t h e s u r g e o n 1 0 ( 2 0 1 2 ) 1 1 4 e1 1 9

Available online at www.sciencedirect.com

The Surgeon, Journal of the Royal Colleges

of Surgeons of Edinburgh and Ireland
www.thesurgeon.net

Review

A practical review of the Glasgow Coma Scale and Score

Philip Barlow*
Institute of Neurological Sciences, Southern General Hospital, 1345 Govan Road, Glasgow G51 4TF, UK

article info abstract

Article history: Since the Glasgow Coma Scale was introduced in 1974,1 it has become the most common
Received 28 November 2011 method of describing a patient’s level of consciousness. However, despite its almost
Accepted 25 December 2011 universal use, there remain a number of misunderstandings, particularly regarding the
Available online 31 January 2012 appropriate situations in which to use the Glasgow Coma Score rather than the Scale, and
also in the correct way to elicit and record the motor responses. This article, aimed at non-
Keywords: neurosurgeons, addresses these problems, and provides a reference for those wishing to
Coma Scale learn or teach the Glasgow Coma Scale and Score.
Motor response ª 2012 Royal College of Surgeons of Edinburgh (Scottish charity number SC005317) and
Royal College of Surgeons in Ireland. Published by Elsevier Ltd. All rights reserved.

Introduction authors have sought to modify the way the responses,

The Glasgow Coma (GC) Scale was introduced in 1974 as
a measure of a patient’s level of consciousness.1 Prior to this,
terms such as stuporose, comatose, semi-comatose, obtun-
ded, and decerebrate were in common use, but because such
terms were ill defined, comparison of different series of head
injuries was difficult, and communication about individual
patients was often imprecise.
The GC Scale (Table 1) provided a relatively simple and
reliable measure of conscious level. Studies showed good (i.e.
low) inter-observer variability amongst medical and nursing
staff once they had been trained.2e4 The GC Scale went on to
be accepted and used by most Neurosurgical Units in the
English speaking world, and most of Europe.5
However, its use has not been without its problems and
critics. For example, Wiese criticised the 1976 modification of
the motor response6 without “an explanation and...clar-
ifying change of name”.7 Bassi8 and Marion9 suggested the
inter-observer variability was poor (although in their studies
the GC Score was assessed (see below) not the GC Scale). Other
particularly the motor responses, are elicited and described,
presumably due to dissatisfaction with the original descrip-
tions. Also, a number of other Coma Scores have been devised,
including the GlasgoweLiege Scale, the Full Outline of UnRe-
sponsiveness (FOUR) Score, the Reaction Level Scale, the
Innsbruck Coma Scale, and the Maryland Coma Scale, all of
which have been summarised by Laureys.10 However, none
appear to have gained widespread acceptance outside their
area of origin. The author’s experience of teaching the GC
Scale over a 30-year period suggests that many of the
perceived problems with the GC Scale and Score arise from
a poor understanding of how the Scale, particularly the Motor
Responses, should be applied. Once such misunderstandings
have been overcome, most clinicians find that the GC Scale is
accurate and reliable, and just as quick as other methods of
assessing responsiveness.
Other reviews of the GC Scale have been published,
including the comprehensive review by Matis and Birbilis.5
Matis describes the origins of the GC Scale, its use in the
assessment of reduced consciousness from all causes, not just

* Tel.: þ44 141 201 2026, þ44 7968 802 443 (mobile).
E-mail address: philbar@doctors.org.uk.
1479-666X/$ e see front matter ª 2012 Royal College of Surgeons of Edinburgh (Scottish charity number SC005317) and Royal College of
Surgeons in Ireland. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.surge.2011.12.003
 t h e s u r g e o n 1 0 ( 2 0 1 2 ) 1 1 4 e1 1 9 115

Table 1 e The Glasgow Coma Scale and Score. Taken from Eye opening
SIGN Guideline No. 46.15
Feature Scale responses Score There is rarely any difficulty in assessing Eye Opening. If the
notation patient does not have spontaneous eye opening on approach,
Eye opening Spontaneous 4 speaking in a loud voice whilst gently shaking him will
To speech 3 determine if they eye open to speech. (If they do, then obvi-
To pain 2 ously the verbal response, and whether they obey commands,
None 1 will be assessed at the same time.)
Verbal response Orientated 5
If there is no eye opening to speech, pain is applied, either
Confused conversation 4
Words (inappropriate) 3
by squeezing the ear lobe or by pinching the upper inner fold
Sounds (incomprehensible) 2 of the trapezius muscle, or if these fail, by pressure to the
None 1 finger tip (see below e testing motor response). Applying
Best motor response Obey commands 6 pressure to the supraorbital ridge is not the best way to test for
Localise pain 5 eye opening, as this can cause a reflex screwing up of the
Flexion Normal 4
eyelids. In the case records the responses are written as:
Abnormal 3
Extend 2
None 1 EO Spont;
Total Coma ‘Score’ 3/15e15/15 EO Speech;
EO Pain;
EO None.
trauma, the problems inherent in using the GC Scale in
ventilated patients, and the statistical implications of the If it is not possible to test for eye opening because of, for
different weightings applied to each of the three components; example, swelling of the eyelids, this is recorded on the
he also describes the Paediatric Coma Scale. observation chart as “C”, not “None” (Fig. 1), and in the case
Many descriptions of the GC Scale will also be found in record the full reason is given.
textbooks, guidelines, and on internet sites. However, few of
these describe accurately and in detail how each of the
responses should be elicited and recorded; virtually none Verbal response
explain the importance of distinguishing between the GC
Score and the GC Scale in clinical use. If the patient appears drowsy or unresponsive, some effort may
This article aims to provide a detailed and practical be required to get him to answer questions. The best verbal
account of the GC Scale and Score for a general (non-neuro- response, “Orientated”, requires the patient to be orientated to
surgical) audience, and to act as a reference for those wishing time, place and person, and so this requires the patient to
to learn or teach the GC Scale. answer at least three questions accurately. A “Confused”
response requires the patient to be able to speak in phrases or
sentences, but to be disorientated in at least one of time, place or
The GC Scale person. In milder injuries there may be some advantage in
grading the degree of orientation e i.e.: time, place, person (x3),
The three aspects of consciousness are reckoned to be place and person (x2), person (x1). This can be written long-
arousal, awareness, and activity, and the GC Scale uses hand in the case records, but anything less than “Orientated
a description in words to describe these components as Eye x3” would be recorded on the coma chart as “Confused”.
Opening, Verbal Response, and Motor Response respectively. The next worse response after “Confusion” is “Inappro-
The GC Score (see below) is the numerical equivalent of the GC priate Words”. Almost always these will be swear words or
Scale. It is important to understand that the GC Scale other expletives, often in response to stimulation such as pain.
(description in words) should be used to describe individual If no recognisable words are uttered, but the patient cries,
patients in a clinical situation, whereas the GC Score (a groans, screams, or makes any other kind of noise, this is
number) was designed for research and audit purposes. recorded as “Incomprehensible Sounds”. Usually such sounds
When using the GC Scale, all three components are recorded are produced when pain is applied. If there are no sounds,
separately. However, with a little experience it is often easy to even when pain is applied, the verbal response is “None”. In
assess two or three components at the same time, thus the case records the response is written as:
increasing efficiency. For example, if a patient does not open his
eyes, or speak, or obey commands on initial approach, whether VR Orient;
he then eye opens to pain, or makes incomprehensible sounds VR Conf;
in response to pain, can be assessed when pain is applied to test VR Words;
the motor response; there is no need to apply pain three times. VR Sounds;
When assessing responses the clinician should “start at the VR None.
top” and work down. So, for example, a good effort should be
made to see if the patient obeys commands before applying It can sometimes be difficult to distinguish confusion from
pain to test for a localising response. dysphasia and, if the patient is cooperative enough, formal
116 t h e s u r g e o n 1 0 ( 2 0 1 2 ) 1 1 4 e1 1 9
Fig. 1 e Neurosurgical observation chart showing deterioration in a patient with a left extradural haematoma, and recovery
after surgical evacuation.
tests for expressive or receptive dysphasia (naming objects,
carrying out complex commands) may be tried. If a patient
appears bright and alert but has difficulty speaking, they are
more likely to be dysphasic than confused, especially if they
have a right sided weakness, or evidence of left hemisphere
damage on the CT scan. On the observation chart, dysphasia is
recorded as “DYS”, written vertically (Fig. 1). If the patient has
a tracheostomy or is intubated, this is recorded on the chart as
“T”, not “None” (Fig. 1).
Motor response
The motor response (MR) is the most important of the three
responses as it carries the greatest prognostic significance, but
is also the response that causes the most difficulties in
assessment. Numerous variations of both the description of
the motor responses and the methods of eliciting them have
appeared in the literature, adding to the confusion.
Jennett and Teasdale in their original description in 19741
listed five motor responses, namely:
Obeys commands;
Localises pain;
Flexion to pain;
Extension to pain.
None
A second paper in 19766 included a further response of
“Spastic (or Abnormal) Flexion”, and the authors have been
criticised for making this change without adequate explana-
tion.7 The reasons for this response being considered some-
what differently are given below.
The term “Best” MR is used to describe the conscious level,
“Best” referring to the response in the best limb, not the best
response over a period of time. This implies that both upper
limbs need to be assessed in the event of there being
a difference between the two sides. For example, if the
patient is flexing on the right but extending on the left, their
conscious level is “Flexing” (the “Best” MR); the extensor
response on the right is an important focal, or localising, sign,
which is useful clinically, but is recorded separately under
limb responses (Fig. 1). The word “Best” also refers to the best
level of response shown in the better limb at the time of
assessment. Occasionally, determined testing during a single
examination will produce a higher response, or a mixed
response, and the best response is the one taken to represent
the conscious level.
In Jennett and Teasdale’s original paper, it was suggested
that pain be applied in several ways during any single
assessment to gauge the MR. Experience has shown that the
best way to learn and teach the MR is to follow the sequence
described below. The end result is the same, the point being
that there are standard ways of eliciting each of the responses
to pain.
 t h e s u r g e o n 1 0 ( 2 0 1 2 ) 1 1 4 e1 1 9
Obeys commands
This should always be tested for before applying pain, even if
the patient initially appears comatose, or is on a ventilator. If
the patient has a cervical injury, it is even more important to
test carefully for this response.
Assuming the patient can use both upper limbs, the most
efficient way to test for obeying commands is to ask the
patient to raise both arms, and hold them elevated. In this
way, any weakness of one upper limb should be evident (an
important focal sign). However, obeying any kind of command
will suffice, as long as the observer is sure that the movement
carried out by the patient is not simply a reflex action.
Localises pain
If the patient does not obey commands, test next for a local-
ising response. This is done by applying pressure to the
supraorbital ridge, sufficient to produce significant pain. If
positive, the patient will raise one hand to the site of stimu-
lation. By convention, if the hand gets above the clavicle, this
is classed as a localising response.
Some clinicians prefer not to apply supraorbital pressure,
and in the case of significant periorbital injury it would not be
appropriate anyway. A suitable alternative is pinching the
upper inner border of the trapezius muscle (the Mr Spock
death grip). Applying pain to the sternum, or indeed
anywhere below the clavicle, is not appropriate, for obvious
reasons.
The patient will almost always use one arm preferentially
to localise the pain. To test the other arm, hold the “preferred”
arm firmly down on the bed, and apply pain again. After
attempting to move the pinned limb, the patient will then
move the other arm if able to do so.
If the patient localises with both arms as described, there is
no need to test any further for the motor response.
Flexion to pain and extension to pain
These responses are tested by applying pressure with a pen or
pencil to the fingernail bed (Fig. 2). (Concern has been raised by
some that this can damage the nail, causing it to fall off later,
although the author has never seen this. Therefore, pressure
Fig. 2 e The standard method of applying pain when
testing for flexing or worse.
117
on the side of the terminal interphalyngeal joint is an alter-
native.) The arm is rested on the body with about 30e40 of
elbow flexion as the stimulus is applied. A good flexion
response is characterised by flexion of the elbow, often
accompanied by lifting the elbow clear of the body. An
extension response is shown by extension of the elbow,
usually accompanied by pronation of the forearm.
Both upper limbs are tested in the same way, and any
asymmetry of response noted (although the “Best” response
will be recorded as the Coma Scale).
If there is no movement of the upper limb in response to
nail bed pressure, record “No Motor Response”.
If a patient is deteriorating with raised intracranial pres-
sure, the time at which he starts extending to pain, or prog-
resses to no MR, usually correlates with the development of
brain herniation and the pupil(s) becoming fixed and dilated.
Progression to apnoea and death is not too far away.
However, a patient may give no MR because of high cervical
cord injury, brachial plexus or other severe limb injury, or due
to the effects of alcohol or other drugs, but in all these situa-
tions the pupils are likely to be reactive to light. In other
words, if the pupils are reactive, look for other reasons for an
apparent no MR.
Abnormal (or spastic) flexion
The response of spastic flexion sits between “flexion” and
“extension” in the Coma Scale. It is not a commonly-observed
response, and, if a patient with increasing intracranial pres-
sure deteriorates from obeying to localising to flexing, their next
worse response will usually be extending; they will not nor-
mally exhibit spastic flexion between flexing and extending.
Abnormal flexion seems to be correlated with damage to
the basal ganglia of the brain, and is often associated with
a severely disabled outcome. It is a response that is easier to
recognise than to describe. Nevertheless, the definition is11:
Abnormal flexion (spastic flexion) is present if there is
either:
1) Preceding extension movement in arms or
2) Extension in leg or
3) Two of these:
i. Stereotyped flexion posture;
ii. Extreme wrist flexion;
iii. Adduction of arm;
iv. Fingers flexed over thumbs.
In Glasgow, the response of abnormal flexion is not
included in the section of the observation chart for the Coma
Scale, but can be recorded under the individual limb
responses (Fig. 1). It is a response that usually only affects one
side anyway, and is therefore often more useful as a focal sign
than as a measure of conscious level. If there is doubt whether
a response is (normal) flexion or spastic flexion, it is best to
record “Flexion”.
Other terms used
It should be noted that phrases such as “reacts to pain”,
“responds to pain”, and “withdraws to pain” are not part of the
118 t h e s u r g e o n 1 0 ( 2 0 1 2 ) 1 1 4 e1 1 9
Glasgow Coma Scale (although the term “withdrawal” was
used once6 by authors from Glasgow to describe the difference
between flexion and abnormal flexion). The inclusion of such
terms in variations of the GC Scale is common, but causes
confusion, particularly when the terms are not defined. Like-
wise the words “decerebrate” or “decorticate” are not in the GC
Scale and are best avoided, implying as they do a “specific
physioanatomical correlation” for which there is no
evidence.1
The GC Score
Most readers will be aware that it is possible to ascribe
a numerical value to each of the responses of the GC Scale
(Table 1). The sum of each of the three numbers allocated to
the three responses gives the total GC Score. The numerical
value for the motor response depends on whether the
response of spastic flexion is included in the list of motor
responses, which in turn determines whether the highest GC
Score is 14 or 15.
The GC Score was devised to allow the information about
head injured patients to be stored and analysed on
a computer. The GC Score was not intended to be used to
describe individual patients in a clinical situation, but rather
to be used for research, comparison of series of head injured
patients, the development of guidelines, audit, and so on. It
may indeed be helpful on occasions to calculate the GC Score
for individual patients if the Score is required for other prog-
nostic indices, such as the Acute Physiology and Chronic
Health Evaluation (APACHE) Score. However, in these situa-
tions, the description of the patient in the case notes and
during handovers or referrals should always be in words (the
GC Scale), not numbers (the GC Score).
Nevertheless, despite pleas from the Glasgow group,12e16
many clinicians still use the GC Score in everyday clinical
practice. Therefore, it is perhaps worth emphasising that
there are at least five reasons why the GC Score is not
appropriate to describe individual patients in a clinical
situation:
1) It is frequently the case that the GC Score cannot be
calculated because one or more individual components of
the Score are not measurable. For example, if the eyes are
closed by severe swelling, the eye opening cannot be
assessed; if the patient has severe lower facial injuries, or is
intubated, or has a tracheostomy, the verbal response
cannot be assessed. To try to compose a Score in these
situations is unhelpful and misleading.
2) There are frequent mistakes in calculating the Score. Many
clinicians will admit in private that they use the GC Score as
a kind of analogue scale, and make a “guestimate” as to
where the patient sits within the range. Use of the Score
rather than the Scale encourages this sloppy clinical prac-
tice, and is perhaps one reason why some studies suggest
poor inter-observer variability.8,9
3) Using the GC Score rather than the GC Scale leads to loss of
useful information. For example, if a patient is described as
“GC Score ¼ 10”, this might mean that the patient obeys
commands, opens eyes to speech, and has no verbal
response; or, the patient may be localising to pain, eye
opening to pain, and speaking inappropriate words. Thus
the same GC Score can represent significantly different
clinical situations, which would clearly be described better
by using words, not a number.
4) To calculate the GC Score (if it is done properly) takes more
time than simply stating each individual response.
5) There is variation in practice regarding whether spastic
flexion is included in the motor response, which in turn can
lead confusion as to whether a 14- or 15-point Score is being
used.7 This can be overcome by using the denominator, but
it is better, surely, to describe the motor response in simple,
well-defined words.
Summary
The Glasgow Coma Scale remains the most commonly used
method of assessing conscious level, and, once learnt, is easy,
quick, and reliable to use. Of the three components of the GC
Scale, the motor response is the most important. Problems of
communication and inter-observer variability arise when
clinicians substitute non-standard, undefined, terms to
describe different motor responses, and when the GC Score is
used rather than the GC Scale.
The GC Score is useful for research, audit, prognostic
calculations, and other types of data collection that require
digitising and grouping of clinical information. The GC Score is
best avoided for everyday clinical use.
Conflict of interest
The author has no conflict of interest arising from the prep-
aration or publication of this article.
Acknowledgement
I am very grateful to Professor Sir Graham Teasdale for his
helpful advice during the preparation of this article.
references
1. Teasdale GM, Jennett B. Assessment of coma and impaired
consciousness: a practical scale. Lancet 1974;2:81e4.
2. Teasdale G, Knill-Jones R, van der Sande J. Observer
variability in assessing impaired conciousness and coma.
JNNP 1978;41:603e10.
3. Braakman R, Avezaat CJJ, Maas AIR, Roel M, Schonten HJA.
Interobserver agreement in assessment of the motor
response of the Glasgow ‘Coma’ Scale. Clin Neurol Neurosurg
1980;80-3:100e6.
4. Teasdale G, Knill-Jones R, Jennett B. Assessing and recording
conscious level. JNNP 1974;37:1268.
5. Matis G, Birbilis T. The Glasgow Coma Scale e a brief review:
past, present, future. Acta Neurolog Belg 2008;108:75e89.
6. Teasdale GM, Jennett B. Assessment and prognosis of coma
after head injury. Acta Neurochir 1976;34:45e55.
 t h e s u r g e o n 1 0 ( 2 0 1 2 ) 1 1 4 e1 1 9
7. Wiese MF. British hospitals and different versions of Glasgow
coma scale: telephone survey. BMJ 2003;327:782e3.
8. Bassi S, O’Reilly G, Buxton N, Punt J. Glasgow Coma Scale:
a help or hindrance. Br J Neurosurg 1999;13:526.
9. Marion DW, Carlier PM. Problems with initial Glasgow Coma
Scale assessment caused by prehospital treatment of patients
with head injuries: results of a national survey. J Trauma
1994;36:89e95.
10. Laureys S. Consciousness scales employed in the acute
setting. www.coma.ulg.ac.be/medical/acute.html.
11. Barlow P, Murray L, Teasdale G. Outcome after severe head
injury e the Glasgow model. In: Corbett WA, editor. Medical
applications of microcomputers. London: John Wiley & Sons Ltd.;
1987. Ch 6.
119
12. Teasdale G, Jennett WBJ, Murray G. BMA News 2004, January
17th:10.
13. Teasdale GM, Murray L. Revisiting the Glasgow Coma Scale
and Coma Score. Intensive Care Med 2000;26:153e4.
14. Jennett B. The Glasgow Coma Scale: history and current
practice. Trauma 2002;4:91e103.
15. Teasdale GM, Gentleman D, Andrews P, Blaiklock C,
Cruikshank E, Donnelly T, et al. Early Management of patients
with a head Injury. A national Guideline. SIGN Publication No. 46.
Edinburgh: Scottish Intercollegiate Guidelines Network, Royal
College of Physicians; 2000.
16. Griffiths J, Chandra-Bose RB. Different versions of the
Glasgow coma scale in British hospitals. Correspondence BMJ
10th January 2004;328:110.