Author: Nikki Abela Editor: Liz Herrieven / Codes: GC3, RP7, SLO3, SLO5 / Published: 17/06/2019


The content you’re about to read or listen to is at least two years old, which means evidence and guidelines may have changed since it was originally published. This content item won’t be edited but there will be a newer version published if warranted. Check the new publications and curriculum map for updates

This blog is the narrative of the author’s talk at #DFTB19

Warning: This blog is not talking about assessment of children who come from developing countries as I unfortunately do not have enough experience/expertise to be able to talk about that.

When I was asked to talk about assessment of (de)hydration in children, about 6 months ago now, I accepted without a second thought. “How hard could it be?”, I guessed, “I do this every single day”.

And, of course, we do do it every single day, but are we getting it right?

I’m not going to talk about scoring systems, because pragmatically these are based on very subjective assessments like “skin turgor” and “sunken eyes” (after some night shifts I’m sure some airlines would like to charge me for the extra baggage under my eyes) and have been proven to not change the variability of assessment compared to physicians’ more conventional methods.

In fact, clinicians are getting it right 25-81% of the time. It probably won’t surprise you to find out that seniority makes you more likely to think the child’s dehydration is not so severe, and being more junior will make you more likely to overestimate the severity of dehydration.

I don’t know about you, but I thought 25-81% right is a bit of a guess really.

What is it that we are looking at anyway? The best way I found of classifying the signs and symptoms we look out for is by dividing it into 3:

  1. Dessication of tissue (e.g. skin turgor, dry mucous membranes)
  2. Compensatory mechanisms to maintain perfusion (e.g. tachycardia, tachypnoea)
  3. Combination of both (e.g. prolonged CRT)

We have all seen those tables asking us to rate the percentage of dehydration based on a combination of the above signs, but what does the evidence say?

Studies have found that most of the things we look at in our assessment of dehydration are imprecise, but if we did have to look at clinical signs, CRT, skin turgor and respiratory rate were the most important. They all have an average (50-60%) sensitivity to rule out disease if negative (which doesn’t make them very good), but they score better for ruling in disease if positive, with specificities of about 80-85%.

The problem with these studies though is that the gold standard assessment is the surrogate marker of percentage weight loss based on pre-illness weight, (and we know this is correct because early studies used muscle biopsies (yes muscle biopsies) in children to prove plasma volume loss correlated with percentage weight loss). I’m sure you may have guessed by now, a pre-illness weight isn’t usually readily available for those presenting in the ED. (Here is the formula if today is your lucky day: percent dehydration = (baseline weight minus admission weight) / baseline weight x 100). Moreover, the studies (especially assessing scales) are heterogenous, involve rather small populations, and haven’t had much external validation.

Although I must say I’m a bit of a sucker for evidence, when it comes to something pragmatic like (de)hydration assessment, I’m inclined to think that PEM people, who see patients day in day out, have a knack to telling whether a child needs further investigation (even if that is good old observation) or treatment.

So I asked a host of PEM people I hold in high regard, and they had a few disagreements on some points, but this is what they all agreed on:

But they all said it was hard to describe, so my job was to try and tease out of all of them (and my subconscious), what this look entailed.

The first important thing to note is that real dehydration in a healthy child in a developed country is rare – if you remember this you will be biased and miss that one child, so please forget I told you that. However, remember that your seeing a whole host of well children in the PED/ED/GP practise where you work means that you probably know what a well child looks like. The dehydrated child looks different to that.

Look at their vital signs, a normal heart rate and respiratory rate make dehydration less likely.

Ask about oral intake: I find that if a child is eating, (food) they are unlikely to be dehydrated. Reports of decreased urine output are concerning, but need to be viewed holistically.

Look (before you approach too), is the child active and playing, are they drooling? This makes dehydration less likely. Are they weak, disinterested and quiet? Start to worry.

Look at their eyes: are they dark and sunken, are they crying without tears? This isn’t good. Do they look pale or mottled – well this depends on where you practise.

When I first moved to the UK, I was alarmed to find most children I saw looked pale, and mottling quite common. I noticed, with time, that mottling in fair skin tends to happen when cold, fairly commonly. I learned that if I observed that child when wrapped in a warm blanket, this mottling commonly disappears in a well child. In Mediterranean children/adults, skin doesn’t mottle easily, so if a child from this background presents mottled, be concerned.

Now there is a slight difference here between children who are isotonically dehydrated – these children commonly just need treatment. Many of them will have this “look” but remain quite plump looking, and the only reason for investigation is if you are gaining IV access anyway (according to NICE guidance).

It was generally agreed that investigation should be the exception, not the rule, and if you are investigating, you should probably ask yourself why am I doing this? Is this for the child I have a bad hunch about? And what am I going to do if my hunch is wrong? (the answer to this is probably to observe). Things that may help are a blood gas with a highly deranged base deficit and a very high (about 4 mmol/l) lactate. Don’t be tempted to measure ketones as these have been proven to be a bad marker of dehydration.

Those who have a deranged sodium or glucose tend to be a bit different. They may just look more severe, but they can have neurological signs like increased tone and jitteriness, and they do tend to be a bit sleepier, more dehydrated and sometimes even confused. These children need investigation, there is no doubt about it.

Most importantly, if you consider that severe dehydration in a first world country is unlikely, if you are seeing a child with severe dehydration, then you have to think about the possibility of underlying illness –  especially DKA.

Be especially cautious with children less than 6 months old and children with learning difficulties or special needs, as their signs may be softer and therefore easier to miss!

And if you want to open that Pandora’s box about treatment, here is a tweet from Damian Roland to start you off:

References / Further Reading:

  1. ROWLAND, D., CLARKE, C., BORLAND, M.L., PASCOE, E.M. (2010) Does a standardised scoring system of clinical signs reduce variability between doctors’ assessments of the potentially dehydrated child?Journal of paediatrics and child health. 46(3):103-107.
  2. STEINER, M.J., DEWALT, D.A., BYERLEY, J.S. (2004) Is this child dehydrated?. JAMA. 2004 Jun 9;291(22):2746-54.
  3. Tagg, A. (2016) Assessing Dehydration, Don’t Forget the Bubbles, 2016.
  4. FREEDMAN, S.B., VANDERMEER, B., MILNE, A., HARTLING, L. (2015) Diagnosing Clinically Significant Dehydration in Children with Acute Gastroenteritis Using Noninvasive Methods: A Meta-Analysis.The Journal of Pediatrics.166(4): 908-916.
  5. NICE: Diarrhoea and vomiting caused by gastroenteritis in under 5s: diagnosis and management Clinical guideline [CG84]