Imagine yourself in the resus room. You’ve just got Return of Spontaneous Circulation (ROSC) on a patient who was in cardiac arrest. They had excellent bystander CPR and an early shock from an Automatic External Defib. The patient has been packaged and rapidly transported by paramedics to the ED where they are reasonably stable. You’re planning to get the patient to the cath lab (as you know this is the right thing to do) and as part of that process you are getting some lines in, exchanging the Laryngeal Mask Airway for an Endo-Tracheal tube and sorting out some sedation.
At the head end of the bed the airway doc pipes up….
‘Pupils are fixed, this is not going to end well’
You can feel the mood in the team drop at this news. The positive feeling about a patient with ROSC is dented by the news about the pupils, but is this right? It’s ingrained in us that a lack of pupillary response is a bad prognostic sign as it is a feature of brain death, but it’s not as clear cut as we once thought. We have previously looked at the prognostic influence of pupillary light responses (PLR) in trauma following Mark Wilson et al’s great paper on fixed dilated pupils in head injury1. That paper showed that PLR did not always predict a terrible outcome and that in fact many patients with an extradural and fixed pupils can have an excellent outcome.
In cardiac arrest patients the picture is unclear. There are plenty of papers out there that have looked at PLR as a prognostic factor but they are all over the place in terms of patient population and findings2–7 . The dogma still exists in clinical practice that it’s a terrible findign, but we don’t really know and there are perhaps many reasons why a patient might have fixed pupils but still have a positive outcome. Pupils may be dilated as a result of medications, toxins, drugs or environment, and it’s worth noting that Adrenaline causes dilatation and we do use that a lot in ALS protocols (for now at least). The bottom line is that the picture is unclear and we need to know more before we start altering our management of patients based on what we think we know, but don’t really know.
This week we review a paper that might help us answer the question. A French paper in Resuscitation looks at the PLR as a factor in the outcome from over 11,000 cardiac arrests8. The abstract is below, but as always we recommend you read the full paper.
What kind of paper is this?
It’s a retrospective cohort paper. The data used in the analysis is routinely collected as part of an ongoing project looking at the management of cardiac arrests. This is great in some respects as it’s routine, well established and did not need to be set up as a separate project. The authors had access to lots of data and that meant that the time from conception of the study to analysis to publication is shorter than if this was set up as a new study. That’s great for the researchers, but retrospective studies have their problems. Missing data, reporting and recording issues and a lack of calibration and focus on the study question are all problems with retrospective cohorts and the authors did find this to a degree. On the other hand cardiac arrest is not that common, and survival from cardiac arrest even less so. You need a big database to get decent sample size for analysis and that would probably not happen if you wanted to set up a study just to look at this question. The bottom line is that this is not a perfect design but it is appropriate and reasonable in the circumstances. You can hear more about cohort studies on our podcast here.9
Who was studied.
This is a French study with data collected between 2011 and 2017. The French system is different to the UK and that makes a difference here. The patients in the cohort are those attended by a mobile medical team that contains an emergency physician. This means that the vast majority of the 73,000 plus patients in the database are declared dead at scene and not transported to hospital. Those in this study are the ones who get ROSC and are brought to the ED. So not really the same as UK practice and we need to be mindful of that. These patients are ROSC patients and not those still in cardiac arrest. For me this means that it’s the group who are brought to hospital with an output or those who arrest in ED and who get an output back.
Patients were excluded if the PLR, GCS or outcome data was unavailable. We should also note that the age ranges are not the same as UK. In this study they only looked at adults as defined as over 18. In truth the difference between that cohort and the UK definition of over 16 is likely to be irrelevant as there are so few cardiac arrests in that differential age grouping.
What were the outcomes?
The outcome in the study was the Cerebral Performance Score (CPC) at 30 days10. The CPC rates outcome on a 5 point scale with 1 being pretty much normal through to 5 which is brain dead. In this study a score of 1 or 2 was considered a positive result, which is consistent with other studies in this area.
They also looked at other factors that influence outcome in cardiac arrest such as age, type of arrest, intial rhythm etc. They used that data in a multivariate model to see if PLR was an independent risk factor. Do remember that this study is about EARLY PLR and not late PLR when you might be assessing prognosis at day 3. We’ve talked recently about prognostication in other conditions such as neurological injury and it’s generally unwise to do it too early11.
What are the main results?
The headline figure is that PLR is a bit rubbish as a single test. The sensitibity for a positive outcome was 72.2% and the specificity was 68.8%. Neither figure is good enough to base a clinical decision on and it would be easy to simply conclude that we should abandon it, but as ever it’s a bit more complicated than that.
The multivariate analysis showed that PLR is an independent predictor of outcome and quite an influential one too with an Odds Ratio of 3. This is important as it tells us that it is a sign that is independently associated with outcome. As clinicians we rarely get a single sign to tell us what to do with a patient, we look at a wider picture and use that to make a judgement. In this case it would seem that PLR is a factor in outcome that we could use in balance with other factors such as age, rhythm, co-morbidity to help guide our judgement.
In other words if we took PLR as a single test then many patients with an absent PLR will get a positive outcome, but who ever looks at a single test?
The authors also looked at whether drugs like adrenaline and noradrenaline alter PLR and the answer is that they do. However, that factor was not associated with a better prognosis as you might think. I suspect this may be because Adrenaline use will be associated with more prolonged resuscitation attempts which as we know have a worse prognosis (Ed – one shock on an AED and no drugs for me please).
Will this change my practice?
I started reading this paper in the belief that PLRs had no place in assessment for the reasons stated above. Now I’m not so sure. It’s clear that the single finding is not enough to guide my clinical treatment and I will continue to manage my ROSC patients as best I can without this influencing how we do things in the resus room. However, it might change the emphasis and way in which I explain likely prognosis to families in the relatives room.
I think it might change some of my more difficult interactions with specialities though. I’ve often been challenged about patients going to cath lab or ITU on the basis of pupillary responses where that particular finding appears to hold more sway than others (even though we know that we should12). I think this paper will help me challenge those beliefs and ensure that we get a more rounded assessment of the patient.
So what would I say to my airway doc friend at the head of the bed? I think it might go something like this….
Her: ‘Pupils are fixed, this is not going to end well’
Me: ‘You’re probably right, but then cardiac arrest is never a great event and we are not starting from a great position. As for the pupils then sure that’s not a great sign but there are many people who have a really great outcome who come in with them fixed. Let’s crack on and do the best we can for this patient here and we can revisit prognosis when we have a bit more time and info.’
I’d like to think it would go like that (Ed – in the real world it probably wouldn’t be quite so calm 🙂 ).
So this is not quite the #dogmalysis I had hoped for but useful none the less. Let me know if it changes your practice.
Before you go please don’t forget to…
- Subscribe to the blog (look top right for the link)
- Subscribe to our PODCAST on iTunes
- Follow us on twitter @stemlyns
- PLEASE Like us on Facebook
- Find out more about the St.Emlyn’s team
- Come join us at our conference in October 2018