No place for the rigid cervical collar in pre-hospital care

The cervical collar was introduced to pre-hospital practice in 1967. It has been widely acknowledged that evidence supporting its use has been lacking, and that its use was therefore based upon hypothetical benefit.

However, modern evidence challenges the very principle of immobilisations on which the collar's use was justified, and argues that the cervical collar is an ineffective immobilisation tool regardless. Cervical collars have always been thought to be a relatively harmless measure, and their precautionary use therefore encouraged in order to protect those few patients who have sustained an injury. Yet, there is an increasing volume of evidence to say that the application of cervical collars can be harmful to patients. An assessment of today's evidence justifies an immediate change of practice, and EMS providers across the world are adapting accordingly. The UK remains entrenched in rigid trauma protocol, which is not consistent with the principles of patient safety, evidence-based practice or the HCPC's Standards of Proficiency for Paramedics. Practice change is well overdue.

This review article describes how the rigid cervical collar was introduced to practice, and explains the rationale for its inception. The same rationale is then critically analysed against a range of modern evidence challenging the validity and safety of the rigid collar. Practice changes across the world are highlighted, and relevant UK based guidance reviewed.


In 1967, Dr JD Farrington, an orthopaedic surgeon at Lakeland Memorial Hospital in Wisconsin, published a paper entitled ‘Death in a Ditch’ (Farrington, 1967). He explained that the purpose of his paper was to standardise pre-hospital training, equipment and practice which he appeared to attribute to an increased mortality\morbidity rate within his community. In particular, Farrington explains that physicians must apply a cervical collar in order to prevent the ‘sloppy and inefficient removal of a victim from their vehicle’. Whilst immobilisation principles and makeshift collars were used in the Vietnam war during the early 1960s (William Marsh Rice University, 2012), Farrington appears to be the first to recommend its practice in civilian medicine (Bledsoe, 2013).

Contrasting greatly with modern articles, Farrington's paper does not refer to any definable evidence, nor does it offer even the briefest of explanations or justifications. As Farrington offers so little for us to deconstruct, critically analysing his paper is difficult. Instead, Benger and Blackham (2009) explain that the application of a rigid cervical collar is based upon the following four principles: Injured patients may have an unstable injury; further movement will exacerbate that injury; cervical collars prevent movement; application of the collar is a safe process, and can therefore be applied to a large number of patients as a precautionary measure.

The evidence

We now know that many of these assumptions are false – especially in the alert and co-operative adult patient. For example, alert and stable patients are unlikely to have a clinically significant cervical injury (Stiell et al, 2003; Deasy and Cameron, 2011) with an incidence rate as low as 0.4% for cervical fractures and 0.14% for cervical spinal cord injury (Rhee et al, 2006). Even in presence of an injury, additional, minor movement of the cervical spine is likely without consequence (Plumb and Morris, 2013) and is unlikely to cause a secondary injury (Hauswald et al, 1998; Manoach and Paladino, 2007). Even if additional movement was likely to cause an injury, rigid cervical collars don't do a very effective job of immobilising the cervical spine (Thomas, 2013) and correctly fitted rigid collars still allow for more than 30 degrees of flexion, extension and rotation (James et al, 2004). The reality is that many rigid collars are often applied sub-optimally and therefore allow for an even greater range of movement (et al, 2009). Consequently, there is no evidence that cervical spine immobilisation prevents secondary injuries (DeMond, 2010).

Rigid cervical collars have always been thought to be a relatively harmless measure and their precautionary use therefore encouraged in order to protect those few patients who have sustained an injury (Benger and Blackham, 2009). Yet, there is an increasing volume of evidence to say that the application of rigid collars can be harmful.

For example, we know that the presence and function of the cervical plexus between C1 and C4 means that c-spine injuries to this area are amongst the most catastrophic. We also know that patients with head injuries are eight times more likely to have a C1-C4 injury (Morris et al, 2004; Sporer, 2012) and that current practice likely requires immobilisation with a cervical collar if a head or neck injury is suspected. Yet, we also know that rigid collars increase intracranial pressure (Davies et al, 1996; Mobbs et al, 2002; Stone et al, 2010) and that they forcibly increase the distance between C1 and C2 (Ben-Galim et al, 2010) – both of which can be hugely detrimental to patients with head or high c-spine injuries (Kolb et al, 1999). Additionally, this upwards movement significantly reduces mouth opening and therefore hinders definitive airway management, increasing the likelihood of choking and aspiration (Goutcher, 2005), which is particularly concerning given that airway compromise remains a leading cause of preventable death in trauma patients (Kwan et al, 2001).

Ben-Galim et al (2010) identified that in general, the application of collars had the effect of pushing the head away from the shoulders and specifically noted an average increase of 7.3 mm between C1 and C2. Although their study was conducted on fresh cadavers, this phenonomen was first reported in real patients undergoing diagnostic imaging (Bledsoe, 2015). This is not insignificant; severe neurological deterioration and even death has been reported in patients with ankylosing spondylitis as a direct result of immobilisation (Papadopoulos et al, 1999), and even those with generalised conditions such as spinal stiffness are at risk (Thumbikat et al, 2007). Children too, are at additional risk, with cervical collar complications occurring in as many as 10% of severely injured paediatric patients (Chan et al, 2013).

Even without these additional risk factors, when rigid collars are used in conjunction with full length spinal immobilisation equipment, respiration is restricted (Sugarman and Totten, 1999) and the application of immobilisation equipment delays treatment and transport to definitive care, which may increase morbidity and mortality (Morrissey, 2013). It also causes unnecessary discomfort, pain, distress and anxiety (Biller and Jacobsen, 2012) which reduces the effectiveness of trauma assessment at scene, during transport and at hospital (Hauswald and Braude, 2002). Patients who are immobilised are more likely to develop additional pain (Kwan and Bunn, 2005; March et al, 2002) which means they're also more likely to be imaged unnecessarily (Bledsoe, 2013; Hauswald and Braude, 2002) – which itself isn't without consequence.

In most countries, road traffic collisions are the main cause of cervical spine injuries (Chen et al, 2013) and the application of a rigid cervical collar has been used for decades as part of a wider extrication process, with the intent to limit cervical spine motion. But again, evidence for this practice is poor (Dixon, O'Halloran and Cummins, 2013) and modern evidence argues that alert, stable patients are able to limit mobility of their cervical spine best whilst extricating themselves (Cowley, 2014; Dixon et al, 2015), with some studies citing four times more cervical spine movement during extrication than controlled self-extrication (Dixon, O'Halloran and Cummins, 2013). Other studies highlight that self-extrication is far faster than emergency service led mechanical extrication.

Whilst some highlight that the presence of a cervical collar may serve as a useful reminder to clinicians that a spinal injury is suspected, others argue that there are better ways of reminding staff and that a soft collar would serve this purpose well without many of the complications provided by their rigid counterparts (Bledsoe, 2015).

Of course, weakness does exist in the range of modern evidence, for example, James et al (2004) state that healthy participants used “active movement” in order to achieve the 30 degrees range of motion they describe whilst wearing cervical collars. The authors do not appear to define the term active movement and recognise that the force applied was therefore subjective and unregulated. Critics could argue both that the injured patient is unlikely to use a comparative level of strength, and that an immobilisation device should, nor could not prevent the determined actions of a patient.

International context

Although further research is clearly required, ethical considerations mean that definitive level 1 evidence is unlikely to be easily achievable (Deasy and Cameron, 2011). However, an assessment of current evidence suggests that a practice change is strongly warranted (Sundstrøm et al, 2014) and EMS services around the world are beginning to respond.

Services in Norway, the Netherlands, Northern California and Australia have all abandoned use of rigid cervical collars pre-hospitally (Emergency Care Institute, 2016). The Australian Queensland ambulance service (QAS) is a particularly interesting example; in 2014, they undertook their own comprehensive evidence-based review, within which, they directly compared the evidence base for rigid and soft collars, noting that the weight of evidence demonstrates that soft collars ‘provide minimal or no significant restriction in cervical motion’ (QAS, 2014).

The report recommended cessation of rigid collar use, and instead advocated the use of soft collars. Having already stated that soft collars provide no immobilisation benefit whatsoever, they explain the purpose of soft collars is purely to act as a reminder to patients to self-protect their neck, and as a visual cue to emergency department staff indicating the need for spinal clearance. QAS's subsequent clinical practice procedures further justify the abandonment of rigid collars, stating that they lead to ‘significant complications and morbidity when used to immobilise the c-spine’ (QAS, 2016).

UK Context

The UK has an enviable history of leading evidence-based practice, but it is generally accepted that if the rigid cervical collar was a new device, the current evidence base would not be sufficient to support its introduction to practice. In the meantime, collars continue to be applied because it remains a protocol based instruction in all of the major UK pre-hospital guidelines from organisations such as the Joint Royal Colleges Ambulance Liaison Committee (Brown et al, 2016), National Institute for Clinical Excellence (NICE, 2016) and is present in the Pre-hospital Trauma Life Support (PHTLS, 2010) and the International Trauma Life Support (Campbell, 2011) syllabuses.

Much of this guidance, including that from Royal College of Surgeons (Connor et al, 2015) acknowledges the key evidence cited in this article. It accepts, and even details that the vast majority of evidence available to us tells us that cervical immobilisation isn't needed, and even if it was, highlights that rigid collars do not achieve immobility. UK guidance refers to evidence stating that cervical immobilisation using a rigid collar may complicate assessment and cause harm (Kwan, 2001). Yet, the same guidance still instructs clinicians to apply cervical collars to patients, itself admitting that the instruction is based upon ‘historical rather than scientific’ practice (Connor et al, 2015).

Historically, the presence of a protocol driven service and the lack of practice autonomy meant that UK based pre-hospital clinicians have always been taught to immobilise, rather than being taught to determine who may need immobilisation and how best to accomplish it for that patient. However, increased paramedic education based upon autonomous, evidence-based practice has enabled the profession to move from a strictly protocol driven service, to an autonomous, professional healthcare provision (Lovegrove and Davis, 2013).

Of course, scope of practice is directly proportionate to professional accountability and more of both are now afforded and expected from UK paramedics. In particular, standard four of the Health and Care Professionals Council (HCPC) Standards of Proficiency for Paramedics (HCPC, 2014) requires that paramedics act as an autonomous professional; exercise their own professional judgement; make reasoned decisions to initiate (and cease) the use of techniques or procedures and finally; recognise that they are personally responsible for the care they provide.


Current UK guidance instructs paramedics to apply rigid cervical collars despite awareness about current evidence-base. This instruction places the paramedic in a position that is incompatible with the HCPC Standards of Proficiency for Paramedics. It is also contrary to the principles of patient safety and evidence-based practice.

UK guidance must be amended to reflect the now overwhelming modern evidence base. At the very least, this means ceasing to apply rigid cervical collars to the alert and stable patient, regardless of how they score on clinical decision tools such as CCR or NEXUS. Instead, self-protection and manual inline stabilisation should be advocated. Serious consideration must also be given to use of an alternative device, such as a soft collar for patients with reduced consciousness.

In short, clinicians must be supported and enabled to provide a holistic, patient-centred assessment, utilising their education and the principles of evidence based practice to make individualised, clinical decisions firstly regarding whether or not to immobilise, and if so, how best to do so for that particular patient and their unique circumstances.