7.1. Good practice in imaging of patients with a head injury
It is assumed that general principles of good practice in imaging will be adhered to, as outlined in publications by the Royal College of Radiologists.14 On the basis of consensus, the Guideline Development Group has made the following recommendations.
All CT scans of the head should be reviewed by a clinician who has been deemed competent to review such images.
All plain radiographs of the cervical spine should be reviewed by a clinician who has been deemed competent to review such images.
Where necessary, transport or transmission of images should be used to ensure that a competent clinician review the images.
All imaging performed on patients with head injury should have a full or interim written report for the patients’ notes within an hour of the procedure having been performed.
Imaging of any kind should not delay neurosurgical or anaesthetic referral in patients with severe head injury. (D)
These recommendations are based on level five evidence and are considered to be grade D recommendations.
7.2. Urgency in performing CT of the head
Given the demands on CT scanners and radiologists trained in their use it is important to distinguish between those patients for whom CT imaging is required ‘urgently’ and those where CT can be performed ‘within a reasonable period’.
Given that it is proposed that selection for head imaging be based upon the Canadian CT-head rules, it is possible to distinguish between those patients at high risk for neurosurgical intervention (the five point rules) and those at high risk for non-neurosurgical clinically important brain injuries (the seven point rules). The former set of patients will need CT imaging to be performed urgently (that is, within one hour of the request having been received) whereas the latter patients can wait for a reasonable period (8 hours) before imaging.
[Amended] CT imaging of the head should be performed (that is, imaging carried out and results analysed) within 1 hour of the request having been received by the radiology department in those patients where imaging is requested because of any of the following risk factors:
- -
GCS less than 13 on initial assessment in the emergency department.
- -
GCS less than 15 at 2 hours after the injury.
- -
Suspected open or depressed skull fracture.
- -
Any sign of basal skull fracture (haemotympanum, ‘panda’ eyes, cerebrospinal fluid leakage from the ear or nose, Battle’s sign).
- -
More than one episode of vomiting in adults; three or more episodes of vomiting in children.
- -
Post-traumatic seizure.
- -
Coagulopathy (history of bleeding, clotting disorder, current treatment with warfarin) providing that some loss of consciousness or amnesia has been experienced; patients receiving antiplatelet therapy may be at increased risk of intracranial bleeding, though this is currently unquantified – clinical judgement should be used to assess the need for an urgent scan in these patients.
- -
Focal neurological deficit.
[Amended] Patients who have any of the following risk factors and none of the risk factors above should have their CT imaging performed within 8 hours of the injury (imaging should be performed immediately in these patients if they present 8 hours or more after their injury):
- -
Amnesia for events more than 30 minutes before impact (the assessment of amnesia will not be possible in pre-verbal children and is unlikely to be possible in any child aged under 5 years).
- -
Age 65 years or older, providing that some loss of consciousness or amnesia has been experienced.
- -
Dangerous mechanism of injury (a pedestrian struck by a motor vehicle, an occupant ejected from a motor vehicle or a fall from a height of greater than 1 metre or five stairs) providing that some loss of consciousness or amnesia has been experienced.
These recommendations are based on level two evidence and are considered to be grade B recommendations.
7.3. Cervical spine imaging urgency
The demands on X-ray facilities are not as pressing as those on CT facilities and there is no consequent need to discriminate between different categories of patient requiring cervical spine imaging. Cervical spine imaging if indicated should be carried out urgently as these patients will often need CT of the head once the cervical spine has been cleared.
[Amended] Imaging of the cervical spine should be performed within 1 hour of a request having been received by the radiology department or when the patient is sufficiently stable. Where a request for urgent CT imaging of the head (that is, within 1 hour) has also been received, the cervical spine imaging should be carried out simultaneously.
This recommendation is based on level five evidence and is considered to be a grade D recommendation.
7.4. Involving neurosurgical care
The care of all patients with new, surgically significant abnormalities on imaging should be discussed with a neurosurgeon. The definition of ‘surgically significant’ should be developed by local neurosurgical centres and agreed with referring hospitals. An example of a neurosurgical referral letter is shown in Appendix L.13.
This recommendation is based on level five evidence and is considered to be a grade D recommendation.
Examples of abnormalities not surgically significant have been produced by a survey of neuroradiologists and emergency physicians in Canada.25 However, these criteria have not to date been accepted by UK neurosurgeons, and a survey carried out in 2003 by the Society of British Neurological Surgeons found substantial concern about the Canadian criteria. The UK survey was carried out specifically to complement the development of this guideline. It would be desirable if the criteria to be used in this area could be based on the opinion of UK neurosurgeons.
7.4.1. Recommendations for research
The GDG identified the following priority areas for research in the original guideline as well as in this update.
7.4.1.1. Research Question
Research is needed to develop consensus on criteria for lesions not currently considered to be surgically significant following imaging of a patient with head injury.
Although most neurosurgeons agree about which extradural and subdural haematomas should be removed, there is controversy about whether or not to remove traumatic intracerebral haemorrhage (TICH) and cerebral contusions (CC). A prospective randomised controlled trial (PRCT) should be set up to discover if early surgery improves the outcome in these lesions compared to initial conservative treatment.
7.4.1.2. Why this research is important
One option in the management of traumatic intracerebral haemorrhage (TICH) and cerebral contusions (CC) is to monitor the patient clinically or with intracranial pressure monitoring and other forms of brain tissue monitoring such as brain tissue oxygen (BtO2) or microdialysis. When the patient deteriorates, he or she is rushed to the operating theatre. The problem is that this approach has never been validated in a prospective randomised controlled trial (PRCT). Waiting until there is deterioration in the level of consciousness (LOC) or until there is deterioration in the monitoring parameters builds delay into the management and results in secondary brain damage occurring and becoming established before surgery in all such cases. The principle of early surgical evacuation of spontaneous intracerebral haemorrhage (SICH) has been investigated in the surgical trial in intracerebral haemorrhage (STICH) and reported in the Lancet (2005). The results of such a PRCT in TICH would fundamentally alter the recommendations made by NICE, in terms of which patients should be referred to neurosurgery and, more importantly, how they should be managed there. There is no level 1 evidence about what to do with these patients and the need for such a PRCT in head injured patients is urgent. This research question should immediately be put to UK Research Funding bodies.
7.5. Other reasons for discussing a patient’s care with a neurosurgeon
Other criteria for discussing a patient’s care with a neurosurgeon were developed by both Guideline Development Group consensus and recommendations from previous guidelines.13
Regardless of imaging, other reasons for discussing a patient’s care plan with a neurosurgeon include:
- -
persisting coma (GCS less than or equal to 8) after initial resuscitation.
- -
unexplained confusion which persists for more than 4 hours
- -
deterioration in GCS score after admission (greater attention should be paid to motor response deterioration)
- -
progressive focal neurological signs
- -
a seizure without full recovery
- -
definite or suspected penetrating injury
- -
a cerebrospinal fluid leak.
This recommendation is based on level five evidence and is considered to be a grade D recommendation.
7.6. Criteria for neurosurgical interventions
These guidelines assume best practice will be followed once neurosurgeons have become involved with a particular patient. The exact nature and timing of the interventions is beyond the scope of the guidelines.
7.7. Transfer from secondary to tertiary care settings
The risk of a further injury to patients during transfer to tertiary care is well established.128 In the previous guideline transfer of the patient between a general hospital and a neurosciences unit were advised to follow the principles set out by the Neuroanaesthesia Society of Great Britain and Ireland and the Association of Anaesthetists of Great Britain and Ireland.129 The recommendations are listed below see section 7.9.7 with slight modifications to wording so that they fit the style of these guidelines. The PaCO2 targets recommended for intubated patients are based on recent literature in this area.130–132 Since the original guideline there has been an update of the guidance from the Association of Anaesthetists133 which has been reviewed in this update and recommendations have been revised accordingly see section 7.8.6.
7.8. What are the benefits for patients of receiving treatment at a neurosciences centre who have suffered a clinically important brain injury that does not require surgical intervention?
7.8.1. Introduction and rationale for the clinical question
There is no uncertainty about the management of patients with operative lesions; they must be transferred to the neurosciences unit for their operation. However, there is concern that patients who have suffered a clinically important brain injury, who are initially referred to an emergency department but do not have an operable lesion, may have a poorer outcome if they are not referred to a neurosciences centre. The dilemma for hospital staff at the DGH is whether to keep the patients at that location or to transfer them to a neurosciences unit to continue with their treatment. This question is relevant for clinicians at both types of hospitals. It is important to address whether the patient will receive better non – operative treatment if they are transferred to a specialist neurosciences centre than if they remained at the initial DGH.
An emergency department is described as a local, regional district general hospital with no neurosciences unit or a non specialist centre whereas a neurosciences unit is described as a specialist centre or a unit that has neurosurgical and neurointensive care facilities.
The main outcome measures for including studies in this review were mortality, neurological outcome, disability and hospital duration and at least one of these outcomes were reported in the studies. Studies were excluded where;
data on head injury patients were not provided,
the patient group was less than 50% head injured patients,
intervention was pre hospital care rather than transfer and
the outcomes reported only duration of transfer and no other outcomes.
7.8.2. Clinical evidence
One study134 was identified that looked at interhospital transfer (secondary transfer from one hospital to another). Three additional studies66,135,136 looked at direct transport from the injury scene to a DGH or transfer to a neurosciences unit from a DGH.
The first study134 a prospective observational study (level 2+ evidence) included patients of any age who were injured by blunt trauma between 1996–2003 (n=6921). These patients were treated by participating hospitals in the Trauma Audit and Research Network (TARN), in the United Kingdom. The intervention group (n=4616) patients received care at a neurosurgical centre (including those who had been transferred which was 53% (2677/4982)). The control group (n=2305) patients received all their care in hospitals without neurosurgical facilities on site. The mortality rate for all patients that were transported to a neurosciences unit was 35% (95% CI, 34–37%) and for those that were transported to the emergency department was 61% (95% CI, 59–63%). The mortality rate for the subgroup (n=894) of patients with isolated, non-surgical severe head injury who were transported to a neurosciences unit was 26%, (95% CI, 22–29%) and for those that were transported to the emergency department the rate was 34% (95% CI, 39–40%), p=0.005.
The second study66 a retrospective observational cohort study (level 2+ evidence) examined the issue of bypass, which obtained data from the New York State Trauma Registry from 1996–1998. The population consisted of adults more than 13 years of age with a GCS less than 14. A sub group of 2763 head injured patients from the data set of 5419 trauma patients was analysed. The patients in the intervention group (n=1430 (51.8%)) were transported to a regional trauma centre. These patients were assessed via the American Triage system (pre hospital care) and referred directly to the emergency department of a regional centre. The comparison group (n=1333 (48.2%)) were transferred to an area/non trauma centre. These patients were assessed via the American Triage system (pre hospital care) and referred directly to either an area centre or a non trauma centre. The mortality for transfer to regional centre versus non trauma centre was OR of 0.67 (95% CI, 0.53–0.85).
In another study135, a low quality study (level 3 evidence), where patients were directly transported to neurosurgical care or secondarily transferred from a DGH, the population group were neurosurgical unit patients with an extradural haematoma requiring surgery (n=104). Group 1 patients (n=71) had a mean age of 22 years (±2SE) were directly transported to a neurosurgical centre. Group 2 patients (n=33) had a mean age 20years (±3SE) and were transferred from the DGH to a neurosurgical centre. The results using the Glasgow Outcome Scale (GOS) show that mortality in group 1 was 4% (3/71) and in group 2 was 24% (8/33). The moderate/severe disability in group 1 was 10% (7/71) and group 2 was 27% (9/33). Recovery was good in 86% (61/71) of group 1 patients and 49% (16/33) in group 2, with p≤0.0002.
The final study136 was a well designed cohort study (level 2++ evidence) looking at mortality outcomes between patients directly transferred to a trauma centre and those who were transferred first to a non-trauma centre, and then on to a trauma centre. This cohort study included severely traumatic brain injured patients. The data was collected as part of a multi-centre online database designed to track pre-hospital and in-hospital severe TBI patient data, called TBI-trac. All patients passing through the trauma centres were included, and selection criteria were applied. Therefore, out of 1449, only 1123 patients were included; the remainder were excluded on the basis of a well-defined criterion, which included the mechanism of injury, death, brain death, or otherwise not benefiting from the care on offer. The authors compared, using a logistic regression model, two-week mortality outcomes between patients directly transferred to a trauma centre (n=864, 77.3%), and those who were transferred first to a non-trauma centre, and then on to a trauma centre(n=254, 22.7%). The model controlled for baseline characteristics and clinical data including hypotension status on day one, if the patient was less than or more than 60 years old, pupil status on day 1, and the initial GCS. Admission time and time by transport status were found to not affect the significance of the results. Patients were found to have a significantly lower chance of mortality with direct transfer with an odds ratio of 1.48 (CI 1.03–2.12) and p=0.04.
7.8.3. Economics Evidence from 2007 update
There was no new economic evidence for this question found in the update.
7.8.4. Summary of evidence from 2007 update
Only one study134 provides some good evidence that all patients with severe head injuries (GCS 8 or less) would benefit from receiving treatment in a neurosurgical unit irrespective of any need for a neurosurgical operation instead of receiving treatment at the emergency department. This study found data which suggests that treatment in a neurosciences centre offers a better strategy for the management of severe head injury. This study did not address direct transfer from the scene, only interhospital transfers. There is evidence135,136 which suggests good recovery, better mortality and morbidity rates amongst severely injured patients who bypass the DGH and go to the neurosciences unit. However another study66 suggests very little difference.
7.8.5. Rationale behind recommendation
A slight amendent to the previous recommendation was required (see 7.8.6). The GDG felt that there is evidence to support a recommendation for severely head injured to receive treatment in a neurosurgical unit irrespective of any need for a neurosurgical operation and have included an amendent to the recommendation below 7.8.6. The GDG agreed that the studies66,135,136 did not provide enough evidence for this question to demonstrate that all patients should be sent directly to receive treatment in a neurosurgical unit irrespective of any need for a neurosurgical operation. This is because the GDG recognises that this would require a major shift of resources of between an additional 84,000 and 105,000 bed days to neurosurgery from the existing general surgical, orthopaedic, emergency department, paediatric and geriatric services that currently care for these patients. The GDG agreed that whilst there are not enough resources for all head injury patients to go to a neurosciences centre, we should aspire to improve the rate of transfer. The GDG opinion therefore is to propose this area for further research (see section 7.9.1).
7.8.6. Recommendation
For adults
[Amended] Local guidelines on the transfer of patients with head injuries should be drawn up between the referring hospital trusts, the neuroscience unit and the local ambulance service, and should recognise that:
- -
transfer would benefit all patients with serious head injuries (GCS ≤ 8), irrespective of the need for neurosurgery
- -
if transfer of those who do not require neurosurgery is not possible, ongoing liaison with the neuroscience unit over clinical management is essential.
[NEW] The possibility of occult extracranial injuries should be considered for the multiply injured adult, and he or she should not be transferred to a service that is unable to deal with other aspects of trauma.
There should be a designated consultant in the referring hospital with responsibility for establishing arrangements for the transfer of patients with head injuries to a neuroscience unit and another consultant at the neuroscience unit with responsibility for establishing arrangements for communication with referring hospitals and for receipt of patients transferred.
[Amended] Patients with head injuries requiring emergency transfer to a neuroscience unit should be accompanied by a doctor with appropriate training and experience in the transfer of patients with acute brain injury. The doctor should be familiar with the pathophysiology of head injury, the drugs and equipment they will use and with working in the confines of an ambulance (or helicopter if appropriate). They should have a dedicated and adequately trained assistant. They should be provided with appropriate clothing for the transfer, medical indemnity and personal accident insurance. Patients requiring non-emergency transfer should be accompanied by appropriate clinical staff.
The transfer team should be provided with a means of communication with their base hospital and the neurosurgical unit during the transfer. A portable phone may be suitable providing it is not used in close proximity (that is, within 1 metre) of medical equipment prone to electrical interference (for example, infusion pumps).
[Amended] Although it is understood that transfer is often urgent, initial resuscitation and stabilisation of the patient should be completed and comprehensive monitoring established before transfer to avoid complications during the journey. A who is patient persistently hypotensive, despite resuscitation, should not be transported until the cause of the hypotension has been identified and the patient stabilised.
All patients with a GCS less than or equal to 8 requiring transfer to a neuroscience unit should be intubated and ventilated as should any patients with the indications detailed in the recommendation below.
[Amended] Intubation and ventilation should be used immediately in the following circumstances:
- -
Coma – not obeying commands, not speaking, not eye opening (that is, GCS ≤ 8).
- -
Loss of protective laryngeal reflexes.
- -
Ventilatory insufficiency as judged by blood gases: hypoxaemia (PaO2< 13 kPa on oxygen) or hypercarbia (PaCO2> 6 kPa).
- -
Spontaneous hyperventilation causing PaCO2 < 4 kPa.
- -
Irregular respirations.
[Amended] Intubation and ventilation should be used before the start of the journey in the following circumstances:
- -
Significantly deteriorating conscious level (one or more points on the motor score), even if not coma.
- -
Unstable fractures of the facial skeleton.
- -
Copious bleeding into mouth (for example, from skull base fracture).
- -
Seizures.
[Amended] An intubated patient should be ventilated with muscle relaxation and appropriate short-acting sedation and analgesia. Aim for a PaO2 greater than 13 kPa, PaCO2 4.5 to 5.0 kPa unless there is clinical or radiological evidence of raised intracranial pressure, in which case more aggressive hyperventilation is justified. If hyperventilation is used, the inspired oxygen concentration should be increased. The mean arterial pressure should be maintained at 80 mmHg or more by infusion of fluid and vasopressors as indicated. In children, blood pressure should be maintained at a level appropriate for the child’s age.
Education, training and audit are crucial to improving standards of transfer; appropriate time and funding for these activities should be provided.
Carers and relatives should have as much access to the patient as is practical during transfer and be fully informed on the reasons for transfer and the transfer process.
These recommendations are based on level five evidence and are considered to be grade D recommendations.
7.9. Transfer of children
The recommendations in section 7.8.6 above were written for adults but the principles apply equally to children and infants, providing that the paediatric modification of the Glasgow Coma Scale is used.
Service provision in the area of paediatric transfer to tertiary care should also follow the principles outlined in the National Service Framework for Paediatric Intensive Care. These do not conflict with the principles outlined in section 7.5 above137.
Transfer of a child or infant to a specialist neurosurgical unit should be undertaken by staff experienced in the transfer of critically ill children.
Families should have as much access to their child as is practical during transfer and be fully informed on the reasons for transfer and the transfer process.
These recommendations are based on level five evidence and are considered to be grade D recommendations.
[NEW] The possibility of occult extracranial injuries should be considered for the multiply injured child, and he or she should not be transferred to a service that is unable to deal with other aspects of trauma.
7.9.1. Recommendations for research
The GDG also identified the following priority areas for research.
7.9.1.1. Research Question
Do patients with significant traumatic brain injury who do not require operative neurosurgical intervention at presentation, but are still cared for in specialist neurosciences centres, have improved clinical outcomes when compared to similar patients who are treated in non-specialist centres?
7.9.1.2. Why this research is important
Traumatic brain injury (TBI) is amongst the most important causes of death in young adults, with an overall mortality for severe TBI of over 50%. TBI care consumes one million acute hospital bed-days, and over 15,000 ICU bed-days annually, and patients who do survive significant TBI experience an enormous burden of long term physical disability, neurocognitive deficits, and neuropsychiatric sequelae. The financial impact is significant: the NHS spends over £1 billion on just the acute hospital care of the 10,000 patients with significant TBI. The costs of rehabilitation and community care are difficult to estimate, but probably total many multiples of the figure provided for acute care. These considerations make TBI a national healthcare priority and its outcome impact is consistent with its inclusion in the National Service Framework for Long Term Neurological Conditions.
Current referral of patients with acute traumatic brain injury practice is still dominated in many parts of the United Kingdom by the need for operative neurosurgical intervention at presentation. This may be inappropriate, since many patients with severe head injury have evidence of raised intracranial pressure in the absence of surgical lesions, and suffer morbidity and mortality equal to those with surgical lesions. Further, several studies provide strong circumstantial evidence that managing such “non-surgical” patients in specialist neurosciences centres may result in substantial improvements in mortality and functional outcome, probably due to specialist expertise in areas of non-operative management, such as neurocritical care. However, these results may be confounded by case-mix effects and referral bias, and the cost-effectiveness of such specialist management remains uncertain. There is a strong case to address this question in the context of a formal study, since a change in practice could have a major impact on death and disability in a condition that is a major contributor to mortality in healthy young adults. Importantly, the results of such a study could fundamentally alter the recommendations made by NICE, in terms of where patients with head injury are treated within the healthcare system, and result in better optimised (and potentially more cost-effective) patient flows within the NHS.
The available evidence in this area has been addressed in the systematic review that contributed to the revision of NICE Guidelines on the early management of head injury. This review could find no high quality clinical evidence on the topic. This is unsurprising, since any study that addressed these issues would have to be undertaken within the context of a healthcare system and include ambulance services, district general hospitals and neuroscience referral centres. Such a study would therefore require the organisational backing of a body such as NICE and careful design to account for confounds and biases. However, we believe that given careful design, such a study would be both ethically and logistically feasible. The patient group is well defined, and adequate numbers would be available to provide a definitive result within a reasonable time frame. While circumstantial evidence may support transfer of such patients to neurosciences centres, current practice is not influenced by this view in many regions, and many would argue that there is still clinical equipoise in this area. There are clear risks from transfer, and there could be clear harm, both in terms of clinical outcome and health economics, if the anticipated benefits were not realised. On the other hand, if the benefits from observational studies were confirmed by the trial, the resulting changes in management could potentially reduce case-mix adjusted mortality by 26% and increase the incidence of favourable outcome in survivors by nearly 20%.
