Predicting paediatric traumatic brain injuries

Cite this article as:
Dani Hall and Mieke Foster. Predicting paediatric traumatic brain injuries, Don't Forget the Bubbles, 2021. Available at:
https://doi.org/10.31440/DFTB.30993

The biggest challenge in managing a child with a mild to moderate head injury is deciding whether to organise a CT scan or not. Balancing the risk of ionising radiation (and with it the small, but definite, risk of a future brain tumour or leukaemia) against the risk of missing a significant brain injury is mitigated to some extent by using a clinical decision rule, like the PECARN, CATCH or CHALICE rules. These rules are extremely sensitive with very few false negatives and excellent negative prediction values, meaning if you follow them, you’re unlikely to miss a clinically important brain injury (cTBI). Their problem is their specificity is low with plenty of false positives, meaning most of the children who have a scan won’t actually have a brain injury. (If you’d like a refresher on sensitivity, specificity, NPV and PPV in head injury decision rules, check out Damian’s critical appraisal talks in DFTB Essentials.)

Over the last 6 years, Australasia’s PREDICT network has been a publishing powerhouse on paediatric head injuries from their Australasian Paediatric Head Injury Research Study (APHIRST for short). In their cohort of 20,000 children the team have been able to tell us that of PECARN, CATCH and CHALICE, the PECARN rule has the highest sensitivity. They’ve also shown that planned observation leads to significantly lower CT rates, with no difference in missed cTBI. And probably most telling of all, they’ve told us  that, without using any rules, their clinicians are already very good at identifying children with a cTBI with a sensitivity almost as high as PECARN’s, but with a very low baseline CT rate.

Nonetheless, clinical decision rules do play their role. And so, when they asked their network what an ideal decision rule would tell them, their clinicians highlighted the gaps in the existing guidelines: What should we do with a child with a delayed presentation up to 72 hours after the head injury? What about a child with a bleeding disorder and a head injury? What about a child with a VP shunt and a head injury? Or an intoxicated child with a head injury? The list goes on.

And so, in true PREDICT style, they decided to develop their own guideline.

This week marks a landmark day for paediatric head injury management worldwide as PREDICT launch their guideline for mild to moderate head injuries in children. The risk criteria from the PECARN rule, the best performing prediction rule in the APHIRST study, play a central role, supported by an extensive literature search, including studies from PECARN and PREDICT on the risk associated with VP shunts and bleeding risks. PREDICT have pulled all the data into one comprehensive, evidence-based guideline for managing, what has previously been considered, some of the less clear-cut paediatric head injury presentations. Let’s explore the algorithm and run through a series of cases.

Babl FE, Tavender E, Dalziel S. On behalf of the Guideline Working Group for the Paediatric Research in Emergency Departments International Collaborative (PREDICT). Australian and New Zealand Guideline for Mild to Moderate Head injuries in Children – Algorithm (2021). PREDICT, Melbourne, Australia.

How was the guideline derived?

Building on the existing high-quality clinical decision rules, the PREDICT group conducted a systematic review of the literature to include more recently published evidence. To develop the new PREDICT guideline, they used a GRADE-ADOLOPMENT approach, adopting, adapting or developing new recommendations, which are labelled in the main guideline as ‘evidence-informed recommendations’, ‘consensus-based recommendations’ or ‘practice points’.

What does it say?

This guideline is here to tell us what to do with children with a mild or moderate head injury, with a GCS of 14 or 15, or a child with a GCS ≤ 13 with a normal CT scan. The ‘who to discharge, who to observe and who to scan’ part of the guideline is succinctly summarised with a two-page algorithm. Page 1 has an easy to follow flowchart, supplemented by footnotes and Appendix with modified guidance for special conditions on page 2.

Page 1
Page 2

The bottom line

What I like so much about this guideline is that it answers so many of our “what about the child with a head injury plus…?” questions. With the evidence-based recognition that senior clinicians who choose to observe rather than scan a child reduce the CT rate without increasing the number of missed cTBIs, this guideline also allows senior clinicians to make a risk assessment on a case by case basis, while remaining fluid enough to upgrade or downgrade a child’s risk if their clinical picture changes. Although designed for use in Australia and New Zealand, I can see it being immensely useful outside Australasia and am looking forward to putting its pearls of wisdom to use.

Case 1

Case 2

Case 3

Case 4

Case 5

Cases 6 and 7

Case 8

Cases 9 and 10

Case 11

Case 12

Case 13

Case 14

Case 15

References

 Babl FE, Tavender E, Dalziel S. On behalf of the Guideline Working Group for the Paediatric Research in Emergency Departments International Collaborative (PREDICT). Australian and New Zealand Guideline for Mild to Moderate Head injuries in Children – Algorithm (2020). PREDICT, Melbourne, Australia.

Babl FE et al. Accuracy of PECARN, CATCH, and CHALICE head injury decision rules in children: a prospective cohort study. 2017. 389;10087:2393-2402. DOI: https://doi.org/10.1016/S0140-6736(17)30555-X

Babl FE et al. A prospective observational study to assess the diagnostic accuracy of clinical decision rules for children presenting to emergency departments after head injuries (protocol): the Australasian Paediatric Head Injury Rules Study (APHIRST). BMC Pediatr. 2014. 13;14:148. DOI: 10.1186/1471-2431-14-148

Singh S et al. The Effect of Patient Observation on Cranial Computed Tomography Rates in Children With Minor Head Trauma. Acad Emerg Med. 2020. 27:832–843. DOI: 10.1111/acem.13942

Borland M et al. Delayed Presentations to Emergency Departments of Children With Head Injury: A PREDICT Study. Ann Emerg Med. 2019. 74:1-10. DOI: 10.1016/j.annemergmed.2018.11.035

Head Injuries Module

Cite this article as:
Team DFTB. Head Injuries Module, Don't Forget the Bubbles, 2020. Available at:
https://doi.org/10.31440/DFTB.27374
TopicHead injuries
AuthorChris Odedun
DurationUp to 2 hours
Equipment requiredNone
  • Basics (10 mins)
  • Main session: (2 x 15 minute) case discussions covering the key points and evidence
  • Advanced session: (2 x 20 minutes) case discussions covering grey areas, diagnostic dilemmas; advanced management and escalation
  • Sim scenario (30-60 mins)
  • Quiz (10 mins)
  • Infographic sharing (5 mins): 5 take home learning points

We also recommend printing/sharing a copy of your local guideline.

Basics of head injury assessment 

RCEMLearning module on head injury (September 2018)

Managing more serious head injuries 

OPENPediatrics “Introduction to Traumatic Brain Injury” (February 2016)

Neuroprotective strategies for severe traumatic brain injury (Paediatric FOAM)

DFTB “Traumatic Brain Injury” (2013)

Your department/region’s guideline for managing head injuries in children.

Head injuries form a wide spectrum of clinical presentations. At their most simple, they can be defined as any impact to the body, proximal to the cervical spine & neck, excluding trivial impact to the face. Practitioners seeing any patient with head injuries should devote time to understanding the primary injury – the mechanism, including its biomechanics. 

They should also aim to develop expertise at identifying the cohort of patients at risk of secondary injury, from deviations of ICP, blood pressure, CO2, O2 & glucose.

Head injuries are generally defined by conscious level (Glasgow Coma Score/GCS) post-injury.  Head injuries are a very common presentation for children to emergency departments. The vast majority are trivial or minor, requiring observation and/or discharge advice only.

Head injuries remain one of the most common causes of serious morbidity & mortality in children (and young adults). Practitioners need to become skilled at selecting the cohort who require imaging – which is well established as CT. This is the best modality commonly available to detect more serious injuries – typically contusions, intra/extra-cerebral bleeds & skull fractures. Practitioners should become familiar with clinical guidelines & decision-support resources (eg. NICE) to help guide which patients need imaging.

An even smaller proportion of these injured children will go on to require neurosurgical intervention. Here, the practitioner’s role is to mitigate secondary injury, as above – with neuroprotective strategies.

Lastly, practitioners should be aware of the possibility of non-accidental injury, especially with regard to drowsy or unconscious infants, and remember that they have a role in safeguarding all children presenting to the ED, regardless of reason for presentation.

NICE clinical guideline CG176 – head injury: assessment & early management esp. 1.3 & 1.4.9, 10, 11 + this review of the 2014 changes to indications for CT and more [Tessa Davis, Anna Ings (BMJ)]

NICE clinical guideline CG176 – head injury: assessment & early management esp. 1.3 & 1.4.9, 10, 11 + this review of the 2014 changes to indications for CT and more [Tessa Davis, Anna Ings (BMJ)]

CT imaging became the imaging modality of choice during the 2000s/2010s in developed economies.  Since then, access to CT has generally widened, and become ubiquitous. Key to this in the UK has been the development of guidance by NICE, in 2014, with updates since – it guides management of head injury in children and adults.  A trio of cohort studies looking at outcomes of children with head injury were key to the development of the paediatric part of this guideline – please see references. Included below is a flowchart guiding CT use in children courtesy of NICE.

CHALICE (UK/2006) – highly sensitive but significantly less specific rule developed in the UK, later incorporated into the NICE guidance

PECARN (US/2009) – cohort study looking to identify low-risk group of paediatric patients who could safely not be imaged
CATCH (Canada/2010) – prospective multicentre cohort study from Canada looking to establish features for medium & high-risk for clinically significant traumatic brain injury

A 6 year old girl is brought in by ambulance to the ED you work in. She was playing on a climbing frame and fell off the top onto concrete, onto her head. Handover states that she was briefly knocked unconscious, then returned to a GCS = 15, but has become more drowsy en route to hospital.

On your initial assessment, there is a large swelling to the left side of her scalp and forehead, and there appears to be some blood leaking from her left ear. Her GCS is 12 (E3V4M5) but the rest of her vital signs are within normal limits.

Outline your management steps.

How soon do you want this child to have CT imaging?

The scan shows an extradural haematoma. How can you direct your team to prevent secondary brain injury?

  • This child needs immediate CT imaging of the head and their cervical spine – they are ideally managed by a trauma team, where the primary survey should ensure detection of any other injuries. If as likely, the cervical spine cannot be cleared clinically, they will need immobilisation until this is completed. A written report from an appropriate radiologist will ideally be available within 60mins of the scan.
  • Significant CT findings (see ‘Basics’) will need urgent discussion with a neurosurgeon, to determine if the child needs emergency surgery. If not, a clear management plan – who will monitor the child, and where? will need to be agreed.
  • This child may require intubation, for airway, oxygenation & ventilatory control, or for secondary transfer. Tranexamic acid may be used. Attention should be paid to pain management, and neuroprotective initiatives should be put in place (control of ICP, blood pressure, CO2, O2 & glucose – see the referenced paediatric FOAM article which provides a good summary of clinical management) 

You see a 20 month old boy in your ED, who was playing with his 6 year old cousin when he ran into an opening door at home. He cried immediately, and vomited around 10 minutes later. 

Having been brought into the ED, which is 20 minutes from his home, he has vomited twice more. There was no LOC or seizure activity, and other than looking nauseated he appears to be behaving normally.

To scan or not to scan?

What guidance do the parents/nursing staff looking after this child in the ED need?

How long will you observe for, and what if the child vomits again?

  • This child can probably be safely observed without immediate CT scanning – this management approach would be supported by NICE (see sections 1.4.9 and 1.4.10).
  • This case will hopefully provoke discussion about what constitutes a ‘vomit’, and whether there are any other plausible causes of vomiting, other than the injury itself. 
  • Learners could discuss what local provision they have for more extended observation of a child.
  •  Discussion of provision of verbal + written advice would also be pertinent.

A 9 month old child presents after rolling off a bed onto the floor. You see a 7cm swelling on his occiput. In the trauma call, he is held in mum’s arms and is crying.

You are unsure over how to proceed – the child definitely needs CT imaging, but how should we ensure they keep still?

  • Recap of CT guidance – “For children under 1 year, presence of bruise, swelling or laceration of more than 5 cm on the head”
  • Options for CT sedation: benzodiazepines vs. diamorphine/opiates vs. ketamine vs. intubation & ventilation – given a significant CT finding is possible. This would be a good opportunity to mention the 2020 revisions to RCEM ketamine sedation guidance (with associated DFTB commentary)

An 8 year old girl is brought in by her dad. She clashed heads with another player at basketball two days previously, and did not initially seek medical advice as she was ‘fine’. She had to leave school early today because she had trouble seeing the board & teacher, and felt sick. There are no focal neurological findings but there is a bruise on the parietal part of the scalp on the right, and you cannot feel the scalp.

  • Need for detailed history-taking around the delay in presentation – actively look for any safeguarding concerns
  • Should we have an altered threshold for CT imaging when presentation is delayed? This DFTB post is a useful summary of a paper relating to this cohort of patients – finding of a nonfrontal scalp haematoma or strong suspicion of a basal skull fracture were significantly associated with a clinically significant brain injury.

A 15 year old girl re-attends 10 days after being knocked unconscious for 10-15 seconds while jumping for a header playing football. She passed a pitchside concussion test and continued to play, but was substituted after saying she felt dizzy, and was seen in an ED. A CT scan was performed – which showed no bleed, contusion or fracture. 

She says she found it hard to concentrate on schoolwork for a week afterwards, but this is now normal. She wants to know exactly when she can go back to playing as she has an important match in 3 days.

What do you do?

  • Concussion describes the symptoms & abnormal function experienced by patients after a head injury, without any evidence of macroscopic brain injury. Its management is commonly misunderstood and poorly explained to patients and carers.
  • Management focuses on cognitive rest, avoidance of activities that trigger symptoms, and graduated return to cognitive activity & education.
  • Return to sport should also be graduated, with trial of light activity, and avoidance of sport with a risk of head impact until the patient has been reviewed by a clinician.
  • There is a significant risk of secondary concussion if sport/normal activity is returned to too soon after the initial injury 

Some excellent resources from:

You see a 4 year old with a head injury. All of the following are an indication for urgent CT imaging except:

A: GCS<14

B: Sign of a basal skull fracture

C: Focal neurology on examination

D: Post-traumatic seizure

E: Loss of consciousness for a few seconds

The correct answer is E.

NICE guidelines mention all of the above as indications for immediate CT except for LoC – if brief this is not an indication. If more prolonged (>5min), this would mandate observation in the ED for at least 4h after the time of injury.

In an intubated child with an extradural haematoma causing mass effect, the following are important considerations in managing intracranial pressure:

A: Managing untreated pain

B: Using RR or tidal volume to control pCO2

C: Keeping O2 saturations 94-98%

D: Keeping blood glucose tightly controlled between 4-8

E: Removing any constrictive neck devices (tube ties, cervical collars etc)

The correct answer is D.

Evidence for tight glycaemic control has been superseded by the risks associated with hypoglycaemia for the injured brain. Prevention of hyperglycaemia would be a more sensible aim. All of the other answers would minimise increases in intracranial pressure, including aiming for a low-normal pCO2.

When managing children with head injuries, which of the following statements are true?

A: If the mechanism of injury is dangerous, the cervical spine should be CT imaged along with the head.

B: It is good practice to discuss management of delayed presentations with a senior before discharge

C: 3 vomits in 10 mins constitutes separate ‘episodes’ of vomiting.

D: CT imaging is essential in those with haemophilia.

E: Intranasal diamorphine can be used to manage pain and keep a child still for scanning.

The correct answers are A, B, and E.

If the child’s head is being imaged with CT, best practice would be to extend to the cervical spine if concern exists regarding injury. Teams should use judgement of the mechanism, presence of abnormal neurology and GCS to help make this decision). 

Evidence would suggest that presentation >24h after a head injury is associated with more significant findings, thus the threshold for scanning may need to be altered. 

Clinical judgement needs to be exercised with regard to vomiting. NICE refers to a vomit being a ‘single discrete episode’ but does not explicitly define timing. Our practice would suggest significant time for recovery should be allowed between episodes eg. 20mins.

Diamorphine is a good analgesic for young children and its sedative effects can be harnessed when attempting to safely CT in mild agitation – although other options are more safe if airway protection is required.

Congenital bleeding diatheses such as haemophilia require a lower threshold for imaging, and would need urgent supplementation of clotting factors – but very minor trauma to the head may still be managed without CT.



Please download our Facilitator and Learner guides