CT head and C-spine, decided properly.

The CT head in minor head injury is a test for a treatable structural lesion — an extradural or subdural haematoma, a contusion large enough to matter, a depressed fracture. It is not a test for concussion, and a normal scan does not mean the brain was not injured. Most people with a normal CT and persistent symptoms have exactly what it looks like: a concussion, which CT cannot see.

NICE is explicit that CT is the test of choice for this question, and equally explicit that MRI is not the primary investigation and skull X-rays are not a way of deciding whether a brain injury is present. NG232 1.5.1, 1.5.2, 1.5.4

Keep the outcome in view when you read the numbers in the Evidence tab. The rules were built to be near-perfect at finding patients who need neurosurgery, and only moderately good at finding every trace of blood. Those are different targets, and a rule tuned for one will look poor when judged against the other.

The 1-hour list NG232 1.5.8 is a list of features that suggest a lesion big enough, or a brain sick enough, that the answer changes management now: GCS 12 or less at the door; GCS below 15 two hours after injury; suspected open or depressed skull fracture; any sign of basal skull fracture; a post-traumatic seizure; a focal neurological deficit; more than one episode of vomiting.

The 8-hour list NG232 1.5.9 applies only to people who lost consciousness or have amnesia, and then asks for one of: age 65 or over, any current bleeding or clotting disorder, a dangerous mechanism, or more than 30 minutes of retrograde amnesia. These raise the probability of finding something without implying it is expanding while you talk.

The hour is a surgical window

An extradural haematoma expands; a patient with a lucid interval declares themselves by deteriorating, and the interval is not reliably long. NICE reinforces this on the other side of the scan: a provisional written radiology report must be available within an hour of the scan NG232 1.5.14, and anyone with new surgically significant abnormalities is discussed with a neurosurgeon. NG232 1.4.15 An hour-target CT that sits unreported for three hours has not achieved anything.

If your department cannot deliver the scan in the indicated window — no scanner, or a patient too agitated to lie still — that is itself an admission criterion, not a reason to downgrade the decision. NG232 1.9.1

Two of the 1-hour criteria are GCS criteria, and they are deliberately different from each other. One is GCS 12 or less on initial assessment in the emergency department. The other is GCS less than 15 at 2 hours after the injury. NG232 1.5.8 The second is not a restatement of the first; it exists because a patient can arrive at 15 and be 14 an hour later, and that trajectory is the finding.

This is why NICE asks for a trained assessment within 15 minutes of arrival NG232 1.4.7, immediate assessment for anyone with a GCS below 15 NG232 1.4.4, and re-examination by an ED clinician even for people triaged as low risk. NG232 1.4.9 The pathway assumes somebody looks twice.

Two traps in reading the number

The baseline. In dementia, chronic neurological disease or learning disability the pre-injury GCS may not be 15, and NICE asks you to establish it and take it into account. NG232 1.3.5 It works in both directions: a GCS of 14 that was 14 last week is not a deterioration, and a GCS of 15 in someone whose baseline is 15 tells you nothing has changed yet.

Alcohol and drugs. The most quotable sentence in the guideline is also the most ignored: “Only assume a depressed conscious level is due to intoxication after an important traumatic brain injury has been excluded.” NG232 1.4.2 Intoxication is a reason the assessment is unreliable, never an explanation for it.

Vomiting. For adults the threshold is more than 1 episode, so a single vomit is not a criterion and a second one converts the patient to a 1-hour scan. For under-16s the equivalent is 3 or more discrete episodes, and it is one of the second-tier factors, not an automatic scan. NG232 1.5.8, 1.5.11 Two different numbers for two different populations is not an inconsistency; using the adult number on a child is.

Dangerous mechanism is defined, and the definition is narrower than the phrase feels. For the adult CT head criteria it is a pedestrian or cyclist struck by a motor vehicle, an occupant ejected from a motor vehicle, or a fall from more than 1 m or 5 stairs. NG232 1.5.9 The C-spine list is longer and includes axial load, rollover, high-speed collision, motorised recreational vehicles and bicycle collisions. NG232 1.6.2 Five stairs is a specific number; “fell down the stairs” is not the same statement.

Retrograde versus anterograde. The 30-minute criterion is retrograde amnesia — loss of memory for events before the impact — and is easy to mistake for the far commoner post-traumatic (anterograde) gap. Ask specifically what the last thing they remember before the fall was, and time it.

Recommendation 1.5.13 covers people with no other indication for a CT head who are on an anticoagulant — vitamin K antagonists, DOACs, heparins — or an antiplatelet excluding aspirin monotherapy. For them, NICE says consider a CT within 8 hours, or within the hour if they present beyond 8 hours. NG232 1.5.13

Two features of that wording carry the clinical weight. “Consider” in NICE language is a weaker recommendation than “do”, and it invites a conversation rather than a reflex. And aspirin alone is outside the recommendation altogether — a patient on aspirin with no other criterion has no NG232 indication for a scan, which surprises people who have absorbed “blood thinner equals scan”.

NICE gives two worked reasons for scanning within that window: it is hard to do a reliable risk assessment, or the person might not come back if they deteriorate. Both are social and practical as much as physiological, and both are much more often true of the frail, isolated older patient than the well-supported one.

The second question — the repeat scan. NG232 does not ask for routine repeat imaging or prolonged observation after a normal CT in an anticoagulated patient, and local practice varies widely. In a retrospective single-centre cohort of 596 anticoagulated adults with mild injury and a negative first CT who all had a second scan at 24 hours, delayed intracranial haemorrhage was found in 2% (12 patients), and complications of hospitalisation — mostly infection and delirium — occurred in 66% of those admitted with it. Giamello 2025 Retrospective, single centre Admission is not a neutral act in an 83-year-old.

If the scan does show haemorrhage, reversal is a separate decision with its own guidance; NG232 points to NICE’s blood transfusion guideline for vitamin K antagonists and to MHRA advice for DOACs. NG232 1.5.13 The bleeding and reversal module runs that pathway.

The adult C-spine pathway asks, in this sequence: are there high-risk factors? If not, is there any low-risk feature that makes it safe to assess movement? Only then: can the patient actively rotate 45° each way?

The high-risk factors are GCS 12 or less, intubation, an urgent need for definitive diagnosis, blunt polytrauma involving the head and chest, abdomen or pelvis in someone alert and stable, and — where there is clinical suspicion of C-spine injury — age 65 or over, a dangerous mechanism, focal peripheral neurological deficit, or paraesthesia in the limbs. Any of these means CT within an hour. NG232 1.6.2

In a patient with neck pain or tenderness and no high-risk factor, three things each mean CT within an hour: movement cannot be safely assessed, safe assessment shows they cannot rotate 45° each way, or they have a condition predisposing to cervical spine injury such as axial spondyloarthritis. NG232 1.6.3 The ankylosed spine deserves its own reflex: it fractures through the fused column after apparently trivial force, the fracture is often unstable, and plain films are poor at showing it.

Delayed presentation. Both C-spine recommendations open by explicitly including people who present late. NG232 1.6.2, 1.6.4 The patient who has had neck pain since a collision three days ago is inside the criteria, not outside them.

Clearing the neck. After normal imaging, NICE considers the risk low enough for transfer to the community if the GCS is 15, examination is normal, there is no other reason to admit, and there is competent supervision at home. NG232 1.10.3 Note how much of that is about the person’s circumstances rather than their scan.

The paediatric CT head criteria have a structure adults do not have. Some factors mean a scan within the hour on their own — suspected non-accidental injury, post-traumatic seizure, GCS below 14 at initial assessment (below 15 for babies under 1 year), GCS below 15 at 2 hours, suspected open or depressed fracture or tense fontanelle, any sign of basal skull fracture, focal neurological deficit, and in babies under 1 year a bruise, swelling or laceration over 5 cm on the head. NG232 1.5.10

Six others — witnessed loss of consciousness lasting more than 5 minutes, abnormal drowsiness, 3 or more discrete vomits, a dangerous mechanism (for the paediatric head criteria a different list: a high-speed road traffic accident as pedestrian, cyclist or vehicle occupant, a fall from more than 3 m, or a high-speed injury from a projectile or object — not the adult 1 m or 5 stairs), amnesia over 5 minutes, or a bleeding or clotting disorder — only mandate a scan when more than one is present. NG232 1.5.11

The neck is the bigger divergence. For under-16s, CT is reserved for a defined list — GCS 12 or less, intubation, focal peripheral neurological signs, limb paraesthesia, an urgent need for definitive diagnosis, other body areas being scanned with clinical suspicion of C-spine injury, strong suspicion despite normal X-rays, technically inadequate X-rays, or X-rays showing significant bony injury. NG232 1.6.4 Otherwise, for a child with neck pain or tenderness and no CT indication, the pathway is 3-view X-rays first — for a dangerous mechanism (here the 1 m or 5 stairs list, as for adult necks), when safe assessment of movement is not possible, or when the child has a condition predisposing to cervical spine injury such as collagen vascular disease, osteogenesis imperfecta or axial spondyloarthritis; and also when safe assessment shows the child cannot rotate 45° each way. NG232 1.6.5, 1.6.6 Reaching for CT first in a child is a real deviation from the guideline, not a cautious version of it.

Safeguarding is not an add-on. Suspicion of non-accidental injury is a 1-hour CT criterion, and NICE separately asks clinicians to consider abuse or neglect as a cause of any head injury and to involve someone trained in safeguarding. NG232 1.4.12, 1.4.13 An injury that does not match the history, or a head injury in a non-mobile baby, is a safeguarding presentation that happens to involve a scanner.

NG232 gives explicit admission criteria: new clinically important abnormalities on imaging; a GCS that has not returned to 15 or to the person’s baseline regardless of the scan; an indicated CT that cannot be done in time; continuing worrying symptoms such as persistent vomiting, severe headache or seizures; and other sources of concern including suspicion of ongoing post-traumatic amnesia, intoxication, other injuries, shock, suspected non-accidental injury, meningism or CSF leak. NG232 1.9.1

It also defines what deterioration looks like in numbers: agitation or abnormal behaviour; a sustained drop of 1 point in GCS for at least 30 minutes, weighted towards the motor score; a drop of 3 or more in eye or verbal, or 2 or more in motor; severe or increasing headache or persistent vomiting; new or evolving signs such as unequal pupils or asymmetry of movement. NG232 1.9.13 A confirmed change triggers reappraisal by a supervising doctor and consideration of an immediate scan. NG232 1.9.15 And if a normal-CT patient is still not GCS 15 at 24 hours, consider further imaging. NG232 1.9.16

Discharge has a hard floor. Do not discharge until the GCS is 15 — or back to the documented pre-injury baseline — and, in preverbal children, until consciousness is normal on the paediatric GCS. NG232 1.10.4 Everyone discharged gets verbal and printed advice NG232 1.10.8; the estate’s head injury discharge leaflet exists for that.

1. Age 78. Age 65 or over is an 8-hour CT head risk factor when there has been LOC or amnesia, and a C-spine high-risk factor when there is clinical suspicion. NG232 1.5.9, 1.6.2
2. A fall of about 2 metres down the stairs. NG232 defines a dangerous mechanism as a fall from a height of more than 1 m or 5 stairs, so this qualifies on height — note that a fall down exactly five stairs would not meet the stairs half of that wording. NG232 1.5.9, 1.6.2
3. She cannot remember the fall. Amnesia opens the 8-hour pathway — and, read through the guideline’s own definition of focal neurological deficit, arguably the 1-hour one. Document which reading you used. NG232 1.5.8, 1.5.9
4. One vomit. Not yet a criterion in an adult. A second one makes it a 1-hour scan, so this is a patient to re-examine rather than park. NG232 1.5.8
5. Apixaban. Even with no other indication, 1.5.13 applies — but here there already are other indications, so the scan is not a “consider”. NG232 1.5.13

Neck. She has neck pain, is 65 or over and had a dangerous mechanism, so if there is clinical suspicion of C-spine injury she meets a high-risk factor: CT cervical spine within 1 hour. The fact that she walked in and rotates her neck does not enter the decision, because rotation may only be assessed when there are no high-risk factors. NG232 1.6.1, 1.6.2

Head. LOC or amnesia plus age 65 or over plus a dangerous mechanism: CT head within 8 hours — brought forward if she vomits again, if her GCS is below 15 at 2 hours, or if you take the focal-deficit reading of her amnesia. NG232 1.5.8, 1.5.9

And the GCS of 15 in the triage note is true, unhelpful, and the reason the pathway needs a second look at 2 hours.

1. The outcome defines the rule. “Need for neurological intervention”, “clinically important brain injury”, “any intracranial traumatic finding” and “clinically significant head injury” are four different targets used across these papers. A rule is only sensitive for its own outcome.
2. Derivation flatters. Recursive partitioning finds the combination that fits the derivation set. The Canadian rule’s 100% sensitivity for neurosurgical intervention held in a Dutch cohort, but for any neurocranial traumatic finding it was 83–87%. Smits 2005
3. Sensitivity without specificity is not a virtue. NEXUS avoided imaging in 12.6% of patients; a rule that scans nearly everyone is sensitive by construction. Hoffman 2000
4. Interobserver agreement is part of accuracy. Kappa 0.85 versus 0.47 for two rules on the same patients means one of them is being applied consistently and the other is not. Stiell 2005
5. Applicability is a hidden denominator. CHALICE applied to 99% of a paediatric cohort and CATCH to 25%. A more accurate rule you cannot apply is worth less at the bedside. Babl 2017
6. Case mix travels badly. These cohorts are Canadian, American, Dutch, Australasian and British, with different thresholds for attending and scanning. A UK guideline validated abroad is being tested in a system that does not select patients the same way.
7. None of this measures patient outcome. Every study here is diagnostic accuracy or observational risk. No randomised trial in these sources shows that using any rule changes death or disability — only how many scans are done and how many findings are caught.

A single-page implementation of the NICE NG232 imaging criteria for head and cervical spine injury, with the reasoning and the underlying evidence alongside it. The Decision tool tab runs the criteria; the Teaching and Evidence tabs explain and appraise them.

What it does not do. It does not decide whether someone has a head injury, manage raised intracranial pressure, cover penetrating head injury beyond the imaging criteria, give drug doses, replace a discussion with your neurosurgical centre, or reproduce NG232 in full — only the recommendations that bear on imaging, observation and discharge. It is not a substitute for reading the guideline.

NG232 recommendations were taken from the full recommendations text on the NICE website and quoted directly; every recommendation number on these pages points at the numbered recommendation it quotes. For every research study cited, the figures on this page come from the published abstract, retrieved from PubMed — not from full texts, review articles or secondary summaries. Where a study is described as retrospective, single-centre, or from a particular country, that description is taken from the abstract itself.

NG232 does not name the Canadian CT Head Rule, NEXUS, the Canadian C-Spine Rule, PECARN, CHALICE or CATCH in its recommendations. They appear here as the evidence background, tagged as decision rules rather than as guidance, because in UK practice NG232 is what applies.

1. National Institute for Health and Care Excellence. Head injury: assessment and early management. NICE guideline NG232, 2023. nice.org.uk/guidance/ng232
2. Stiell IG, Wells GA, Vandemheen K, Clement C, et al. The Canadian CT Head Rule for patients with minor head injury. Lancet 2001;357(9266):1391–6. PMID 11356436.
3. Haydel MJ, Preston CA, Mills TJ, Luber S, et al. Indications for computed tomography in patients with minor head injury. N Engl J Med 2000;343(2):100–5. PMID 10891517.
4. Stiell IG, Clement CM, Rowe BH, Schull MJ, et al. Comparison of the Canadian CT Head Rule and the New Orleans Criteria in patients with minor head injury. JAMA 2005;294(12):1511–8. PMID 16189364.
5. Smits M, Dippel DW, de Haan GG, Dekker HM, et al. External validation of the Canadian CT Head Rule and the New Orleans Criteria for CT scanning in patients with minor head injury. JAMA 2005;294(12):1519–25. PMID 16189365.
6. Foks KA, van den Brand CL, Lingsma HF, van der Naalt J, et al. External validation of computed tomography decision rules for minor head injury: prospective, multicentre cohort study in the Netherlands. BMJ 2018;362:k3527. PMID 30143521.
7. Stiell IG, Wells GA, Vandemheen KL, Clement CM, et al. The Canadian C-spine rule for radiography in alert and stable trauma patients. JAMA 2001;286(15):1841–8. PMID 11597285.
8. Hoffman JR, Mower WR, Wolfson AB, Todd KH, Zucker MI. Validity of a set of clinical criteria to rule out injury to the cervical spine in patients with blunt trauma. National Emergency X-Radiography Utilization Study Group (NEXUS). N Engl J Med 2000;343(2):94–9. PMID 10891516.
9. Stiell IG, Clement CM, McKnight RD, Brison R, et al. The Canadian C-spine rule versus the NEXUS low-risk criteria in patients with trauma. N Engl J Med 2003;349(26):2510–8. PMID 14695411.
10. Kuppermann N, Holmes JF, Dayan PS, Hoyle JD Jr, et al. Identification of children at very low risk of clinically-important brain injuries after head trauma: a prospective cohort study. Lancet 2009;374(9696):1160–70. PMID 19758692.
11. Dunning J, Daly JP, Lomas JP, Lecky F, et al. Derivation of the children’s head injury algorithm for the prediction of important clinical events decision rule for head injury in children. Arch Dis Child 2006;91(11):885–91. PMID 17056862.
12. Babl FE, Borland ML, Phillips N, Kochar A, et al. Accuracy of PECARN, CATCH, and CHALICE head injury decision rules in children: a prospective cohort study. Lancet 2017;389(10087):2393–2402. PMID 28410792.
13. Mason SM, Evans R, Kuczawski M. Understanding the management of patients with head injury taking warfarin: who should we scan and when? Lessons from the AHEAD study. Emerg Med J 2019;36(1):47–51. PMID 30065073.
14. Giamello JD, Martini G, Fulcheri C, Visconti GL, et al. Impact of anticoagulant therapy on delayed intracranial haemorrhage after traumatic brain injury: a study on the role of repeat CT scans and extended observation. Injury 2025;56(9):112523. PMID 40537351.

• Acute extradural haematoma on CT — Jpogi, via Wikimedia Commons, CC BY-SA 3.0.
• Fracture of the base of the dens on CT — James Heilman, MD, via Wikimedia Commons, CC BY-SA 3.0.

Both images are of unnamed patients and are used illustratively; neither depicts a patient of this site’s author.