Guest Post: How do you decide that a child has sepsis (or not)?


This Guest Post is asking for your help – your thoughts – on the identification of an Unwell Child. Please – read on and link to the survey at the end ….

The introduction of Paediatric Sepsis 6 along with the recently released guidance notes has caused clinicians and organisations to look at the way that sepsis is identified and treated.
It is a problem that presents different challenges to different disciplines within medicine. A primary care or ED clinician must avoid over-diagnosis while at the same time referring a cohort of possible sepsis and presumptively starting treatment when sepsis is most likely. The secondary care paediatrician will reassess the referrals that may include paediatric sepsis and decide when to treat and when to observe or discharge. The paediatric intensivist will be involved for the most significant cases on admission. Those caring for inpatients will have to use a totally different threshold for considering sepsis, especially when the child is already ill or has compromised immunity.

Medicine has so far failed to find a test for paediatric sepsis. Guidelines point towards red flags but all of them ultimately require a clinician to make a decision.

Abnormal temperature and altered physiological parameters identify children who may have sepsis. Unfortunately fever is commonly accompanied by tachycardia even when a child has a self-limiting viral infection. Together without further evidence they are a poor indicator of sepsis. One of the effects that repeated observation of an association has on people is that they become de-sensitised to the information. This is like looking for a needle in a haystack. Even when I see something that looks like a needle, it turns out to be a piece of straw again. Eventually I stop expecting to see a needle, even though that was what I was looking for initially.

Box 1 – Some rough figures from the Sheffield Children’s Hospital (SCH) showing the patients who present directly to the Emergency Department (ED):

Number of self-referred patients attending SCH ED per year 50,000 Proportion of total
Number of ED patients with temperature above 38.5 1900 ~1 in 25
Number of ED patients with a possible diagnosis of septicaemia at time of transfer to ward 60 ~1 in 800
Number of ED patients with a final diagnosis of septicaemia at discharge from hospital 15 ~1 in 3500


How do we apply this to the hunt for the septic child, as we assess many, many children who are febrile and tachycardic but do not have sepsis? One way is to make the haystack smaller.
Paediatric sepsis 6 and the NICE febrile child guidelines both rely on a clinician to make a global assessment of the probability that a child has sepsis. These clinicians do not rely on the features of SIRS (Systemic Inflammatory Response Syndrome – see box 2) alone. If they did, there would be no need for a decision maker. The decision maker is also taking into account other factors such as risk variables, symptom progression and the effect of the illness on the child. Perversely, the first question for the decision maker is, “Can I exclude this child from the hunt for possible sepsis?”
Box 2 – Features of paediatric Systemic Inflammatory Response Syndrome (SIRS), taken from Wikipedia, September 2015

·        Heart rate is greater than 2 standard deviations above normal for age in the absence of stimuli such as pain and drug administration, or unexplained persistent elevation for greater than 30 minutes to 4 hours. In infants, also includes heart rate less than 10th percentile for age in the absence of vagal stimuli, beta-blockers, or congenital heart disease or unexplained persistent depression for greater than 30 minutes.

·        Body temperature obtained orally, rectally, from Foley catheter probe, or from central venous catheter probe less than 36 °C or greater than 38.5 °C. Temperature must be abnormal to qualify as SIRS in paediatric patients.

·        Respiratory rate greater than 2 standard deviations above normal for age or the requirement for mechanical ventilation not related to neuromuscular disease or the administration of anaesthesia.

·        White blood cell count elevated or depressed for age not related to chemotherapy, or greater than 10% bands plus other immature forms.

Removing the background noise of ill children who will not have a life-threatening infection is one of the most important steps in the hunt for sepsis. 

Knowing when a child does not have sepsis still leaves you with a cohort of children with possible sepsis though.  The smaller that cohort is, the easier it is to spot the true sepsis.  To illustrate the value of having features which thin the crowd, I turn to a popular children’s puzzle book: “Where’s Wally?”

When the page is first turned, there are hundreds of characters, most of which are clearly not Wally.  Occasionally there is a figure that looks like Wally but only partly.  Meanwhile, I haven’t found Wally because I am distracted by the sheer number of faces to screen. If I could limit my attention to just those half-dozen that look a bit like Wally, my screening would be much more efficient and I could find Wally far faster. (But it wouldn’t be as much use to distract a small child while cannulating.)

So, what are the features that help clinicians to decide that a child almost certainly does not have sepsis, even when febrile and possibly tachycardic?

I have listed a few variables above such as risk factors.  In the case of ruling out sepsis, the absence of risk factors (such as immunocompromise) is important.  One dimension of the global assessment made by the clinician is the child’s behaviour and activity.  A child’s behaviour while ill seems to be a significant part of decision making for clinicians.  Research into determinants sometimes falls into the trap of seeing this as an unquantifiable and unreproducible gestalt assessment.  In fact, many of the features that can be considered are binary and entirely measurable.  Did the child walk or talk when assessed?  Did they run or play?  Often, these activities are strong factors in our decision to clinically exclude sepsis.

So the question is “what activities and behaviours reassure you that a febrile child is very unlikely to have sepsis?”  Perhaps we are all thinking the same things are reassuring and perhaps not.  What I would like is to get some of you decision makers to take part in a modified Delphi which would help to answer that question.  Linked below is a survey which will take just a couple of minutes of your time.  If you have to answer the question “Could this child have sepsis?” in your day to day clinical work, I would very much appreciate your input into this so that we can learn more about what matters when deciding who we are not so worried about.


(Round was was at )

Thank you

Edward Snelson


Consultant, Paediatric Emergency Department

Sheffield Children’s Hospital, UK


(Visited 114 times, 1 visits today)