Head injury: assessment and early management

People 16 and over

There was evidence from 2 RCTs for tranexamic acid, both of which included adults with no suspicion of extracranial bleeding. One trial was in a pre-hospital (out-of-hospital) setting and the other was in a hospital setting. There was no evidence available for people under 16 (1 trial included a few people aged 15 and 16).

In the pre-hospital setting, a single 2 g bolus dose of tranexamic acid was given within 2 hours of a head injury. The evidence suggested that it reduced all-cause mortality at 28 days and 6 months in people with moderate or severe traumatic brain injury. But there was no clinically important difference between tranexamic acid and placebo for hospital-free days at 28 days, degree of disability at discharge and after 6 months (Glasgow Outcome Scale–Extended more than 4), and serious adverse events (that is, myocardial infarction, pulmonary embolism, deep vein thrombosis and stroke).

In the hospital setting, a 1 g bolus dose of tranexamic acid was given within 3 hours of injury, followed by an 8‑hour intravenous infusion of 1 g of tranexamic acid. The evidence suggested that this reduced death from traumatic brain injury at 28 days in groups with mild or moderate traumatic brain injury. Evidence suggested reduced mortality related to severe traumatic brain injury at 28 days in high-income countries, but there was no difference in mortality related to traumatic brain injury in low- and middle-income countries. No evidence was available separating the mild from moderate severity groups for mortality related to traumatic brain injury and all-cause mortality, but communication with the study authors suggested significant uncertainty about tranexamic acid's effect in mild traumatic brain injury. There was no clinically important difference between tranexamic acid and placebo for serious adverse events (that is, myocardial infarction, pulmonary embolism, deep vein thrombosis and stroke) and disability rating scale scores in a mixed severity group (mild, moderate and severe) across all income groups.

The committee considered that, despite the uncertainty in the clinical evidence, there was a benefit with tranexamic acid in terms of reducing all-cause mortality and mortality from traumatic brain injury. They also considered that the evidence showed that it caused very few adverse events. Based on the evidence, they agreed that a 2 g intravenous bolus dose of tranexamic acid, given within 2 hours of a head injury and before imaging, could be considered for people 16 and over with moderate or severe traumatic brain injury. The committee recommended a 2 g intravenous bolus injection of tranexamic acid because this dose was found to be the most safe and effective.

An economic model was developed that looked at each traumatic brain injury severity subgroup separately. It was based primarily on the pre-hospital clinical trial because:

For moderate traumatic brain injury, the health benefits associated with a 2 g bolus given in the pre-hospital setting outweighed the costs in all scenarios. For severe traumatic brain injury, the cost-effectiveness estimate was more borderline. But, when limitations in the modelling were taken into account, the committee concluded that the health benefits were likely to outweigh the costs in this group too.

People under 16

Because of a lack of trial evidence for tranexamic use in people under 16, the committee used extrapolated evidence from the trials in adults, and their expertise and knowledge. In NHS clinical practice, a tranexamic acid dose of 15 mg/kg is used in people under 16 with extracranial injuries. But, in this age group, tranexamic acid is not currently widely used for isolated head injury and dosing is variable (15 mg/kg to 30 mg/kg). Evidence for people 16 and over with a head injury from a pre-hospital setting suggested that a 2 g dose of tranexamic acid reduced all-cause mortality (at 28 days and 6 months), with no evidence of negative effects. So, the committee concluded that it could recommend the equivalent of a 2 g dose of tranexamic acid for people under 16. They discussed that, based on the average weight of people 16 and over being 70 kg, a 2 g dose of tranexamic acid for people 16 and over would equate to a 30 mg/kg dose for people under 16. So, they concluded that a dose range of 15 mg/kg to 30 mg/kg was appropriate for people under 16.

For people with mild traumatic brain injury, it is less clear whether the benefits of tranexamic acid outweigh the potential risk of blood clots. So, the committee made a recommendation for research on tranexamic acid for this group.

People 16 and over

Several diagnostic accuracy studies were identified but there were no diagnostic RCTs. Most of the evidence was of low to very low quality and was in people with a mild head injury (defined as a GCS score of 13 to 15 in many studies, but sometimes limited to a GCS score of 14 to 15). The committee noted that the existing recommendations for clinical decision rules for head imaging in people 16 and over were largely based on the Canadian CT Head Rule (CCHR). This involves identifying high and medium risk factors, with some modifications aimed at improving sensitivity. Updated evidence for this decision rule showed that it has good sensitivity when used as intended but that its specificity values are poor. The committee noted that specificity values of decision rules are often low because they prioritise very high sensitivity.

Evidence identified for other decision rules showed sensitivity values similar to those of the CCHR. However, specificity values were lower compared with the CCHR high and medium risk rule. Limited evidence showed that the NEXUS 2 decision rule has specificity values similar to those of the CCHR.

Only 1 study had assessed the performance of the 2014 update of NICE's head injury guideline recommendations for head imaging. This reported sensitivity values that were poorer than those for the CCHR. But the specificity values were better compared with other decision rules. The committee agreed that it was unclear why the sensitivity of the NICE recommendations would be worse than those of CCHR. They also agreed that it was unclear why the CCHR did not do as well in this study as in other studies. They thought that this possibly suggested some differences between study populations, which may have affected the results. The committee also agreed that, in their clinical experience, the sensitivity of the NICE recommendations was not as low as suggested in this study.

The committee agreed that there was insufficient evidence to support changing the clinical decision rule recommendations for head imaging in people 16 and over. Because the NICE recommendations were largely based on the CCHR rule, this decision was further supported by cost-effectiveness evidence. This showed the CCHR rule to be the most cost effective of the multiple decision rules assessed.

People under 16

The committee noted that the existing recommendations for clinical decision rules for head imaging in people under 16 were largely based on the CHALICE rules with some modifications. These modifications were based on current practice and experience allowing for the option of an observation period with imaging if the condition of some people deteriorated, rather than immediate imaging.

Updated evidence identified for this decision rule showed that it had good sensitivity when considering clinically important injuries or neurosurgical outcomes. In the 2014 update of this guideline, the committee stated that an improvement in specificity relative to the NICE recommendations would be needed to warrant switching to another decision rule for people under 16. They noted that PECARN and CATCH‑7 may have slightly better sensitivity values compared with CHALICE, but agreed that, overall:

The committee noted that, in terms of the content of the rules, PECARN and the NICE guideline are not very different but that PECARN is vaguer and does not give timings. They thought this less useful. In addition, they noted that the PECARN and CATCH-7 rules apply to more specific populations than the NICE guideline recommendations.

The committee agreed that the recommendations in the NICE guideline are in widespread use in current practice, and used with little variation. Their opinion was that they are currently well-accepted and used with good effect. The committee therefore agreed that there was insufficient evidence to support changing the clinical decision rule recommendations for head imaging in people under 16.

The committee agreed to keep the existing recommendations in from the 2014 update of NICE's head injury guideline for people with bleeding and clotting disorders. This was because there was no new evidence to change practice in adults, and the committee agreed to extrapolate from an existing recommendation to make a new recommendation for people under 16. However, they changed the recommendation wording from 'history of bleeding or clotting disorders' to 'current bleeding or clotting disorders'. In children, some disorders are short-lived or resolve in a couple of months. In adults, a history of bleeding or clotting disorders is used to help screen people before surgery. However, this is a crude tool and may not be appropriate in this setting. So, the committee agreed to keep the changed wording for all age groups to help provide a consistent message.

People having anticoagulants or antiplatelets

There was conflicting evidence from cohort studies on whether people who are on anticoagulants or antiplatelets are at higher risk of intracranial haemorrhage than people not on anticoagulants or antiplatelets. CT scans could be limited to people with symptoms of traumatic brain injury such as loss of consciousness or amnesia. However, the committee thought that the new evidence was not strong enough to warrant stopping imaging in people with a head injury who are on anticoagulants but have no other indication for imaging. So, they decided CT scanning should be considered rather than automatically done in this group. They also agreed that antiplatelets other than aspirin monotherapy should be included in this. The review findings suggested that people on anticoagulants (including warfarin and direct-acting oral anticoagulants) or antiplatelets (excluding people on aspirin monotherapy) with low-risk factors (no loss of consciousness, amnesia, a GCS score of 15 and no other indications for CT brain scan) can be risk assessed (including for other injuries, supervision at home, cause of incident and risk of further falls). Then, if there are no risk factors and after shared decision making, they could be discharged safely without a CT scan, with the usual discharge advice (see the section on discharge and follow up). The committee highlighted that clinicians would either scan or admit someone for monitoring if any risks were identified. This might be, for example, if a person (with pre-existing cognitive impairment) may be less likely to return to the emergency department urgently if they have any signs of deterioration. The committee noted that, if an intracranial haemorrhage is not detected at initial presentation, delayed recovery is more likely rather than death. They also discussed that neurosurgical intervention for traumatic brain injury is less likely to be offered for people over 74 because risks outweigh the benefits.

In current practice, in accordance with the 2014 update of NICE's head injury guideline recommendations, a CT head scan is done within 8 hours of a head injury in people with no other indications for a CT head scan who are having anticoagulant treatment. The 2014 update did not make specific recommendations for people on antiplatelets. The committee agreed that, in clinical practice, there is variation, with some services offering imaging to people on antiplatelets. Based on the evidence, they also agreed that antiplatelets other than aspirin monotherapy should be included but did not specify which antiplatelets. This was because they did not want to be prescriptive and exclude any newer antiplatelets in development. Based on their experience and extrapolation of evidence in people presenting within 8 hours of injury, the committee agreed that these recommendations could be applicable to people presenting more than 8 hours after their injury. However, they agreed that imaging should be done within an hour of confirming that the person with head injury is on anticoagulant or antiplatelet medication.

There was limited evidence on aspirin. From their knowledge and clinical experience, the committee highlighted that the risk of intracranial haemorrhage is low with aspirin. This is even so in people with neurological symptoms such as loss of consciousness or amnesia. So, they agreed that people on aspirin monotherapy could be discharged without a CT head scan after shared decision making if there is no other indication for a CT brain scan or hospital admission.

The committee made a recommendation for research on risk of bleeding for people with pre-injury coagulopathy.

People with liver or coagulopathy disorders

There was no evidence for people with liver or coagulopathy disorders. Current practice is variable, with some services offering imaging to people with liver disease who have no symptoms. People with liver or coagulopathy disorders are at increased risk of bleeding, although some people will have a tendency for increased clotting. People with acquired coagulation defects can be a heterogenous and complex group. They can include people with acquired haemophilia through to people with other abnormalities such as disseminated intravascular coagulation.

The committee agreed to keep the existing recommendations from the 2014 update of NICE's head injury guideline for people with bleeding and clotting disorders. This was because there was no new evidence to change practice. But they changed the recommendation wording from 'history of bleeding or clotting disorders' to 'current bleeding or clotting disorders'. In people under 16, some disorders are short-lived or resolve in a couple of months. In adults, a history of bleeding or clotting disorders is used to help screen people before surgery. However, this may not be appropriate in the emergency department setting. So, the committee agreed to keep the changed wording for all age groups to help provide a consistent message.

People with pre-injury cognitive impairment who have a head injury through low-energy impact or low-level falls

Limited evidence from cohort studies suggested that, in people 16 and over, falling from a standing position, being over 70, having a reduced GCS score compared with normal, taking antiplatelet medication including aspirin, and having neurological symptoms (loss of consciousness, vomiting after fall) were risk factors associated with the diagnosis of an intracranial bleed. Anticoagulant medication in this population was not associated with an intracranial bleed. It was not clear if people in the studies had pre-injury cognitive impairment, so the applicability of this evidence is limited.

The committee discussed the challenges in assessing risk in people with cognitive impairment. For example, people with dementia may under-report or may be unaware of symptoms such as loss of consciousness or amnesia. It is also difficult to differentiate head injury symptoms from the pre-existing dementia in these people. There was no evidence for people under 16.

The committee acknowledged the limited evidence for this group. They made a recommendation for research on indications for imaging for people with pre-injury cognitive impairment.

People with recurrent head injuries

There was no evidence for recurrent head injuries in any age group. Recurrent head injuries could occur in people with epilepsy, people with mobility issues at high risk of falls and with some sports activities. Particularly in the context of sports injuries, these injuries can be repeated and lead to cumulative risks. Because of a lack of evidence, the committee decided to make a recommendation for research on indications for imaging for people with a history of recurrent traumatic head injuries.

People presenting more than 24 hours after a head injury

Evidence from 1 observational study suggested that there was an increased risk of any traumatic brain injury or important traumatic brain injury on a CT head scan in babies and children under 2 years:

There was no evidence for people 2 years or over. The committee discussed that people 16 and over presenting more than 24 hours after injury have increased risk factors such as vomiting and loss of consciousness. This is because they would be attending because of worsening symptoms. The 2014 update of NICE's head injury guideline recommendations are for people presenting within 24 hours of injury. But, because of a lack of evidence, the committee agreed that these could be extrapolated to people presenting more than 24 hours after a head injury (see the recommendations on the criteria for doing a CT head scan). They also agreed that this was an important area, so proposed a recommendation for research on indications for selecting people for imaging when they present more than 24 hours after a head injury.

Post-concussion syndrome

The committee agreed that high specificity is needed for brain injury biomarkers for post-concussion syndrome. This was because the population with a mild head injury is large but only a small proportion go on to develop post-concussion syndrome. So, false positives would have a negative effect on resources if biomarkers were to be used to direct everyone towards interventions or monitoring.

Overall, the committee agreed that the evidence was too limited to be able to make recommendations for using biomarkers (including fluid biomarkers or MRI) to predict post-concussion syndrome in people with mild traumatic brain injury. There was no evidence from prognostic test-and-treat studies comparing clinical outcomes, so the committee agreed to highlight the criteria for doing a CT head scan. They also made a recommendation for research on using biomarkers and MRI for predicting post-concussion syndrome.

Using biomarkers for predicting acute complications after a traumatic brain injury

Evidence from diagnostic accuracy studies suggested that there were high sensitivity values for some biomarkers at certain thresholds for predicting acute complications after a traumatic brain injury, but the specificity values were not high enough across the evidence. Also, many biomarkers were only tested in small samples, which led to imprecise estimates. The committee noted that accuracy differed quite widely between different studies looking at the same biomarker test measured with different assays on different platforms. Also, the evidence was heterogenous, with variable thresholds and time points for different biomarkers. Most people with a head injury present to hospital within 3 hours, and the manufacturers recommend this timeframe for optimal test results. Many of the studies assessed biomarkers beyond this time point.

The committee agreed that the specificity values were equally as important as the sensitivity values, given the consequences of unnecessary radiation from CT scans. They thought this was particularly important in people under 16. But, after considering the limitations of the evidence, the committee were unable to make recommendations for using biomarkers to predict acute complications after a mild traumatic brain injury. They did think that biomarker tests had promise, so they proposed a recommendation for research on using biomarkers for predicting acute post-traumatic brain injury complications.

Small intracranial injuries

Limited evidence from cohort studies in people with intracranial injuries suggested that effect sizes for the clinical decision rules were larger overall than those for individual risk factors. So, the committee agreed that clinical decision rules were likely to be the way to identify people who should be admitted in the future. This was because they thought it would be difficult to make decisions based on individual risk factors in clinical practice. But the evidence for clinical decision rules was all retrospective. So, the committee did not think any specific clinical decision rules could be recommended, particularly because they would be new to clinical practice.

For individual risk factors, the committee noted that there was consistent evidence across all studies (including 1 study in people under 16) that GCS scores of 13 and 14 were associated with a worse outcome than a GCS score of 15. But this is already an existing indication for admission (see recommendation 1.9.1).

Evidence for specific thresholds and findings on CT (including thresholds for subdural or epidural haemorrhage size, or findings such as midline shift or mass effect on CT) also indicated larger effect sizes than for some other risk factors. But, for factors such as midline shift, the committee noted that a threshold for degree of shift would be more useful in practice. They also noted that the varying thresholds used for subdural and epidural haemorrhage across the different studies made the ideal threshold to use unclear.

The evidence also suggested that thresholds for age could be associated with a worse outcome in higher age groups. But the committee noted that older age would not solely be used in practice to make decisions about admission. This is particularly because admission in older age groups can also be associated with harms such as hospital-acquired infections. The committee agreed that age or frailty may be a concern but should not be a sole indicator for admission. Overall, the committee agreed that prospective studies are needed in people with a GCS score of 13, 14 or 15 and a head injury of any size confirmed with CT to validate existing clinical decision rules for predicting deterioration. The aim would be to refine indications for admission in this group. So, the committee made a recommendation for research on indications for admission in people with a mild head injury and a confirmed abnormality on a CT scan.

Isolated skull fracture

Evidence from several case series suggested that there was low risk of death, neurosurgery, admission to critical care, unplanned hospital admission and delayed intracranial injury in babies and children (age cut-off varied across studies) with an isolated skull fracture. The evidence suggested that there was a slightly higher risk of seizure (at presentation) and evaluation for suspected non-accidental injury in this group. According to current guidelines (recommendations 1.5.8 and 1.5.10), people with seizures and suspected non-accidental injuries will be admitted to hospital after a head injury.

Based on the evidence, the committee agreed that simple linear non-displaced fractures are not likely to be a clinically important injury. So, they agreed that, after shared decision making, people under 16 with such fractures can be safely discharged if they have normal neurological status and there are no safeguarding concerns.

There was no direct evidence for people 16 and over with an isolated skull fracture. Indirect evidence from cohort studies suggested that there was low risk of clinical deterioration from a simple skull fracture compared with:

Clinical deterioration was measured by a composite of death due to traumatic brain injury, neurosurgery, seizure, a fall in GCS score of more than 1, admission into intensive care for traumatic brain injury, intubation or hospital readmission for traumatic brain injury. The simple skull fracture group included both isolated and non-isolated skull fractures. But the committee agreed that the evidence is still likely to be broadly applicable for people 16 and over with an isolated skull fracture. Based on the evidence and their collective experience, the committee agreed that, after shared decision making, people 16 and over with an isolated skull fracture can be discharged safely (except for people on anticoagulants or antiplatelets) if there are no safety concerns.

Anticoagulants and antiplatelets

The evidence was limited on admission to and observation in hospital of people who have a head injury and are on anticoagulants (warfarin and other vitamin K antagonists, and direct-acting oral anticoagulants) or antiplatelets (including aspirin, ticlopidine, clopidogrel, prasugrel and ticagrelor), and have normal brain imaging or no indication for early imaging.

Limited evidence from non-randomised studies suggested that there was no clinically important difference between pre-injury anticoagulant or no anticoagulant treatment in terms of delayed bleeding. The evidence included at least 1 propensity matched study including about 70,000 people.

The committee agreed that the evidence was not strong enough to recommend using anticoagulation status as a sole indicator for admission for people with a negative initial CT head scan. They highlighted that admission based solely on this could cause harm in people already vulnerable (because of, for example, frailty, an underlying condition such as delirium or risk of hospital-acquired infections). This is particularly so if there is not a large increase in risk of delayed bleeding. Also, when these events do occur, they are usually not clinically important.

Evidence from non-randomised studies for antiplatelet comparisons was more limited than that for anticoagulants. There were no large studies, and all reported effects were based on a difference of only 1 to 2 events between pre-injury antiplatelet and no-antiplatelet groups per study. The committee agreed that the evidence was not strong enough to recommend using antiplatelet status as a sole indicator for admission for people with a negative initial CT or no indication for a CT. The committee did not make a recommendation for research for this group because they did not consider it to be a priority for research.

Concussion symptoms

There was no evidence on admission or discharge of people with concussion symptoms after normal imaging or no indication for imaging. The committee agreed that their discharge is based on clinical discretion, and admission is considered if non-accidental injury is suspected. From their experience, the committee also agreed that most people with concussion symptoms and normal imaging do not need further intervention and are safe to be discharged from the emergency department. They highlighted that current practice is to not admit people with concussion symptoms unless they have any of the indications in recommendation 1.9.1. The committee were unaware of any evidence indicating that current practice was causing harm (coroners reports, safety reports, patient group feedback). They also noted that someone's condition may worsen if they are admitted, for example, because of being in a noisy and unfamiliar environment, and because the risk of hospital-acquired infections could increase. When people with concussion symptoms are discharged, information is provided on when to return to hospital to seek further immediate care and ongoing support for persistent symptoms (see the recommendations on discharge advice). The committee agreed that this is important and did not think there was any evidence on which to base a change in practice.

People with pre-injury cognitive impairment

Pre-existing cognitive impairment such as dementia, Parkinson's disease or stroke was reported in some studies in people taking anticoagulants and antiplatelets, but the studies did not report the effect of pre-injury cognitive impairment on the outcomes. Examples of pre-injury cognitive impairments and neurodivergence seen in people of any age include developmental delay, Down's syndrome, cerebral palsy, fetal alcohol syndrome, a learning disability, autism spectrum disorder and other conditions with altered sensation. Examples of pre-injury cognitive impairment seen only in adults include depression, dementia and medication side effects. The committee noted from their experience that people with pre-existing conditions affecting cognition are less likely to recognise and to raise an alarm about the early signs of a late intracranial bleed (such as a severe headache, drowsiness and vomiting) than people with no pre-existing cognitive impairment. So, in current practice, a short overnight admission for observation is arranged for people with a pre-existing cognitive impairment when no supervision at home is available. The committee agreed that people with a pre-existing cognitive impairment will need to be appropriately supervised and monitored to ensure that their symptoms are not worsening if they are discharged from the emergency department. Supervision and monitoring was also noted to be important for people discharged to custodial settings. The committee noted that, at discharge, it is important for people and their carers to be given a written copy of the head injury discharge advice.