This case is inspired by a recent Twitter discussion with fellow Pads ED enthusiasts – Tessa Davis [ @TessaRDavis ] , Andrew Tagg [ @andrewjtagg ] , Rachael Rowlands [ @rachrwlnds ] and whoever else was reading.
As a background – you probably should read my previous ramblings about Paediatric head injury assessments here: Kids Coconuts and CTs
I think that the PECARN decision tool [MdCalc clickable version HERE ] is probably the most useful and robust data out there for making the call on children with minor head injuries in the ED. In the past we have often taught or been taught to observe kids for hours and hours, or CT them. The PECARN data set certainly has changed my practice in the last 5 years.
I now use it to identify children whom are at extremely low risk and feel confident to discharge them with follow-up instructions of course , one needs to be assured that the parents / carers are comfortable with this and able to access appropriate follow-up if required. The PECARN study also gives us a basis for discussing the risks of observation, imaging and non-investigation in kids who are not in the low-risk group.
Now take a look at this part of the PECARN Algorithm – this is the section for kids aged under 2 years of age.
The majority of kids that I see in this group are little ones who have walked / run / fallen head first into something solid. They usually have a frontal “egg” or laceration over the forehead.
So lets walk through the algorithm. If this kid has a large frontal haematoma then it is going to be tough to say that they do not have a skull fracture, or at least you may think you can feel something. Let’s face it – pushing on a fresh, boggy swelling over a kids head is just plain cruel!
If you think that you can palpate a fracture – then they immediatley jump into the ‘high risk’ group i.e. the group that is recommended to CT early. This was about 1 in 8 of the kids in this age group.
However, if there is no fracture palpable then you are very likely heading down the pathway to simple observation, or possibly into the super low-risk group.
Hence the question of the presence or absence of a skull fracture seems to be a big hinge-point when it comes to making this decision. And here is what I suspect: we are terrible at picking these! OK, I understand that the PECARN trial is a pragmatic set-up. It was a simple clinical call – was the clinician able to palpate a fracture. However, in reality this does seem very subjective and prone to bias.
Enter our friend – the Ultrasound. Is there nothing we cannot do with the probe?
Ultrasound is useful in picking fractures elsewhere in the body – particularly in superficial bones. So can we detect skull fractures with any accuracy?
Well there are a few smallish trials looking gat this question in the literature:
One form New York, Jim Tsung and co. in Paediatrics, June 2013 - “Accuracy of Point-of-Care Ultrasound for Diagnosis of Skull Fractures in Children”
Another from Riera et al in Paediatric Emergency Care, May 2012 – “Ultrasound Evaluation of Skull Fractures in Children: A Feasibility Study” This one had more small kids in it – i.e.. mean age 2 years
These are small trials – 115 patients combined. 19 had a fracture on CT – so the incidence was ~ 16%. Like many POCUS papers they show the usual characteristics of bedside ultrasound techniques – very specific and reasonably sensitive – so if you see a fracture – there almost certainly is a fracture, but you will miss 15 – 20 % potentially based on the numbers here. In the Tsung paper – only one fracture was missed – that was a fracture that was adjacent to the hematoma – rather than directly beneath it. So you could improve this with a more thorough scan field. In terms of likelihood ratios: bedside US give a +LR of 27 (excellent!) and a -LR of 0.13 (pretty good!). As always – we need more data to validate this and make it more generally applicable.
So how would a bedside skull US fit into the PECARN pie?
Hard to say what our “Sensitivity and specificity” is for clinical palpation of skull fractures – I would guess we are 50% sensitive and 75% specific.
So using that as a comparison – we should be able rule in a few more kids who should probably get an early CT.
So will this mean we do more CTs? I think not – as there are a lot of kids who have a nasty looking egg on the head – and we are often biased by the external picture into believing that we can feel a ‘step’ as we are worried and want to rule out badness. So if we scan the kids with the ugly looking hematoma and find no fracture on US – then this would probably be a group where it is safe to push them into the “observe, wait ‘n see” strategy. Although US is not super sensitive – it surely must be better than a subjective prod over a boggy lump.
[Note: if you use a stand-off pad or lots of gel – you can do this US without inflicting much pain at all. No pressure needs to be applied. So I think this is a more humane approach to the kids with a large, boggy swelling of the noggin.]
So overall I think that US would allow us to separate the goats from the sheep – allow us to create a bit of diagnostic daylight between these 2 groups:
(1) Those who definitely have a fracture and may need early imaging of the brain
(2) Those at lower risk of fracture who can be safely observed.
Bedside US probably doesn’t add too much to the kids whom can be classified as very low risk by the PECARN algorithm.
My practice is to promptly discharge kids who meet the “extremely low risk” criteria. I give the parents reassurance and information for what to look out for. But… this is my new pet peeve… forcing a kid / family to sit in the ED for 4 – 6 hours to have “Neuro obs” completed seems like a really antisocial and low-yielding exercise in this group. So if I am satisfied that the parents understand the risk and what to do if… happens, and there is no NAI question – I will send them home from triage.
Love to hear how you manage this common problem in your ED.