Sunday 4 March 2018

Paediatric Sepsis - the Facts, the Myths, How We Got Here and Where We Need to Go Next

You may be wondering if you're the only one confused about what's happening with paediatric sepsis.  You're not.  I work in paediatric emergency medicine and I just about get it.  Just.

I am frequently asked questions about what people (in both Primary and Secondary Care) should be doing now with regards to sepsis.  Everything seems to be changing so quickly that it is difficult to keep up.  As intelligent clinicians, we like to understand the reasons for change as well as what changes are occurring.  In the case of sepsis, research evidence is only part of the explanation for the changes that are happening.  That doesn't seem very reasonable but there are valid reasons.  As you're the one having to change your practice, it's probably time that you received a full and honest explanation for why you should do this in the absence of robust evidence.

Sepsis has always been with us.  We might have changed but sepsis has always been the same - a dysfunctional response to infection, manifesting as an exaggerated response to the illness or as organ dysfunction.

Some things have changed in our understanding of sepsis, while other aspects remain frustratingly beyond comprehension.  What we know and what we don't know are key elements to understanding the evolving approach to paediatric sepsis.

What we know:
  • Sepsis carries a high morbidity and mortality
  • Early treatment with antibiotics reduces morbidity and mortality
  • Undertreatment of sepsis increases morbidity and mortality
That's about it for what we know.  It's a short list.

What we don't know:
  • We don't have a good definition of sepsis in children.  The Third International Consensus Definitions for Sepsis and Septic Shock (1) gave a definition of "life-threatening organ dysfunction caused by a dysregulated host response to infection."  This is in many ways spot on but at the same time suitably subjective so as not to be specific when it comes to a clinical decision.  The other commonly used definition of SIRS (systemic inflammatory response syndrome) is very specific (if you have an abnormal temperature and tachycardia, you fulfil the SIRS criteria) and has the opposite problem.  It is very prescriptive and very often wrong.  Hot, tachycardic children are not hard to come by.  A few of these will have sepsis but most will not.
  • We don't have a good way of deciding if a child has sepsis.  Various approaches have been tried.  Decision tools have always been hampered by poor sensitivity and specificity.  No blood test or combination of blood tests has been found to be reliable either in the ruling in or ruling out of sepsis in children.  If there was a reliable method we would all be using it. Telling the difference between a febrile viral child and a child with early sepsis is as much of a challenge now as it was in years past.
Image taken from http://rolobotrambles.com/ with kind permission from Damian Roland.
  • We don't know the impact of the various paediatric decision tools that have been introduced in recent years.  In an attempt to improve the recognition of sepsis, there have been many campaigns, guidelines and decision tools landing in your inbox over the past few years.  Driven by the fact that sepsis is often diagnosed late and that the consequences can be fatal, these initiatives inevitably push for increased diagnosis, using a lower threshold. Increased diagnosis can only be through a lower threshold at the moment, since we have failed to find a way to improve sensitivity and specificity together.  What nobody knows is whether this move to a lower threshold is having a positive or a negative effect.  If for every case that is diagnosed earlier, 1000 have unnecessary treatment (totally made up numbers of course), is the overall effect good or bad?
  • We don't know how many cases of sepsis can be diagnosed earlier.  The main problem of course is that retrospective analysis of cases of paediatric sepsis are inherently flawed.  The outcome is known, therefore all tachycardia becomes evidence of sepsis, despite the fact that other causes are possible.  Most children who present with sepsis will have had one or more contacts with a clinician in the days preceding their diagnosis of sepsis.  In some cases the symptoms will have been due to a viral infection which preceded the sepsis and in other cases the symptoms may have been due to the developing sepsis.  It is rarely possible to be certain which of these was the case.

It is therefore with a large number of uncertainties that we go forward.  There is no doubt that we are getting better at treating sepsis, but recognising it remains an area that we have a serious amount of work to do.

So why have we changed practice in the absence of evidence?  The answer is that, although we are all coming at the problem from a different point of view and with different agendas, we all want to do this better.  With the evidence pointing towards a benefit from early recognition and more aggressive treatment, changes are afoot to try to achieve these outcomes.  We need to alter our perspective and awareness of the problem of sepsis even if we do not yet have the answer to the question of how to better recognise it in children.

How should I change my practice?

Taking Damian Roland's sepsis spectrum and simplifying it, there are three clinical scenarios that I think are generally true by the end of consultation with an ill child.

Scenario 1 - the child with a febrile illness who has demonstrated their wellness to the point that sepsis is extremely unlikely

For these children, we should no longer be talking in terms of them having 'just a virus.'  They have a virus and they have no signs of sepsis, however, they are now in a group that is at risk of developing sepsis.  Worse still, if sepsis should develop, the signs might not be recognised as they will be attributed to the viral illness.

What should now be standard practice is a safety-netting conversation at the close of the consultation which makes clear what the red flags are and when to seek further assessment or advice.  Parents should feel that they are empowered to seek this without fear of being seen as over anxious people.

Scenario 3 - the child who appears to be seriously unwell

If you work outside of a centre which has acute paediatrics, get them there.  If you are the person at the other end, make sure that the child is seen quickly, a decision is made to treat ASAP and that there is no delay to treatment with a full Sepsis 6 bundle as appropriate.  These children might have had multiple cannulation attempts over a long period of time in the past, but now it's time to get out that drill.

Scenario 2 - all the children in between 1 and 3

In between those two set pieces is where we make the big decisions and make them very carefully.  As mentioned above, there is no good decision tool available, no test and no easy answer.  Decision making in these circumstances is complicated (2) and the explanation of how these decisions are made cannot be oversimplified.

What can be simplified are the guiding principle for assessing these children who occupy the uncertain zone.
  • Gut feel is an important part of the decision making process.  This means that you should listen to your gut feel and to the anxieties of the parents and your colleagues.
  • Persistence of abnormal signs is concerning.  While signs (heart rate, level of activity) fluctuate significantly during an uncomplicated viral illness, children with a significant infection tend to be persistently unwell.  Persistent tachycardia, tachypnoea, lethargy or other signs of illness should be taken very seriously.
  • No one factor will give you the answer.  Making these decisions would be a lot easier if there was a solid rule.  Can you send someone home with a pyrexia? Often.  Can you send someone home with tachycardia? Sometimes.  
  • You can only work with the information that you have.  What do I do if it is 2am, the child is asleep and is suitably grumpy when woken?  In this situation, the doctor-patient banter that I rely on to rule out sepsis is not available to me, so I am forced to look at all the other information that I have.   Admittedly, I look at it very carefully before making a decision.  I can only work with what I've got though and I'm not going to keep the child till morning just to tick the "I've seen them smile and run around" box.
  • The sepsis assessment never ends.  If the initial assessment leaves uncertainty, there are several options including referral, observation and discharge, depending on the degree of suspicion.  In all of these cases, we need to be clear about the need to take persistent, worsening or new symptoms seriously. 

Early sepsis is difficult to detect, so we need to use every tool available to us to make sure that the diagnosis is made at the earliest possible opportunity.  Until we have something better, creating that opportunity means as many clinical assessments as needed depending on the scenario.

What has changed over the past few years?  Hopefully we have.  We now better recognise the importance of early recognition of and immediate treatment of sepsis.  What we shouldn't forget is that while we have set rules for ourselves that achieve the latter, we haven't yet found a way to consistently do the former.  So that leaves us with a need to do what we can to make the best clinical decision possible for each of our patients, while being patient with the academics who will hopefully one day come up with a magic test.

Edward Snelson
Medical Historian or Historical Medic?
@sailordoctor

Disclaimer - very little research into the actual history of sepsis went into this article.  By little, I may mean none.  Still, the leeches bit is probably true.

References
  1. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3), JAMA, February 23, 2016
  2. Roland D, Snelson E ‘So why didn’t you think this baby was ill?’ Decision-making in acute paediatrics Archives of Disease in Childhood - Education and Practice Published Online First: 01 March 2018. doi: 10.1136/archdischild-2017-313199