Clinical exam, GCS, and FOUR score

1. Bedside vignettes: why this matters in the PICU#

Vignette A. The 6-year-old after the MVC#

A 6-year-old arrives in resus 40 minutes after an unrestrained MVC. ETT in place, in-line stabilisation. On your first look he opens his eyes briefly to a sternal rub (E2), the ETT precludes a verbal score so you record V1T, and he localises with the left arm but extends with the right (M5 best). GCS = 8T. You also examine FOUR: eye 2 (open to pain only), motor 4 (localising left), brainstem 3 (right pupil 5 mm sluggish, left 3 mm brisk; corneals intact), respiration 1 (breathes above the vent rate set by the ventilator). FOUR = 10. The right-sided anisocoria and the right-arm extension drive the next 60 seconds: hypertonic saline, head up, CT scanner now. 1 2 3

Vignette B. The 6-month-old with bulging fontanelle#

A 6-month-old infant is brought in floppy and lethargic after 18 hours of vomiting. The fontanelle is full but not tense. She opens her eyes to your voice (E3), cries inconsolably when handled but the cry has no consonants and is not directed at her mother (V3 on pGCS), and withdraws her arm from a heel-stick (M4). pGCS = 10. The pupils are 4 mm and brisk bilaterally. You document the exam, alert the surgical team that you may need a fontanelle ultrasound and a CT, and place a peripheral line for fluids while you decide on imaging. The verbal score in a pre-verbal child is the part that catches people out: a 6-month-old never scores V5; the highest possible V on the pGCS is 5 for "coos, babbles, smiles." 4

Vignette C. The post-arrest adolescent at 72 hours#

A 14-year-old is on day 3 after an unwitnessed out-of-hospital cardiac arrest; rewarmed, sedation off for 8 hours. He does not open his eyes to noxious stimulation (E1), is intubated (V1T), and shows extensor posturing to deep nailbed pressure bilaterally (M2). GCS = 4T. FOUR: eye 0, motor 1, brainstem 0 (both pupils 6 mm and non-reactive, no corneals, no cough on suction), respiration 0 (no spontaneous breaths above the vent). FOUR = 1. This is the case where the FOUR score does the work the GCS cannot: brainstem and respiratory function are quantified, and the floor (FOUR 0) is the only structured exam score that can predict in-hospital mortality with high specificity in comatose adults and is being validated in pediatrics. 5 6

2. What the exam is, and what it is not#

The clinical neurological exam samples behavioural responses to structured stimuli. It is not a measurement of brain activity, of perfusion, or of metabolism. It is a probe of integrated function across the level of the brainstem reticular activating system, the cortex, and the descending motor pathways.

Three things to keep in mind every time you examine.

Arousal is brainstem, content is cortex. A patient with eyes open and no command-following has an intact reticular activating system but absent or disconnected cortex. A patient with eyes closed who follows commands when you ask does not exist; if you encounter it, suspect locked-in or psychogenic unresponsiveness and look harder.

Best response, not worst response. GCS, FOUR, and pGCS all score the best response in each domain across the exam. A child who localises with one arm and extends with the other scores M5 (localises), not M2 (extends). Asymmetry is documented separately as a focal finding.

Serial exams trump single exams. The single most important variable on this page is the trend. A GCS that has fallen from 15 to 13 over an hour is an emergency, regardless of the absolute value. A GCS that has been 10 for 8 hours in a sedated post-operative child is a different clinical situation.

3. The scales: GCS, pGCS, FOUR, mGCS#

3.1 Glasgow Coma Scale#

The original 1974 Teasdale and Jennett scale. Eye, verbal, motor; minimum 3, maximum 15. 1

Operational tips.

• If the patient is intubated, score V as 1T and report the total with the "T" suffix, e.g., "GCS 8T."
• "Localises" (M5) means the limb crosses the midline or moves above the clavicle to remove the stimulus. Simple withdrawal of the limb on the same side as the stimulus is M4.
• Sternal rub is uncomfortable and bruises; supraorbital or nailbed pressure is preferred in most modern units, especially in children.
• Score the best response across the limbs; document asymmetry separately.

3.2 Pediatric GCS#

Same E and M scales; V scale adapted for the pre-verbal child. 4

Children aged 2 to 5 can be scored on either scale depending on developmental expectations; school-age children use the adult scale. The Adelaide Coma Scale uses 14 points instead of 15 (it removed the adult "confused" score) and is largely historical. 7

3.3 FOUR score#

Wijdicks 2005 four-domain alternative. Each domain 0 to 4; total 0 to 16. Designed for the intubated, sedated, or aphasic patient where the GCS verbal component is uninterpretable. 2 8

Why FOUR is the right score for the ICU. Three reasons. First, it scores brainstem reflexes (pupil, corneal, cough), which GCS does not, and the brainstem is the most prognostically relevant compartment in many comatose patients. Second, the respiratory domain distinguishes the patient who is triggering the vent from the patient who is not, which matters for sedation titration and weaning. Third, the floor (FOUR 0) maps to absent brainstem function, the bedside surrogate for the imminent need for ancillary brain-death testing. 3 6

3.4 The motor component alone (mGCS)#

In many trauma triage scores (the Simplified Motor Score, the Reduced Motor Score), the motor component of GCS is used alone. The reasoning: it is the most prognostically informative single component and the most reliable across raters. It is not a replacement for the full GCS in the ICU, but it is a useful single-number proxy when full scoring is impractical. 4

4. The pupillary exam#

Pupils are the most-watched data point on the bedside chart, and the most-misread.

The first-look sequence. Stand at the foot of the bed. Ambient light. Estimate sizes to the nearest 0.5 mm. Then darken the room briefly (or shield with your hand) and time the dilation in seconds. Then bright-light each side directly with a pen-torch, note the brisk constriction, and check the consensual response on the opposite side.

What anisocoria really means. True new-onset anisocoria > 1 mm in a sleepy patient with a head injury is structural until proven otherwise: it is the bedside signature of uncal herniation against the third nerve and requires immediate escalation. Pre-existing anisocoria (Adie's pupil, congenital, post-surgical iris) is benign but should be documented on admission. Pharmacological causes (atropine, scopolamine patch, nebulised ipratropium aerosol drifting into one eye) are common and reversible.

See the dedicated pupillometry page for NPi, constriction velocity, and the ORANGE multicentre data.

5. Brainstem reflex ladder#

In order from rostral to caudal. Each reflex tests a specific brainstem level.

Order of loss in rostral-caudal deterioration. Pupillary first (midbrain), then corneal (upper pons), then oculovestibular (lower pons), then gag and cough (medulla). The loss of the cough reflex on tracheal suction in a patient who had it yesterday is a late but ominous sign.

6. The numbers: what to record, in what order#

For every exam, document this six-pack:

Record both sides separately. Time-stamp the exam. Note whether sedation was on or off. A single number in isolation is uninterpretable.

7. What is normal? Age-banded expectations#

A pre-verbal infant cannot score above pGCS 12 by definition in the first months of life because the verbal repertoire ceilings the score. This is not abnormal; it is the scale, not the child. Use the trend within the same patient and within the same scale.

8. What is abnormal? Pattern library#

Decision tree: "what does the exam tell me right now?"#

flowchart TD
  Exam[Exam complete] --> Sum{GCS total}
  Sum -->|13 to 15| Mild[Mild]
  Sum -->|9 to 12| Mod[Moderate]
  Sum -->|3 to 8| Sev[Severe]
  Sev --> Pups{Pupils symmetric?}
  Pups -->|No| Hern[Herniation suspected]
  Pups -->|Yes| BS{Brainstem reflexes intact?}
  BS -->|No| Deep[Deep coma, FOUR low]
  BS -->|Yes| Cort[Cortical or subcortical insult]
  Hern --> Esc[Escalate: HTS, head up, imaging, NSG]
  Deep --> Ancillary[Consider ancillary testing if criteria met]
  Mod --> Trend{Trend over last hour?}
  Trend -->|Falling| Fall[Re-examine in 15 min; imaging if continued fall]
  Trend -->|Stable| Watch[Routine reassessment]

9. Try it: interactive widgets#

10. Using the exam in management decisions#

The exam does not directly titrate BP, CPP, or sedation; it titrates decisions about all three.

10.1 Sedation depth#

A patient on a propofol or midazolam infusion needs a planned daily sedation interruption (a "sedation hold") so the clinical exam can be assessed at trough. Without this, the GCS / FOUR is the score of the drug, not the brain. Some units use BIS or pSEDline as a real-time proxy when sedation cannot safely be interrupted (refractory raised ICP, refractory status); see the BIS page for the limits of these proxies.

10.2 ICP and CPP#

A falling GCS in a TBI patient with an ICP monitor in place forces a recheck of ICP, CPP, sedation, head position, PaCO2, temperature, and serum sodium. In a patient without a monitor, a falling GCS is the strongest single trigger for placing one, especially when GCS falls to 8 or below.

10.3 Imaging escalation#

A drop in GCS of 2 or more points, new asymmetry on motor exam, or new anisocoria in a child with TBI all justify an urgent re-image (CT in the acute window; MRI for vascular or evolving lesions later).

10.4 Prognostication after cardiac arrest#

The FOUR score, paired with somatosensory evoked potentials and quantitative pupillometry, is one leg of the multimodal prognostication framework after pediatric cardiac arrest. A FOUR of 0 at 72 hours, bilateral absent SSEP N20s, and an NPi of 0 together are very specific for poor outcome. Single-modality prognostication is forbidden by current consensus statements. 11 12

11. Clinical contexts: the exam in eight scenarios#

11.1 Severe TBI#

The GCS is the defining variable of severe TBI (GCS ≤ 8). It triggers airway protection, ICP monitor placement consideration, head-injury imaging protocols, and the entire BTF / pediatric BTF management bundle. A serial fall in GCS in a patient with an ICP monitor demands a recheck of the four pillars: head position, sedation depth, PaCO2 / oxygenation, and osmotherapy timing. In children, the pediatric BTF guidelines tie management thresholds to GCS for triage decisions. 13 14

11.2 SAH#

Pre-treatment grading (Hunt and Hess, World Federation of Neurological Surgeons) is GCS-based. Post-treatment, the exam is the primary DCI screening tool: new focal deficit, new confusion, or a drop in GCS by 2 or more points triggers angiography and DCI workup. TCD and qEEG add resolution but the exam is the gate. 15 16

11.3 Pediatric AIS#

Sudden focal weakness, dysphasia, or facial droop in a child are the bedside signs of an arterial ischaemic stroke. The pediatric NIH stroke scale (PedNIHSS) adds quantitative scoring beyond the GCS for hyperacute decisions about IV thrombolysis or thrombectomy. The full neurological exam guides candidacy decisions; serial post-recanalisation exams catch hyperperfusion bleeds. 17 18

11.4 HIE and post-cardiac arrest#

Neonatal HIE: the Sarnat staging system encodes consciousness level (alert / lethargic / stuporous), tone, primitive reflexes, autonomic function, and seizure activity. Sarnat stage 2 and 3 are the standard inclusion criteria for therapeutic hypothermia. 19 Pediatric post-arrest: serial GCS and FOUR scores at 24, 48, 72 hours post-rewarming, in combination with EEG and biomarker data, drive the prognostication conversation. 20

11.5 Pediatric ECMO#

Daily neurological exam on ECMO is challenging (paralysis, sedation, hard-to-examine eyes through the cannula traffic), but it remains the bedside trigger for the next investigation. New asymmetry, new anisocoria, or new seizure activity prompts head CT or fontanelle ultrasound. The ELSO neurological guidelines mandate documentation of pupil size, reactivity, and motor responsiveness at least every 6 hours on VA-ECMO. 21 22

11.6 Meningitis and encephalitis#

The exam is the primary marker of evolving cerebral oedema, hydrocephalus, or vasculitic stroke in bacterial meningitis and encephalitis. New focal deficit, new seizure, declining GCS by 2 points, or new anisocoria all warrant urgent imaging and consideration of EVD or ICP monitoring. The IDSA encephalitis and meningitis guidelines lean heavily on the exam for serial monitoring. 23 24

11.7 Brain death determination#

The clinical exam is the brain-death determination. The world brain-death project lays out the prerequisites (irreversible cause, normothermia, normonatraemia, no confounders), the exam (deep coma, absent brainstem reflexes, apnoea testing), and the optional ancillary tests. Pediatric brain-death criteria differ slightly by age: the AAP / SCCM / Child Neurology Society guidance requires two clinical exams separated by 12 to 24 hours depending on age, with apnoea testing at each. 25 2

11.8 DKA cerebral oedema#

In a child being rehydrated for DKA, a drop in GCS by 2 points, new headache out of proportion, vomiting, bradycardia, or hypertension at any point between hour 4 and hour 24 of treatment is cerebral oedema until proven otherwise. The exam, not imaging, drives the bolus of mannitol or hypertonic saline. 26 27

12. Multimodal integration: the exam in the MNM stack#

13. Setup and technique#

13.1 Equipment#

• Pen-torch with adjustable focus (or two pen-torches for direct vs consensual comparison).
• Tongue depressor.
• Cotton-wisp or sterile saline ampoule.
• A quiet bedside (turn off the TV; minimise distractions).
• Pupillometer if available (separate workflow; see pupillometry page).

13.2 The structured 90-second exam#

1. Approach the patient from the foot of the bed. Note baseline posture, spontaneous movements, eye opening, breathing pattern at rest.
2. Greet the patient by name (if known). Note response.
3. Command-following. "Squeeze my hand." "Show me two fingers." "Stick out your tongue." Test both sides separately to detect asymmetry.
4. Verbal content (if applicable). Ask name, location, why-here. Score V appropriately.
5. Pupils. Ambient light estimate; then bright direct light each side; check consensual.
6. Brainstem reflexes. Corneal each side (if cooperative); cough on suction (if intubated); gag (cautiously, in deeply comatose); doll's eyes only if cervical spine cleared.
7. Motor. If no command-following, supraorbital pressure or nailbed pressure each limb. Record best response per limb and best overall.
8. Score and document. GCS components, FOUR components, pupil sizes and reactivity, focal findings, time.
9. Compare with previous exam. Note delta and trend.

13.3 Pediatric adaptations#

• Pre-verbal infants. Use pGCS verbal scale; observe interaction with caregiver, response to caregiver's voice, quality of cry (consolable vs persistent).
• Toddlers (1 to 3 y). Combine pGCS or adult GCS depending on developmental level; ask the parent to elicit command-following ("Show mummy the doggy"); use familiar objects rather than abstract commands.
• School-age and older. Standard adult GCS applies.
• Pain assessment: use FLACC (Face, Legs, Activity, Cry, Consolability) for pre-verbal children rather than visual analogue scales; depressed FLACC despite stimulation suggests cortical depression.

13.4 Documenting the trend#

Bedside flow sheets should have GCS components and pupil findings in adjacent columns with consistent timing (often Q1 hour in TBI, Q4 hour in stable post-op neuro). The pattern over 24 hours is more informative than any single point. Many EHRs allow graphing GCS over time; this is the single most useful summary view for the morning round.

13.5 Inter-rater reliability#

Reith et al's 2016 systematic review reports moderate inter-rater agreement for GCS sum (weighted kappa ~0.7), better for motor component alone (kappa ~0.8), and worse for verbal in intubated patients (kappa ~0.5). Structured bedside training improves these by ~10%. The clinical implication: in a multi-shift unit, document components, train new staff, and use the trend within a single examiner where possible. 4 7

13.6 When the exam is not possible#

• Deep sedation that cannot be safely held: use FOUR brainstem domain only; use NPi for pupil; use BIS / pSEDline for sedation depth proxy; document "exam confounded by sedation."
• Therapeutic hypothermia: exam suppressed by ~3 to 4 points; do not prognosticate.
• Paralysis: only brainstem (pupil, corneal, cough) and respiratory triggering are assessable; use cautiously.
• Massive facial trauma: eye and verbal components unscoreable; use motor alone and brainstem reflexes.

14. Pitfalls#

• GCS sum without components. Always document E, V, M separately. The sum hides important detail.
• Scoring the worst limb. Best response per domain is the scoring rule. Asymmetry is a focal finding, not a low score.
• Scoring during paralysis or deep sedation. Document the confound rather than the score, or score after a planned sedation hold.
• Missing the pGCS verbal scale. A 6-month-old never scores adult V5; using the adult scale guarantees an artificially low number.
• Cold caloric in perforated TM or oculocephalic in unstable C-spine. Always check before testing.
• Pharmacological pupils. Atropine, scopolamine, nebulised ipratropium can mimic CN III compression. Take a thorough drug history.
• Anchoring on a single exam. Trend matters more than any single number.
• Confounded prognostication after arrest. Never use the exam alone; pair with EEG, SSEP, imaging, NPi, and time.
• Forgetting that the exam is dynamic. Sedation comes and goes, fevers spike, ICP fluctuates; document the conditions of every exam (sedation status, temperature, time relative to last dose).
• Ignoring the family's report. Parents often notice subtle behavioural changes before the structured exam picks them up; ask, and listen.

15. Combine with…#

• Pupillometry: NPi as the quantitative complement to the bedside pupillary exam.
• ICP: the structural correlate of a falling exam.
• EEG: catches NCSE in a depressed exam.
• BIS: sedation depth proxy when sedation cannot be held.
• TCD: bedside vasospasm and CPP triage when the exam suggests evolving DCI.
• Foundations: autoregulation: the physiology behind why a falling exam often means falling CPP.

16. Evidence summary#

17. Recent literature (2022 to 2025)#

• Alkhachroum 2024: review of GCS limitations in the modern ICU, with proposed adjuncts (NPi, FOUR, automated coma scoring). 7
• Tasker 2023 PCCM review: pediatric severe TBI management, with GCS as the entry criterion for ICP monitoring. 14
• Naim 2023: pediatric post-cardiac-arrest brain injury and prognostication framework. 12
• Figaji 2025 pediatric MMM consensus: positions the exam as the floor of any multimodal stack. 28
• Glaser 2024 DKA review: re-emphasises the exam as the primary trigger for cerebral oedema treatment in DKA. 27

18. Self-check#