Last year, the venerable Glasgow Coma Scale (GCS), the most commonly used scoring system for assessing consciousness after an acute brain injury, celebrated its 40th birthday. But now, at age 41, the GCS seems to be having a bit of a midlife crisis.
It was 1974 when the GCS was originally published by Sir Graham Teasdale, MD, and the late Professor Bryan Jennett, MD, two leading neurosurgeons at the Institute of Neurological Sciences in Glasgow, Scotland. (Lancet 1974;304:81.)
In the intervening four decades, it has become the standard of care for virtually every hospital and EMS unit in the United States and much of the developed world. Whether you've been in a car crash, fallen off a roof, or taken a hit while playing left tackle, chances are that your medical record contains notes about your GCS score.
Dr. Teasdale and colleagues reviewed the GCS and published a new “structured approach” to the tool on the occasion of its birthday last year, including a downloadable “do it this way” guide and video. Despite its widespread use, the GCS has been criticized for being too complicated and unreliable. Last October, Steven Green, MD, the deputy editor of the Annals of Emergency Medicine and a professor of emergency medicine at Loma Linda University, sent a proposal to the American College of Emergency Physicians' EMS committee, calling for the replacement of the GCS with a simplified scale. “Their response was that they are working on it, but are balancing it with other tasks,” he said.
In the letter and in a 2011 review, Dr. Green pointed to numerous studies documenting the limitations of the GCS. (Ann Emerg Med 2011;58:427.)
Unsatisfactory Interrater Reliability
The first line of his editorial left little room for wondering about Dr. Green's view on the GCS: “It is time to abandon the Glasgow Coma Scale,” he wrote, adding that the scale was “confusing, unreliable, and unnecessarily complex,” not to mention “statistically unsound.”
A prospective study he and colleagues at Loma Linda published from their Level I trauma center had two residency-trained emergency physicians independently assess 116 brain-injured patients using the GCS. (Ann Emerg Med 2004;43:215.) Scores were exactly the same in 38 percent of cases but two or more points apart in 33 percent. Other studies have found somewhat better interrater reliability, but still showed a not-inconsiderable number of cases at least two points apart.
“We recently had a patient in the trauma bay who had sustained head trauma in a fall. The head trauma appeared to be the only major injury,” said Lauren Westafer, DO, MPH, the chief emergency medicine resident at Baystate Medical Center/Tufts University, who wrote a detailed post in June criticizing the GCS on her blog, The Short Coat. “The EMS personnel, the person doing the airway management, and me as the trauma leader — we all came up with a different GCS. These differences can be very serious; they can mean trauma activation or not, and change how the patient is managed.”
The 15-point scale “is widely perceived as complicated,” Dr. Green said, noting one study in which only 15 percent of military physicians, all of whom were trained to use the scale, could correctly calculate a GCS, with many unable to identify the titles of the specific categories. (J Neurotrauma 2005;22:1327.)
Even more telling was a 2003 survey that found that one-fourth of British hospitals were using the original 14-point version of the GCS rather than the current 15-point version without anyone noticing. (BMJ 2003;327:782.) And most recently, another study found that EMS staffers accurately scored the verbal component of the GCS 69.2 percent of the time, the eye-opening component 61.2 percent of the time, and the motor component 59.8 percent of the time. More than nine percent of study participants were assigned nonexistent GCS scores. (Prehosp Disaster Med 2015;30:46.) “If many or most clinicians cannot reliably retain knowledge of the GCS scoring sequence, then how can they be expected to correctly apply the tool?” Dr. Green asked.
Indeed, a Scandinavian physician who blogs about anesthesia, emergency medicine, and critical care at ScanCrit (who declined to give his name to EMN) wrote in 2011, “Whenever my colleagues or nurses ask me what GCS a patient has, I just make something up that sounds about right. Judging by their smirks I know they think that ... I didn't check properly. The truth is far worse. I don't know the Glasgow Coma Scale. Despite med school and my 10 years as a doctor having worked thousands of patients where the GCS is relevant, I still haven't managed to learn the damn thing. I guess it in so many ways never made sense to me, and furthermore I can't think of one single instance where it affected my clinical decision-making.”
Limited Prognostic Value
The GCS is “only grossly predictive” and most useful at its margins, Dr. Green said. Initial GCS scores are even less predictive than later scores, a 1991 analysis found. A total of 68.6 percent of patients were predicted to have a good outcome, and 76.5 percent of those predicted to have a poor outcome actually had those outcomes one year later. Later GCS scores more accurately predicted a good outcome (80.6%), but there was little change in accuracy for poor outcomes (78.6%). (Ann N Y Acad Sci 1991;620:82.)
On the other hand, patients with apparently mild traumatic brain injury can easily score a 15 (normal) on the GCS simply by opening their eyes spontaneously, being able to move appropriately on command, and knowing their name, the date, and their location.
With all these problems with the GCS, what should replace it? Dr. Green noted that the individual subscales of the GCS have been shown in various studies to perform just as well as the combined total, with the six-point motor component exhibiting the best predictive ability.
Another scale, known as the FOUR scale (Full Outline of Unresponsiveness), which measures eye response, motor response, brainstem reflexes, and respiration pattern, also has good prognostic capability that appears similar to the GCS, with the added advantage that all components of the score can be rated even on intubated patients. The GCS verbal score cannot be accurately calculated if intubation or any other external factor renders the patient unable to speak, even if conscious. (Mayo Clin Proc 2009;84:694.)
Investigators from the department of anesthesia and critical care at Royal London Hospital describe two four-point scores that also yielded in their review similar accuracy to the GCS. (Anesthesia 2004;59:34.) The first, AVPU, categorizes patients as alert, responds to verbal stimuli, responds to painful stimuli, and unresponsive to all stimuli. The other, ACDU, categorizes patients as alert, confused, drowsy, or unresponsive.
Then there's the three-point Simplified Motor Score (SMS), defined as obeys commands=2, localizes pain=1, and withdraws to pain or worse=0, which also performed about as well as the GCS in predicting TBI outcomes, at least in the out-of-hospital setting. (Ann Emerg Med 2011;58:417.)
“While I like the simplified motor score myself, I believe that the GCS could be suitably replaced by any of the simplified scales mentioned in my editorial,” Dr. Green said.
Many experts agree that the GCS has flaws, but its creators have said they never recommended using it alone to assess the severity of brain damage or predict outcome. Even so, it doesn't seem likely that the widely-accepted scale will go away any time soon.
“There's a lot of history intermingled with evidence in medicine, such that even when literature establishes new evidence, translation of that literature to clinical practice lags very far behind,” said David Thompson, MD, MPH, an assistant clinical professor of emergency medicine at the University of California, San Francisco and the lead author of the Annals article assessing the SMS tool. “Even if we've established that this three-point scale correlates to outcomes as well as a 15-point scale, people are used to the GCS and are hesitant to change.”
What is most important about using the GCS in a clinically effective way, he said, is the detail in which it is calculated. “Although I wrote about the GCS and think about it a lot, I still can't remember it. For that reason, we have it posted on the wall of all our trauma rooms. You have to be very careful to always express not just the total score, but the breakdown of the categories: eyes, motor, and verbal. You might have your eyes open and that gets you a lot of points, but you're not doing anything else. A patient in a stunned, paralyzed condition with his eyes open is very concerning even though the GCS might not be that low.”
Dr. Thompson noted a tendency in the ED to lump patients into a “gestalt” GCS number, without careful calculation. “If a patient is slightly confused, you just give them a 14. If they're unresponsive, they get a 3. If they have a depressed level of consciousness but are still moving, they get an 8. It's much more useful to go through the three categories and actually calculate the score,” he said.
Changes in the GCS are also clinically important, Dr. Thompson added. “A patient with a high GCS whose number is declining is very concerning. We would treat that patient much more aggressively. For example, a patient with an epidural hematoma is classically associated with rapidly deteriorating neurologic function. Those patients need an urgent head CT and may need to have their airway managed. Alternatively, a low initial GCS that is improving is reassuring. That's an important reason the GCS needs to be done serially over time.”
Dr. Green agreed that the specific scoring elements have value in the ED, as opposed to the prehospital setting, because they can serve as a baseline to monitor patient progress. “However, I see no ED value to assigning numbers to these elements and totaling them,” he said. “Every time I hear someone say something in the ED like ‘The GCS is 10,’ the response is almost always, ‘What were the specific components?’ Giving a single number doesn't adequately relay the clinical information. Why not just state the component responses themselves and omit the scoring oversimplification?”
And no score — whether it's GCS or any of its components, FOUR, AVPU, ACDU, SMS, or anything else — can substitute for detailed communication among providers. “There's no golden goose,” said Dr. Westafer. “It's easier to just say that a patient has a GCS of 14 than to describe in detail, whether in the clinical note or while on the phone with another clinician, what the patient is doing, but it would behoove us and benefit the patient to make the effort to go into that detail. Maybe the GCS motor score can predict mortality in situations — that's what the literature shows — but I don't think a score will ever be adequate in our assessment of these patients. We need to be communicating with each other.”
Share this article on Twitter and Facebook.
Access the links in EMN by reading this on our website or in our free iPad app, both available at www.EM-News.com.
Comments? Write to us at email@example.com.