Date: February 3rd, 2015
Guest Skeptics: Dr. Anand Swaminathan (EM Swami). Swami is an assistant program director at NYU/Bellevue Hospital in the department of EM. He is also part of the REBEL EM Alliance and as of January 2015 part of the East Coast Team doing Emergency Medical Abstracts (EMA).
Dr. Emily Junck. She is a third year Emergency Medicine resident physician at University of Washington. Emily created a list of classic practice-changing Emergency Medicine articles called 52 Articles in 52 Weeks designed for EM interns to read throughout their intern year.
Case: 62-year-old woman who has an unwitnessed ground-level fall while walking on ice and hits her head. There was a possible brief loss of consciousness.. Her daughter brought her to the emergency department right away concerned that she might have hurt herself. The patient complains of a headache, but has not vomited and denies other complaints. The daughter feels her mother is acting appropriately. On your evaluation, the patient has a posterior scalp contusion, but no palpable step-off, and otherwise has Glasgow Coma Scale 15 and a non-focal neurological examination.
Question: Does this patient need a head CT to rule out a clinically significant brain injury?
Background: CT scans are frequently done after head injury to evaluate for intracranial hemorrhage, which can be costly and causes radiation-exposure. Much of the time, these are negative, or find injuries for which no intervention is ever done and do not clinically affect the patient. CT Head decision rules help clinicians decide when to order a CT.
Reference: Stiell et al. The Canadian CT Head Rule for patients with minor head injury. Lancet. 2001
- Population: Adult patients presenting to the emergency departments at 10 large Canadian hospitals with Glasgow Coma Scale 13 or greater within 24 hours after blunt head trauma resulting in witnessed loss of consciousness, amnesia, or witnessed disorientation.
- Age under 16
- Minimal head injury with no LOC, amnesia, or disorientation
- Unclear history of trauma as the primary event (ie primary seizure or syncope)
- Obvious penetrating skull injury or depressed fracture
- Acute focal neurological deficit
- Unstable vital signs associated with major trauma
- Seizure prior to ED assessment
- Anticoagulation or bleeding disorder
- Intervention: Standardized clinical assessments were performed on all consecutive eligible patients before performing a CT scan at the discretion of the attending physician.
- Comparison: They took all the pre-CT variables and compared them with the CT and outcomes at 14 days looking for associations. From there they selected the combination of variables that would give the highest sensitivity/specificity for detecting the outcome measures using logistic regression and recursive partitioning. Overall, 44 variables were assessed (24 primary and 20 created by cut-offs/combinations)
- Primary Outcome: Need for neurological intervention, defined as need for neurologic intervention as death within 7 days due to the head injury or need with 7 days for craniotomy, elevation of skull fracture, intracranial pressure monitoring, or intubation for head injury.
- Secondary Outcome: Clinically important brain injury on CT, defined as any acute brain finding on CT that would normally require admission to a hospital and neurological follow-up.
Authors’ Conclusions: “We have developed the Canadian CT Head Rule, a highly sensitive decision rule for use of CT. This rule has the potential to significantly standardise and improve the emergency management of patients with minor head injury.”
- The study population included or focused on those in the ED. Yes.
- The patients were representative of those with the problem. Yes
- All important predictor variables and outcomes were explicitly specified. Yes.
- This is a prospective, multicenter study including a broad spectrum of patients and clinicians (level II). NO.
- Clinicians interpret individual predictor variables and score the clinical decision rule reliably and accurately. Yes.
- This is an impact analysis of a previously validated CDR (level I). NO.
- The variables used were taken from previous literature and generally believed to be variables associated with more likely brain injury. However, the CDR presented in this article and the telephone criteria to rule out those patients with clinically important brain injuries were not previously validated; however the telephone criteria were internally validated during this study.
- For Level I studies, impact on clinician behavior and patient-centric outcomes is reported NO
- The follow-up was sufficiently long and complete Yes.
- The effect was large enough and precise enough to be clinically significant. Yes
- 3121 patients were enrolled and they were able to assess their primary outcome measure, need for neurological intervention, in 100% of patients.
- Mean age of the population was 39 with two-thirds being male
- The most common mechanisms were fall, MVC, assault and head hit by object, sports-injuries, and pedestrian versus vehicle.
- Initial GCS was 15 in 80% of patients.
- 2078 were scanned (67%) meaning 1043 were not scanned (33%)
- A tool (rule) was created including 7 variables formed through logistic regression followed by recursive partitioning.
- 5 high-risk variables (need for neurosurgical intervention)
- 2 medium-risk variables (for brain injury on CT)
- The 5 high-risk criteria had 100% sensitivity and 68.7% specificity to identify need for neurological intervention
- There were 44 patients (1%) who needed neurosurgical intervention and all were picked up with the tool.
- The sensitivity and specificity of the overall rule (all 7 variables) were 98.4% and 49.6%.
- There were 254 patients (8%) were judged to have a clinically important brain injury. The tool identified 250 of the 254 cases. The four patients not identified with the tool were small contusions. None required neurosurgical treatment and none had neurological sequelae.
Ian Stiell and his team published another classic paper. They seemed to have worked their way up the body starting with the Ottawa Ankle Rules, moving to the Ottawa Knee Rules, the Canadian C-Spine Rules and finally the Canadian CT Head Rules.
This group from Ottawa was ahead of the curve on this topic. They started this project in the 1990’s when people were only beginning to talk about increased utilization of CT (cost) and the risk of radiation. Ian and his team were looking for ways to help EM docs choose wisely a decade before the choosing wisely campaign was initiated.
Their methods were outstanding, as you would expect from a group that has been putting out such classic papers. There was no selection bias, the population represented a broad spectrum of patients, it was a multi-site study, results were robust for sensitivity and the primary outcome was patient oriented. In addition, the tool they derived was simple with only 5 high-risk need for neurosurgical intervention items.
What about all the patients who did not have a CT scan? They represented 33% of the population. More than 1,000 patients in total did not get a CT.
All patients, CT or not, were assessed for the primary outcome of need for neurosurgical intervention. The five-high risk variables did not miss any of the 44 patients who had this primary outcome.
All patients who did not have CT underwent telephone assessment at 14-days post-injury, which classified patients as having no clinically important brain injury if had no or only mild headache, no memory or concentration problems, no seizure or focal motor problems, good performance on the Katzman Short Orientation-Memory-Concentration Test, and they had return to normal daily activities.
Telephone criteria had 100% sensitivity for identifying patients requiring neurological intervention and 87% sensitive for patients requiring clinically important brain injury.
The validity of these criteria was confirmed by applying these to a random sample of 172 study patients who had undergone CT.
What about the issue of intoxication either from drugs or alcohol? These were not part of the clinical decision tool.
Their data showed that examination of patients suspected of intoxication was not reliable or discriminating. In addition, blood alcohol level was not associated with important brain injuries. That is why they did not automatically scan patients with CGS of 13 or 14 but waited 2hrs to see if GSC increased to 15.
Not including alcohol as an indication for head CT is in contrast to the New Orleans’ Rule that we will discuss shortly.
Comment on Authors’ Conclusion Compared to SGEM Conclusion: We agree that the tool has good sensitivity, both with and without the medium-risk variables, and it is meant to rule in patients with concerning head injuries, rather than rule out those that don’t.
Overall, we feel it was a good quality study with a decent sample size. Their power calculation was to get 2,500 patients for a precision of 100% sensitivity for clinically important brain injury. The sensitivity of the high-risk criteria was 100%, which is good enough for me.
This rule was developed to insure that all injuries needing neurosurgical intervention were identified. This is a good, important, patient centered outcome. However, while the injury may not need neurosurgery, it may be important to know about other injuries as far as patient expectations for recovery. Also, in a different medico-legal environment, missing significant, albeit non-neursurgical lesions, may be important.
SGEM Bottom Line: The Canadian CT Head Rule is a simple clinical decision tool with 100% sensitivity to identify patients with head injuries in need for neurological intervention. Being a Level 4 derivation study it will need to be validated in large prospective studies with impact analysis demonstrating change in clinician behavior with benefit and ready for prime time.
New Orleans Criteria (NOC): Haydel MJ et al. Indications for computed tomography in patients with minor head injury. NEJM 2000
- Inclusion Criteria: Age >18, GCS 15 and Blunt head trauma occurring within previous 24hr causing LOC, amnesia, or disorientation
- Head CT is required if one or more of the following are present:
- Age >60yr
- Drug or Alcohol Intoxication
- Persistent anterograde amnesia (deficits in short-term memory)
- Visible trauma above the clavicles
- 1st phase: 520 patients presenting to the ED after closed head ingury with GCS 15 and non-focal neurological examination who all underwent CT Head. They used recursive partitioning to find criteria that defined those with CT abnormalities.
- 2nd phase: Sensitivity and specificity were evaluated prospectively in 909 patients.
- 36 (6.9%) of patients in the 1st phase of the study had positive CT scans. Seven criteria were found to be associated:
- Headache, vomiting, age over 60, drug or alcohol intoxication, deficits in short-term memory, physical evidence of trauma above the clavicles, and seizure.
- 57 (6.3%) of patients in the 2nd phase had positive CT scans. Sensitivity of the clinical decision rule was found to be 100%.
- 36 (6.9%) of patients in the 1st phase of the study had positive CT scans. Seven criteria were found to be associated:
- Sample size was smaller compared to the Canadian CT Head Rule, especially of the 1st phase used to identify the rule’s criteria (based on only 36 patients with positive CTs).
- The proportion of patients with positive CT scans was lower in the New Orleans study compared to the Canadian study (6.9% and 6.3% in the two phases of the New Orleans study versus 12% (8% clinically important and 4% clinically unimportant – in the Canadian study).
- Interestingly, the New Orleans Rule found intoxication to be associated with increased risk whereas the Canadian Rule found no significant correlation.
- The New Orleans Rule found seizure to be associated with higher risk of positive CT scan, whereas the Canadian Rule did not assess for this, given all patients who had a seizure prior to ED assessment were excluded and those who had had a seizure anytime prior to the 14-day follow-up were classified as having a potentially clinically important brain injury.
- Headache was found to be predictive of positive CT scan in the New Orleans Rule, however initial headache was not significantly correlated with a higher risk for clinically important brain injury in the Canadian Rule. However, persistent moderate or severe headache during the 14-day follow-up period did classify patients as having a potentially clinically important brain injury.
- The New Orleans Rule was developed to predict positive CT scans, but not necessarily those patients that would require hospital admission, neurosurgical intervention, or other clinically significant outcome measures.
- Papa et al (2012) compared both rules in patients with GCS 15 at a single U.S. Level 1 trauma center for outcomes of “any traumatic intracranial injury,” clinically important brain injury, and need for neurosurgical intervention.
- 431 patients were enrolled; 7% had traumatic injury on CT, 3.5% had clinically important brain injury, and 1% required neurosurgical intervention.
- Both the New Orleans and the Canadian rules had 100% sensitivity, but the Canadian Rule had a higher specificity for all three outcome measures (36.3 versus 10.2 to identify traumatic intracranial lesions on CT, 35 versus 9.9 for clinically important brain injuries, and 80.7 versus 9.6 to identify need for neurosurgical intervention).
- Smits et al (2005) compared both rules in Dutch patients with GCS 13-15 for the same 3 outcome measures.
- 3181 patients were enrolled; positive CT findings were present in 9.8% of patients and 0.5% required neurosurgical intervention.
- Both rules had 100% sensitivity to identify need for neurosurgical intervention. The New Orleans Rule had a higher sensitivity for identifying positive CT scan findings and clinically important injuries (97.7-99.4% versus 83.4-87.2% in the Canadian study).
- The Canadian Rule had higher specificities for all three outcome measures (37.2-39.7% versus 3.0-5.6% in the New Orleans study).
- They concluded the estimated potential reduction in CT scan ordering was 3.0% for their adapted New Orleans Rule versus 37.3% for their adapted Canadian Rule.
A recent smaller study by Kavalci et al was just published last year with 175 patients from a tertiary care center in Turkey. The CCHR had higher specificity, Positive Predictive Value and Negative Predictive Value for important clinical outcomes than does the NOC. (Free access open article)
SGEM Bottom Line: Both rules are highly sensitivity for positive CT findings and clinically important brain injuries, but the Canadian CT Head Tool had higher specificity and may be more clinically applicable given it is designed to predict clinically important brain injuries.
Case Resolution: Using the Canadian CT Head Tool, given the patient does not have any of the 5 high-risk criteria; you decide not to do a CT scan of their head. However, if you used the New Orleans Rule, she would have warranted herself a CT scan given her age was over 60 and she complained of a headache. You discuss it with the patient and her daughter, and opted not to do a CT scan.
Clinical Application: Swami can you put in what you do with this literature? Do you use the Canadian CT Head Tool, New Orleans Criteria, NEXUS or some other clinical decision aid?
A main issue with New Orleans is that the specificity is so low that what we see is increased testing when you use this rule. The Canadian CT rule has a considerably higher specificity and you only miss CT findings that don’t need neurosurgical intervention. These findings don’t really matter . . . except, possibly to a plaintiff’s lawyer.
The bottom line is that I think both rules can be used to help you establish your clinical reasoning early in training but I don’t know that either performs better than a seasoned clinicians evaluation and assessment.
What Do I Tell My Patient? We have done a good job of checking out your head injury. The good news is you do not need a CT scan of your head. There is a special tool called the Canadian CT Head Rule. It has been shown not to miss any patients with a head injury requiring neurosurgery that should have a CT scan.
The bad news is I think you have a concussion and you may experience a range of symptoms due to this injury. Here is some information on concussion and what to watch for in the next few days. If you are getting worse, have any of these new symptoms listed or are concerned please come back to the emergency department to be re-assessed.
Keener Kontest: Last weeks’ winner was Ben Brackett. He knew that the “red shift” occurs in astronomical terms, when the origin of light (a star for instance) is moving away from the observer. This causes a Doppler effect where the wavelength is slightly lengthened, causing a shift visible light towards the red end of the spectrum.
Listen to this weeks episode of the SGEM. If you know the answer to the keener/gunner question then send me your answer to TheSGEM@gmail.com with “keener” or “gunner” in the subject line. The first person with the correct answer will receive a cool skeptical prize.
Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.