in diagnosis of cervical spine injuries has been shown to effect mortality and morbidity, with up to 29% of patients suffering neurological deterioration as a result.1 22 While some missed injuries may be the result of isolated ligamentous disruption, which may not be visible on plain radiography,23 the incidence of this type of injury in this population is probably less than 1%.17


The approach to the common problem of deciding when to discontinue cervical spine immobilisation in unconscious patients with trauma in an intensive care unit varies widely in the South and West region. There is no consistency in the minimum radiological investigations required or whether the patient's consciousness should effect the decision. There is an overdependence on the isolated use of lateral plain radiographs, which have been shown to be an insensitive screening tool. The detection rate of injury can be considerably improved by the use of three plain views (a lateral, anteroposterior, and open mouth peg view), particularly if poorly visualised or suspicious areas seen on these views are targeted with computed tomography.


Funding: No external funding.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
View Abstract