A Glasgow Coma Scale total adds up three separate observations of a person's eyes, speech and movement, and one total can stand for very different responses. We checked every possible combination against the description of coma in NICE's head-injury guideline for England and Wales, which pairs a person "not obeying commands" who is also "not speaking and not eye opening" with "a GCS score of 8 or less". All 10 combinations that fit that description total 8 or less, yet 41 of the 51 combinations totalling 8 or less do not fit it. That count uses our reading of the words: eyes not opening at all, speech no more than sounds, and any movement short of obeying. Stricter readings fit 5 or 6 of the 51, and no reading fits a total above 8. Bodien and colleagues published the full set of 120 combinations in 2021, so our enumeration is a cross-check of theirs.
Middle totals blur different responses into a single number
Eye opening (E) scores from 1, none, to 4, spontaneous. Verbal response (V) runs from 1, no sound, to 5, orientated. Motor response (M) runs from 1, no movement, to 6, obeying commands. Added together they give a total from 3 to 15.
Those steps make 120 combinations. Bodien et al. (J Neurotrauma, 2021) counted them first and found each total "is associated with between one and 18 different combinations", with 9 the total reached 18 ways. Our count agrees: totals 3 and 15 have a single combination each, and every total from 4 to 14 has at least three.
In our count, every total from 8 to 12 includes a combination with M6, a person who obeys commands, and another with M3 or lower, abnormal flexion or worse. The scale's authors give their own example on their recording page: "a total score of 8 could be E2V2M4 or E1V1M6, with very different implications". These are counts of combinations, never of people.
The summed score came into focus in a 1979 paper
Graham Teasdale and Bryan Jennett described the scale in 1974. According to the authors' assessment aid page, the motor part then had five steps; a sixth was added in 1976, separating "normal" from "abnormal" flexion. The authors date the summed score to a 1979 paper titled "Adding up the Glasgow Coma Score". A structured approach published in Nursing Times in 2014 replaced "to pain" with "to pressure" and shortened "inappropriate words" and "incomprehensible sounds" to "words" and "sounds".
The five-step version lasted. In BMJ letters of January 2004, replying to a 2003 telephone survey of British hospitals, two Swansea neurosurgeons wrote that observation charts in their unit and at the Institute of Neurological Sciences in Glasgow "continue to describe the original 14 point scale" (PMC314062). Added up the same way, that version gives 100 combinations totalling 3 to 14, by our count.
Severity bands zoom out to groups of head-injury patients
Rimel et al. (Neurosurgery, 1982) divided head injury into "severe, 3-8; moderate, 9-12; and minor, 13-15". NICE NG232, published in 2023, uses the same cut-offs for mild, moderate and severe traumatic brain injury (injury to the brain from an outside force). Its reporting rules are older, tagged 2003: monitoring rests on the three separate responses, and the components always accompany the total. A 2014 Lancet Neurology review by Teasdale and colleagues says individual patients "are best described by the three components" and the total "should be used to characterise groups".
By our count, 51 of the 120 combinations fall in the severe band, 59 in the moderate and 10 in the mild. People fall very differently. Among 2,455 adults in Bodien's multi-centre head-injury study, 82% had a first emergency-department total of 13 to 15.
The coma description in the opening is NICE recommendation 1.8.8, tagged 2003 and amended 2007, in a section on transfer to a specialist neuroscience unit. Bodien's team applied published coma criteria to all 120 combinations, found coma possible at totals of 3 to 6, and wrote that "a GCS total score of 7 or 8 cannot result in a diagnosis of coma". NICE's wording, on our reading, admits three combinations at 7 or 8. Our CHA2DS2-VA guide takes a similar look at a stroke-risk score.
Untestable responses obscure the total
Some responses cannot be assessed: eyes closed by swelling, speech blocked by a breathing tube, movement stopped by paralysis. The authors' FAQ says "Do not use number '1' to record missing component; use 'NT' (Not testable)" and "Do not report a total score when a component is Not Testable", because a low total could imply someone is more unwell than they are. It adds that estimating a missing component statistically is "probably more relevant to research". In Bodien's study, swelling, intubation or paralysis was recorded in 30,139 of 69,487 assessments, 43% on our arithmetic.
Until 15 October 2026 our Glasgow Coma Scale calculator had no NT option, so a person unable to speak because of a breathing tube could be entered as V1, with a total and a mild, moderate or severe label printed. Its text wrote the notation as "VT" and said the children's scale had the same scoring range. The revision made that day offers NT for each component, leads the result with E, V and M, prints no total and no category when any component is NT, and labels the category as a head-injury one, citing Rimel and NICE. With all three scored, it says how many of the 120 combinations share that total.
A total from any calculator or chart describes three observations. It is not an assessment, a diagnosis or a forecast for one person.
A paediatric version refocuses the scale for young children
The authors' FAQ says the scale "can be applied without modification to children over 5 years old". Younger children usually cannot be orientated or obey commands, so the authors print a paediatric scale described in Adelaide, whose five-step motor part gives a highest total of 14. The FAQ adds that "several other systems have been put forward". NICE's paediatric version, from 2003, includes a "grimace" alternative to the verbal score for children who do not yet speak.
GCS-Pupils sharpens the bottom of the scale
Brennan, Murray and Teasdale (J Neurosurg, 2018) described GCS-Pupils (GCS-P): the total minus a pupil reactivity score (PRS), the number of pupils, 0 to 2, that do not react to light. GCS-P runs from 1 to 15. In pooled head-injury data from the CRASH trial (9,045 patients) and the IMPACT database (6,855), a total of 4 went with worse outcomes than a total of 3, a pattern not seen with GCS-P. The authors' GCS-P page says it is "not intended to replace the role of separate assessment and reporting of each component".
Counting three pupil states against the 120 combinations gives 360 combinations, by our count, with GCS-P 8 the most crowded at 52.
View author profile → · Editorial policy
