PII S0736-4679(00)00182-7

MedicalClassics

THE GLASGOW COMA SCALE

George L. Sternbach, MD

Department of Emergency Medicine, Stanford University Medical Center, Stanford, CaliforniaReprint Address: George L. Sternbach, MD, 539 15th Avenue, Menlo Park, CA 94025

e Abstract—Teasdale and Jennett first presented theGlasgow Coma Scale in 1974 as an aid in the clinicalassessment of unconsciousness. It was devised as a formalscheme to overcome the ambiguities and misunderstand-ings that arose when information about comatose patientswas presented and groups of patients were compared. Sincethen, the Glasgow Coma Scale has been used extensively,being used to grade individual patients, compare effective-ness of treatments, and as a prognostic indicator. It hasbeen incorporated into numerous trauma and critical ill-ness classification systems. However, a number of compet-ing scales have been developed to overcome its perceiveddeficiencies. These scales are generally more complex. Oneof the expressed reservations regarding the Glasgow ComaScale has been its failure to incorporate brainstem reflexes.The scale also includes a numerical skew toward the motorresponse. An important current issue is the appropriateapplication of the Glasgow Coma Scale to intubated pa-tients. A number of approaches have been used to assign theverbal score to such patients. The timing of initial scoring isanother area of discussion. Despite its drawbacks, the Glas-gow Coma Scale remains the most universally utilized levelof consciousness scale worldwide. It seems destined to beused in emergency medicine for some time. © 2000Elsevier Science Inc.

e Keywords—coma; head injury

HISTORY

“Impaired consciousness is an expression of dysfunctionin the brain as a whole,” wrote Teasdale and Jennett in

1974, that “may be due to agents acting diffusely . . . orto the combination of remote and local effects producedby brain damage which was initially focal” (1). Theauthors were from the Glasgow University Departmentof Neurosurgery Institute of Neurologic Sciences. Intheir view, the clinical assessment of unconsciousnesssuffered from the practice by many physicians to “retreatfrom any formal scheme in favor of a general descriptionof the patient’s state, without clear guidelines as to whatto describe and how to describe it” (1). This in turn ledto “ambiguities and misunderstandings when informa-tion about patients is exchanged and when groups ofpatients treated by alternative methods are compared”(1).

Did the Glaswegians have a remedy for this state ofaffairs? Indeed they did. They presented the GlasgowComa Scale (GCS), and rationalized its three compo-nents by the following reasoning:

● Motor response: “The ease with which motor re-sponses can be elicited in the limbs, together with thewide range of different patterns which can occur,makes motor activity a suitable guide to the function-ing state of the central nervous system” (1).

● Verbal response: “Probably the commonest definitionof the end of a coma, or the recovery of consciousness,is the patient’s first understandable utterance” (1).

● Eye opening: “Spontaneous eye opening . . . indicatesthat the arousal mechanisms in the brainstem are ac-tive” (1).

Medical Classicsis coordinated byGeorge Sternbach, MD, of Stanford University Medical Center, Stanford, California

RECEIVED: 6 July 1999; FINAL SUBMISSION RECEIVED: 20 January 2000;ACCEPTED: 8 February 2000

The Journal of Emergency Medicine, Vol. 19, No. 1, pp. 67–71, 2000Copyright © 2000 Elsevier Science Inc.Printed in the USA. All rights reserved

0736-4679/00 $–see front matter

67

The clinical scale consisted of 14 points, four for eyeopening, five each for best verbal and best motor re-sponse. A sixth point on the motor response scale (“with-drawal from painful stimulus”) was to be added 2 yearslater (2). The scale was simple, a feature the authorsdeemed essential, though they cautioned that “the searchfor simplicity must not be the excuse for seeking abso-lute distinctions where none exist” (1).

Teasdale and Jennett did not, at this point, assignnumerical scores to the various aspects of their scale.This was to follow in 1976 (2). Nor did they present anydata demonstrating the scale’s validity or reliability, ex-cept to report that physician and nurse examiners at theirinstitution were highly consistent in its use. Disagree-ment among these observers was rare. This was desir-able, inasmuch as “repeated observations of consciouslevel are usually made by relatively inexperienced doc-tors and nurses” (1).

The authors anticipated criticism of a system thatseemed to “undervalue the niceties of a full neurologicexamination” (1). They were quick, therefore, to reassurethe reader that it was “no part of our case to deny thevalue of a detailed appraisal of the patient as a whole,and of neurologic function in particular” (1).

DISCUSSION

The Glasgow group was not the first to have formulateda coma scale. In 1966, Ommaya described a five-pointlevel of consciousness scale he had used in conjunctionwith a clinical study of head trauma (3). The stages ofthis scale ranged from “the state of normal conscious-ness” at the top, down to “totally unresponsive to allstimuli.” The three levels in between were defined bydescending combinations of orientation and responses tostimuli. This scale has subsequently been faulted forbeing too simple (4). Moreover, it appears to suffer fromthe “general description of the patient’s state” deficiencythat Teasdale and Jennett raised. It has, in any event,never achieved widespread use.

The GCS (see Table 1), on the other hand, has pros-pered. It has enjoyed extensive acceptance as an impor-tant tool in the care of trauma patients, becoming anintegral part of the American College of Surgeons Ad-vanced Trauma Life Support courses. The GCS has beenused to grade individual levels of consciousness, com-pare effectiveness of treatment, and as a prognostic in-dicator. It is used to codify injury severity by the Inter-national Coma Data Bank and the United States NationalTraumatic Coma Data Bank (5). Champion et al. incor-porated it into the Trauma Score in 1981 and the RevisedTrauma Score 8 years later (6,7). It is a component of theAcute Physiology and Chronic Health Evaluation

(APACHE) II score, the Trauma and Injury SeverityScore (TRISS) and the Circulation, Respiration, Abdo-men, Motor, Speech (CRAMS) Scale.

The GCS was initially developed to gauge coma de-terioration or improvement as well as a predictor ofultimate outcome (1). It has subsequently been utilized asa clinical indicator for management. The best knownexample of this use is the well-known recommendationthat a patient with a GCS score of 8 or less is unable toprotect the airway and requires endotracheal intubation(8). Another use is the categorical division of severity ofhead injury by GCS score of mild (13–15), moderate(9–12), and severe (8 or less) (9,10).

Despite the widespread acceptance of the GCS, thescale has not been embraced entirely without reservation.It has been accused of occupying a “privileged but un-warranted position in clinical and investigative contexts”(11). A number of competing scales have been developedto overcome its perceived deficiencies.

The Reaction Level Scale (RLS85) represents onesuch attempt to improve upon the GCS (see Table 2). Aneight-point scale that ranges between “alert” and “uncon-scious,” the RLS85 bears some resemblance to Om-maya’s level of consciousness scale. It has been praised

Table 1. Glasgow Coma Scale

Eye openingSpontaneous 4To speech 3To pain 2None 1

Best verbal responseOriented 5Confused conversation 4Inappropriate words 3Incomprehensible sounds 2None 1

Best motor responseObeys commands 6Localizes pain 5Withdrawal (normal flexion) 4Abnormal flexion (decorticate) 3Extension (decerebrate) 2None 1

Table 2. Reaction Level Scale (RLS85)

1 Alert; no delay in response2 Drowsy or confused; responsive to strong stimulation3 Very drowsy or confused; responsive to strong stimulation4 Unconscious; localizes but does not ward off pain5 Unconscious; withdrawing movements on pain stimulation6 Unconscious; stereotype flexion movements on pain

stimulation7 Unconscious; stereotype extension movements on pain

stimulation8 Unconscious; no response to pain stimulation

68 G. L. Sternbach

for improving discriminatory ability by combining eye,verbal and motor responses into a single scale (11). TheRLS85 has been touted as demonstrating greater accu-racy and higher inter-observer agreement than the GCS(11). The information content of the GCS and the RSL85is similar, though their construction differs. There is ahigh correlation between the two scales, indicating asimilar ranking order of neurologic defect severity (12).The Swedish Societies of Neurosurgery, Anesthesiologyand Intensive Care have recommended replacement ofthe GCS with the RLS85 in that country’s hospitals (13).Consequently, this scale has been used in this way inSweden since 1984 (13,14).

One of the expressed reservations regarding the GCShas been its failure to incorporate brainstem reflexes. Anumber of investigators have disagreed with Teasdaleand Jennett that spontaneous eye opening is sufficientlyindicative of brainstem arousal systems activity and havefashioned coma scales that include brainstem responses(11,15). The resulting scales generally have been morecomplex than the GCS. The Comprehensive Level ofConsciousness Scale is a behavioral scale that includesbrainstem indicators. In addition to eye opening, motorfunction, and communicative effort, it assesses posture,eye position at rest, abnormal ocular movements, pupil-lary light reflexes, and general responsiveness. Althoughit is only an eight-item scale, the maximum score foreach category ranges from 4 to 8 points, allowing for apotential high score of 48 (16). The Clinical NeurologicAssessment Tool is a 21-item scale that includes variousmotor, reflex, and communication categories. The max-imum possible score is 101 points (17). Another 100-point head injury scale was developed by Bouzarth in1968 and modified in 1978 (14,18). This relies heavilyon vital signs and variables of the GCS, as well aspupillary findings.

The Maryland Coma Scale incorporates aspects of theGCS, brainstem (pupillary, corneal, and oculovestibular)reflexes, as well as a grading of the nature of the stimulusrequired to elicit motor and verbal responses (19). TheInnsbruck Coma Scale contains pupillary size and reac-tion, movement and position of the eyes, reaction toacoustic stimuli, eye opening, and motor responses, butexcludes verbal response. Its eight items allow for amaximum of 23 points (20).

A simpler system is the Glasgow-Liege Scale (seeTable 3), in which the presence of five brainstem reflexeshas been appended to the GCS. The five reflexes selecteddisappear in descending order during rostral-caudal de-terioration. The disappearance of the last, the oculocar-diac, coincides with brain death (21).

The GCS has been criticized for lacking reliability inmonitoring levels of consciousness or predicting out-come in patients with middle range (GCS 9–12) scores

(11). It also has been critiqued on a purely mathematicalbasis. Bhatty and Kapoor note that, because there areonly four units assigned to the eye responses, versus fiveto the verbal and six to the motor responses, the scaleincorporates a numerical skew toward motor response(4). In achieving a sum score, eye response has the leastinfluence on the total and the motor response the most,with verbal response being intermediate. They have sug-gested weighting individual scores for eye, verbal, andmotor responses in such a way that each would have aminimum contribution of one point and a maximum offive.

This effort to provide mathematical parity for thethree components of the GCS has abutted against studiesthat have stressed the particular importance of the motorportion of the score. The conclusion of a number ofinvestigators has been that the motor score is more im-portant than either of the other two components in pre-dicting the magnitude of neurologic injury for patientswith severe head injury (22–25). The motor responsealone is considered by some to be the best predictor ofshort-term outcome from head injury, with verbal andeye scores being more pertinent in patients who are not,in fact, comatose (i.e., those with a GCS of 9–15) (22).

There is some controversy over whether accuracy isdiminished by adding the numerical scores of the threecomponents (22,26). The Glasgow group admits to “lim-itations inherent in the summation of the three re-sponses” (27). They view a total score as “merely aconvenient method for summarizing data.” Indeed, theypoint out that “in Glasgow patients under treatment arealways described by the three separate responses andnever by the total” (27).

One of the most important current issues regarding theGCS is its application to intubated patients. The use ofearly intubation and administration of neuromuscularparalyzing agents in the pre-hospital phase of care has

Table 3. Glasgow-Liege Scale

Glasgow Coma Scale score: and the presence of thefollowing brainstem reflexes:

PointsFronto-orbicular* 5Vertical oculovestibular† 4Pupillary light 3Horizontal oculovestibular† 2Oculocardiac‡ 1No response 0

* The reflex is considered present when percussion of the gla-bella produces contraction of the orbicularis oculi muscle.† Deviation of at least one eye is induced by repeated flexion andextension (vertical) or horizontal neck movement (horizontal). Ifthe cervical spine is immobilized, an attempt is made to elicitocular motion by simultaneous external auditory canal irrigationusing iced water.‡ Pressure on the eyeball causes the heart rate to slow.

The Glasgow Coma Scale 69

rendered the verbal response portion of the scale unmea-surable in these cases. Early treatment was uncommonwhen the GCS was first described, but has since gainedgreater acceptance. A number of approaches have beenused to overcome this limitation. Some of the alternatescales described above, most notably the RLS85, do notinclude a verbal response criterion.

Several techniques have been proposed to designatethe verbal score in intubated patients. One method in-volves assigning an arbitrary score of one point to allpatients on mechanical ventilation (5). Another sugges-tion has been that a “pseudo-score” be applied, in whichthe average value of the testable scores is calculated andadded to the sum in lieu of the verbal score (28). Linearregression prediction of the verbal scores based on theother two scores also has been utilized (30). Many majorhead injury centers in the United States use a non-numerical designation of “T” for the verbal score inintubated patients (14,29).

The timing of the initial evaluation is another area ofdiscussion. In 1977, Jennett and Teasdale specified thatthe initial score be assigned 6 h after head injury hadbeen sustained (29). They felt this time interval allowedfor the diagnosis and management of other injuries thatmight affect neurologic function. Waiting 6 h avoidedoverestimating the extent of brain damage produced bytransient influences such as shock and respiratory insuf-ficiency. This time period was chosen to exclude patientswho might display coma temporarily, due to factors otherthan head injury, such as hypoxia, hypotension, or alco-hol intoxication (31). Progress since then in pre-hospital

and inter-hospital transport, as well as advances in diag-nostic and treatment modalities, have markedly reducedthe time between injury and definitive treatment. Thistreatment is now far more likely to occur in under 6 h.

Assigning the GCS score too soon after injury mayalso be problematic. A study of victims of blunt headinjury for whom paramedic response times averaged 5min revealed that GCS scores calculated by paramedicson the scene had no prognostic value (30). These assess-ments were apparently performed too early in the courseof the injury. As such, they reflected an initial general-ized depression of neurologic status that did not neces-sarily correlate with the severity of head injury or withoutcome (14). A favorable outcome, on the other hand,did correlate with significant improvement in emergencydepartment GCS score over that calculated in the field(32).

Despite its drawbacks, the GCS remains the mostuniversally utilized level of consciousness scale world-wide. Though some of the more detailed scales may bemore suitable for assessment of inpatients, the GCS, byvirtue of its simplicity, seems destined to be used inemergency medicine for some time. The popularity ofRLS85 in Sweden notwithstanding, the GCS is unlikelyto be supplanted soon, both because it has been widelydisseminated and because it “is so entrenched in all of themathematics and all the computer studies” of headtrauma (29). A presenter at the 1997 meeting of theAmerican Association for the Surgery of Trauma, whenasked his opinion of the RLS85, admitted no knowledgeof it. “It sounds like a hand cleaner,” he replied (28).

REFERENCES

1. Teasdale G, Jennett B. Assessment of coma and impaired con-sciousness. A practical scale. Lancet 1974;2:81–4.

2. Teasdale G, Jennett B. Assessment and prognosis of coma afterhead injury. Acta Neurochir 1976;34:45–55.

3. Ommaya AK. Trauma to the nervous system. Ann Roy Coll SurgEngl 1966;39:317–47.

4. Bhatty GB, Kapoor N. The Glasgow Coma Scale: a mathematicalcritique. Acta Neurochir 1993;120:132–5.

5. Marshall LF, Becker DP, Bowers SA, et al. The National Trau-matic Coma Data Bank. Part I: design, purpose, goals, and results.J Neurosurg 1983;59:276–84.

6. Champion HR, Sacco WJ, Carnazzo AJ, et al. Trauma score. CritCare Med 1981;9:672–6.

7. Champion HR, Sacco WJ, Copes WS, et al. A revision of theTrauma Score. J Trauma 1989;29:623–9.

8. Chesnut RM. The management of severe traumatic brain injury.Emerg Med Clin North Am 1997;15:581–604.

9. Rimel RR, Giordani NP, Barth JT, et al. Disability caused by minorhead injury Neurosurgery 1981;9:221–8.

10. Rimel RR, Giordani NP, Barth JT, Jane JJ. Moderate head injury:completing the spectrum of brain trauma. Neurosurgery 1982;11:344–51.

11. Segatore M, Way C. The Glasgow Coma Scale: time for a change.Heart Lung 1992;21:548–57.

12. Starmark JE, Stalhammar, D, Holmgren E, Rosander B. A com-

parison of the Glasgow Coma Scale and the Reaction Level Scale(RLS85). J Neurosurg 1988;69:699–706.

13. Starmark JE, Holmgren E, Stalhammar D, et al. Reliability andaccuracy of the Glasgow coma scale. Lancet 1991;337:1042–3.

14. Marion DW, Carlier PM. Problems with initial Glasgow ComaScale assessment caused by prehospital treatment of patients withhead injuries: results of a national survey. J Trauma 1994;36:89–95.

15. Bouzarth WF, Lindermuth JR. Head injury watch sheet modifiedfor a digital scale. J Trauma 1978;18:571–9.

16. Stanczak DE, White JG, Gouview WD, et al. Assessment of levelof consciousness following severe neurological insult. A compar-ison of the psychometric qualities of the Glasgow Coma Scale andthe Comprehensive Level of Consciousness Scale. J Neurosurg1984;60:955–60.

17. Crosby L, Parsons LC. Clinical neurologic assessment tool: devel-opment and testing of an instrument to index neurologic status.Heart Lung 1989;18:121–9.

18. Bouzarth WF. Neurosurgical watch sheet for craniocerebraltrauma. J Trauma 1968;8:29–31.

19. Salcman M, Schepp RS, Ducker TB. Calculated recovery rates insevere head injury. Neurosurgery 1981;8:301–8.

20. Benzer A, Mitterschiffthaler G, Marosi M, et al. Prediction ofnon-survival after trauma: Innsbruck Coma Scale. Lancet 1991;338:977–8.

70 G. L. Sternbach

21. Born JD. The Glasgow-Liege Scale. Acta Neurochir. 1988;91:1–11.

22. Jagger J, Jane JA, Rimel R. The Glasgow Coma Scale: to sum ornot to sum? [letter] Lancet 1983;2:97.

23. Levati A, Farina ML, Vacchi G, et al. Prognosis of severe headinjuries. J Neurosurg 1982;57:779–83.

24. Choi SC, Narayan RK, Anderson RL, Ward JD. Enhanced speci-ficity of prognosis in severe head injury. J Neurosurg 1988;69:381–5.

25. Colohan ART, Alves WM, Gross CR, et al. Head injury mortalityin two centers with different emergency medical services andintensive care. J Neurosurg 1989;71:202–7.

26. Stalhammar D, Starmark JE. Assessment of responsiveness in headinjury patients. The Glasgow Coma Scale and some comments onalternative methods. Acta Neurochir. 1986;36:91–4.

27. Teasdale G, Jennett B, Murray L, Murray G. Glasgow Coma Scale:to sum or not to sum? [letter] Lancet 1983;2:678.

28. Grahm TW, Williams FC, Harrington T, Spetzler RF. Civiliangunshot wounds to the head: a prospective study. Neurosurgery1990;27:696–700.

29. Meredith W, Rutledge R, Fakhry SM, et al. The conundrum of theGlasgow Coma Scale in intubated patients: a linear regressionprediction of the Glasgow verbal score from the Glasgow eye andmotor scores. J Trauma 1998;44:839–45.

30. Jennett B, Teasdale G, Galbraith S, et al. Severe head injuries inthree countries. J Neurol Neurosurg Psych 1977;40:291–8.

31. Jennett B, Teasdale G. Aspects of coma after severe head injury.Lancet 1977;1:878–81.

32. Winkler JV, Rosen P, Alfry EJ. Prehospital use of the GlasgowComa Scale in severe head injury. J Emerg Med 1984;2:1–6.

The Glasgow Coma Scale 71