Knocked out in New York

We’re at the other side of a busy weekend, we’ve had the usual Saturday afternoon and in to the night with falls from bikes, falls in the pub and what feels like a lot of CT scanning going on.

We’ve talked about head injuries and radiology this year and if you missed it earlier in the year we were lucky enough to have journal club held with the Sue Mason lead author of the AHEAD paper. Make sure you have a read our review and the linked FOAMed too!

Think about how you take your head injury histories. I always ask ‘did you lose consciousness/black out?’ I wonder does it make a difference about how I feel about the likelihood if serious brain injury?
If they say No, but they were beaten about the head with a bat?
Or they fell off their bike whilst not wearing a helmet chasing the downhill Strava segment they’ve been gunning for (the climbs up to our place are a great proving ground but have featured on the Tour de France, so we only have downhills left….)

Clinical Question: Should the presence or absence of a loss of consciousness (LOC) influence my decision to scan, and does that correlate to the presence of something significant on the scan?

This paper then, could help.

Title of the paper: Isolated LOC in head trauma associated with significant injury on brain CT scan
Published: International journal of Emergency Medicine. (online first 25/9/17)

Population: 13-35 year olds presenting with a head injury to a New York Level 1 trauma Centre.
Outcome: Clinically important CT findings (i.e requiring surgical intervention or >24hours in ICU)

This was a retrospective chart review of patients attending a trauma centre in New York between 2010 and 2013, which has an annual attendance of around 120K patients. This is a similar size to our unit, but as with all single centre studies it can be difficult to extrapolate their findings to your own patient population. I’m not sure if folks from the South Bronx are truly different to ours, i mean all human have hard skulls, squishy brains, and blood vessels that bleed but its still a consideration!

Inclusions: ‘Previously healthy’ 13-35 year olds who presented with blunt head injuries were included. They specifically wanted to look at people who were at the lower end of the risk spectrum so excluded anyone on anticoagulants or who had conditions that could affect the CT such as strokes.

Interestingly they excluded those who were intoxicated with drink or drugs.

To me this is interesting, largely as a big swathe of our young head injury patients have sustained their injury whilst under the influence of one or either or both. I can see why they did it, the key question around LOC can be hard to pin down when intoxicants are on board. And its hard to know whether the drink caused the LOC or the head injury. It just narrows the window of ‘who this paper applies to’ even further.

Given they’ve gone to such lengths to ensure they had a clear sight on the presence or absence of LOC then we have to reflect a little on the limitations of retrospective studies. In this case if the presence or absence of LOC wasn’t documented in the card then the patient wasn’t included. It’s not declared in the study how many patients were screened.

They captured 494 patient cards, with 37% of patients reporting a LOC (either themselves of the witness history), of these 8.1% had a positive CT scan. That’s 1 in 12 give or take. What about without LOC?Well 1.3% had a positive CT scan. That’s 1:100 positive CT scans in the patient group without TLOC. Which may sound like a reason to stop scanning these folks. When you look at those 95% Confidence limits though you get 6.73 for LOC Vs No LOC with limits spreading from 2.2-20.6. That’s broad, and reflects the sample size and the relative rarity of the event being studied. In total they had a positive CT rate of 3.8%. I havens looked at out data, but I hope we’re getting more positive scans for our buck than 3.8%. to be clear. I’m not wishing more harm on the fine folks of South Yorkshire, i just hope that we’re scanning fewer folks that have a normal scan…..make sense? (in the AHEAD study they had a 2.7% adverse event rate in those with no LOC and no concerning features)

They decided to further analyse (stated in the secondary outcomes) and found those with a GCS of 15 and no LOC there was a 0.7% postive CT rate, which again sounds impressive but remember the limitations of this study, and therefore the estimation of this event rate.
For me this paper raises some interesting questions but a combination of the retrospective nature, single centre an relatively small sample size for the outcome measure they were wanting to detect (and the associated broad confidence limits) mean its more piqued my interest to read more, rather than answered the question of how significant the LOC is in a patients care, whilst in this cohort there does seem to be a trend towards it being so.



Let me know how you feel about the more brief reviews.

Keep fighting the good fight.


Retrospective Studies and Chart Reviews

Leave a Reply

Your email address will not be published. Required fields are marked *