By Ian Mitchell, MD, FRCP (EM)
@travels2little
travels2much@gmail.com

Emergency Physician, Royal Inland Hospital
Associate Clinical Professor, UBC Dept. of Emergency Medicine.
Site Scholar for EBM - Kamloops Family Medicine Residency Program

18 year old girl who spun out of control and ended up in the ditch. There was significant vehicle damage and a prolonged extrication time. A level 2 trauma alert was called, but cancelled when patient arrived stable in a C-spine collar, texting above her head. She complains of neck pain and a headache. No LOC.

It is important to distinguish between patient oriented outcomes, such as radiation, morbidity and mortality and in this setting, neurosurgical intervention; from disease oriented outcomes, such as abnormalities on the CT scan.

Note the two distinct outcomes. This is something residents often fail to appreciate. The criteria that predict a change in outcome are the first five, the “high risk criteria” and they are relatively objective.

The problem and increase in scan numbers comes from adding in the “medium risk” category, which adds “dangerous mechanism” to the mix, a much more subjective criteria. Adding these criteria significantly increases the number of CT scans performed and predicts only radiological signs of brain injury rather than clinical outcome. While this is critical in the litiginous US environment, it is less important in other countries and needs to be balanced by consideration for radiation exposure issues and patient-oriented outcomes.

It is also important to note that GCS of 15 is only required by two hours, meaning concussed individuals can be observed for improvement before a deferred scan.

Not all injuries visible on CT scan (disease oriented outcome) lead to neurosurgical intervention (drainage, ICP monitoring), or even admission.

As the degree of badness of a CT goes up the rate of intervention rises, once it begins to be associated with increased morbidity and mortality (patient oriented outcomes), until falling off again when surgical intervention becomes futile.

While normal CT head scans in trauma are unavoidable, we should look at our yield similar to the way David Newman has talked about "the 700 Club" with respect to subarachnoid hemorrhage and lumbar puncture.

Even within departments there is a wide variation of CT scanning practice. This is one of the failings and strengths of CT Head clinical decision rules. By identifying patients at a low risk for injury, they can help the physician who orders more CTs than average. However, for a physician (or population) who orders fewer CT scans than average, CDRs can actually lead to increased scans.

Variance can also arise due to resource availability - ie, Whistler

Emergency Department Computed Tomography Utilization in the United States and Canada, was published last month in Annals of Emergency Medicine and has been discussed thoroughly on Academic Life in Emergency Medicine as part of their on line journal club. Their conclusions:
CT was more readily available in US EDs, and US clinicians used the technology more frequently than their colleagues in Ontario for nearly every category of patients, including children. CT utilization increased over time in both jurisdictions, but faster in the United States
The only class of CT scan performed that did not increase was CT scans of the head in children, and that only in Canada. Whats up with that?
From a more Canadian perspective, in large centers we use CT regularly, but how about if a patient meets criteria for scan, but has to be transported for that to happen?

Head Injuries in the Rural Setting - What is the role of the Canadian CT Head Guidelines?

This is an interesting UBC thesis from an M. Sc. researcher in Human Kinetics, Fern Von Der Porten. She looked at the data from all of the patients who presented to the Whistler Health Care Center in 2004 (before they had a CT scanner) with head injury or trauma and applied CCHR criteria to them.

Of the 211 included charts, 51 had CT indicated, and only 4 of these were transferred to a health care facility with CT scan available. A further 9 patients, without meeting the CT criteria were also transferred.

While there are many hazards in retrospectively looking at data of this type, it is likely the case that many patients in rural Canadian EDs do not get transferred for scan. This is probably as it should be, for reasons of resource utilization and transport issues, but it does beg the question of whether outcomes suffer?

I'm not talking about cell phone or smart meter radiation, but what are the real risks of CTs?

How good are we at assessing radiological risk?
How do you clear a trauma room?


"A new study offers the most solid evidence to date that radiation from CT scans increases children's risk of developing leukemia and head and neck cancer.

Children and adolescents who received two to three computed tomography scans of the head were three times as likely to develop brain cancer as those in the general population, according to the study of 176,587 children" WSJ, 2012

There have been concerns for years about cell phone radiation and brain tumors that is probably unfounded, and scare stories about wifi radiation andsmart meters continue to appear in the popular press.

Even within our environment, the best way to clear a trauma bay is to yell out “Xray”. As health care workers flee to the safety of distance and solid objects (or behind heftier coworkers), we are fleeing an estimated 0.2 microgray.

Meanwhile, a CT head scan comes in at around 10 000 times as much radiation.

These two risks obviously have to be taken into context. Hospital staff should avoid radiation and follow ALARA principles. In the same way, we should see to our patients best interests and decrease radiation exposure.

From the physician’s perspective, the risk lies more in malpractice. This is the general argument in the US for not using CT head rules at all and just scanning everyone. Malpractice risk is immediate and real, while radiation risk is distant and unseen.

Cancer is a stochastic risk - not all with the exposure will get the disease.

You can look at each scan as coming with a lottery ticket for cancer. What are the odds in this lottery? Higher than we thought:

Most recently, the Matthewset al Australian data linkage study, with an enormous cohort (11 million) showed that the adjusted overall cancer incidence
for young people exposed to a CT scan was 24% greater than for those who were not exposed.

That is, one in every 1800 scans resulted in an excess cancer case.

The relative risk was higher for many different types of cancers and exhibited a dose-response relationship. This is particularly striking as the mean follow uptime was 9.5 years, suggesting the true lifetime risk may be much higher. The
researchers assumed a typical radiation dose of 40 mGy per scan. They concluded that the increased cancer incidence was mostly due to radiation from CT scans."

- from "Ionising radiation: three game-changing studies for imaging in sports medicine"

AND

"Appropriate CT scan use in trauma evaluation is challenging. It’s use is widespread, and although it changes management it has not decreased trauma mortality. This paper shows that the risk of death from trauma in the elderly outweighs the risk of death from CT scan radiation. However, this gap narrows in younger patients with less serious injuries because of their very low mortality rates. Therefore, we need to focus our efforts to reduce radiation exposure on our young patients with minor injuries." Ryan Radecki, on "Comparison of trauma mortality and estimated cancer mortality from computed tomography during initial evaluation of intermediate-risk trauma patients"

The Swedish study was really the first to connect long term cognitive damage with ionizing radiation in childhood. When you look at the data from the CT utilization study, the only decline in CT scans over the time period studied (2003-2008) was in pediatric head CT scans. It was this article that started to chill our enthusiasm for scanning the heads of children. The cancer risk was long term and nebulous, but cognitive changes made one think twice about scanning the little ones.

We avoid scanning kids because their brains are still developing and the risk of cognitive and cancer risks. While cancer risks decrease with age, at what stage do our brains stop growing and changing? Has yours?

Both cognitive and neurodegenerative risks are likely to be deterministic rather than stochastic. A bit like adding layers of film over your windshield (not a bad analogy, given the link between radiation exposure and cataract formation)

Radiat Environ Biophys. 2013 Mar;52(1):5-16. doi: 10.1007/s00411-012-0436-7. Epub 2012 Oct 26.
Long-term effects of ionising radiation on the brain: cause for concern?
Kempf SJ, Azimzadeh O, Atkinson MJ, Tapio S.

The authors try to establish a link between ionizing radiation and neurodegeneration, which may never be possible given the time delay to Alzheimer’s.

"…around 20 % of the global annual per capita effective radiation dose comes from diagnostic medical and dental radiation regarding the years from 1997 to 2007"

"A proposed model for the role of low- and moderate-dose ionising radiation in neurodegeneration is presented. The model suggests that mitochondria play a central role in the radiation response followed by neuroinflammation and oxidative stress. Subsequent cellular effects lie in the reduction in neurogenesis and cerebrovasculature followed by neurodegeneration"

What is the role of observation as an intermediary between discharge and CT scan? While we tend to see observation as a safe approach, being in contact with the hospital milieu is not always benign, with over 400 000 preventable hospital deaths per year in the US.

Which rule is best?

Will a rule help? Do we get better at making decisions with less information?

These slides are from Ian Stiell's lecture at CAEP 2012.

The three stages of CDRs
1. Derivation
2. Validation
3. Implementation

Impact analysis is designed to measure effect in the real world

No decrease in scans was found when implementing this tool. Scans actually increased during the time of the implementation

Although the authors of the systematic review estimated that the PECARN rules would require approximately 50 CT scans to identify 1 child with clinically significant intracranial injury and more than 200 to identify a neurosurgical injury, it should be noted that the published PECARN algorithm does not recommend that a head CT be performed on all children that do not meet the low risk criteria.

Rather, observation is specifically encouraged as an option for the clinician to consider based on other clinical factors.




this is a follow-on study attempting to determine whether the severe mechanism portion of the decision instrument was predictive of significant TBI, or whether scans could be avoided if mechanism was the only positive feature in their decision instrument. And, yes, a severe injury mechanism in isolation - at least in the 35% of their cohort who received a head CT - had only a 0.3% chance of significant injury in age 2 years. Severe injury mechanisms associated with additional PECARN criteria, however, had 4% and 6% incidence of TBI, depending on age.

Probably the most important aspect of these numbers is they allow for a better discussion of risks with parents and families. While 1 in 150 or 1 in 300 sound like pretty good odds, when you practice long enough, those odds will catch up with you. Even with severe mechanism and additional features, 19 of 20 CTs will be negative - you can still make a reasonable case for observation rather than knee-jerk scanning.

observing children in the ED for a short period, rather than making an immediate decision regarding CT use, resulted in decreased use of CT. Thusly, the press releases state “Waiting and Watching Can Reduce Use of Brain Scans for Kids in the Emergency Department”.

But, watching and waiting doesn’t benefit the children in this cohort – other than preventing avoidable harms. The eight children who had CT scans showing clinically important injuries were easily identified by clinicians as requiring immediate CT. The period of observation doesn’t change the short-term clinical outcome of any of the patients – it only “treats” the risk-aversion of clinicians and parents. ”Watching and waiting” may reduce scans – but discharging the entire observation cohort immediately would have reduced scans even further, without missed cTBI (although the study is underpowered to truly detect all events down to an appropriate “zero-miss” threshold).

Be cautious with the pan scan in the patient who is GCS 15, particularly in the young, but consider for all.

Consider the patient oriented outcome.

Some restrictions apply...

These recent studies raise the question of whether the risks are now established to the point that the patient needs to be specifically informed of them at the time of referral, scan or both.

The existing evidence of a lack of public understanding about radiation doses suggests that patients need additional education about risks compared to benefits in order to give informed consent.

...If a patient is able to understand and give informed consent and has no focal neurological defects, a CT scan of the head is unlikely to yield any information that will lead to a change in outcome. Is it still worth the harms?

Kamloops CT Head Tool for Trauma: Texting indicates an advanced level of cognitive function, that is unlikely to be improved by neurosurgical intervention. In the absence of focal neurological signs or coagulopathy, consider deferring or not performing a CT head scan on the texting trauma patient.

How else to cut down on CTs?

New pediatric trauma forms - cut the Pan Scan?