Adding my uncertainty to your uncertainty – how to decide when there are conflicting signs

It’s always nice when a clinical assessment ends with a clear way forward.  If a child presents in the community with a severe exacerbation of asthma, there are clear guidelines for pre-hospital treatment and referral.  There is usually no dilemma in what to do if a child presents with fever and signs of meningitis to Primary Care.  Similarly, it is usually obvious what to do with a child brought with signs of an uncomplicated viral upper respiratory tract infection.

What is more challenging is when a child has conflicting or paradoxical signs and symptoms.  This is especially true when all the evidence points towards an illness that can and should be managed in Primary Care, yet one small piece of information is inconsistent.  What should you do with that?

Let me give an example:

A four year old boy presents with a history of diarrhoea and vomiting for 24 hours.  It is now 6pm.  The history of the illness is typical of a straightforward viral gastroenteritis.  The child looks well and is alert, interactive and surprisingly willing to mobilise when he spots the toys available.  He has wet mucous membranes and has no signs of dehydration.  Abdominal examination is normal as is the rest of his systemic exam.  All physiological observations are normal.  There is an unexpected snag: the parent with him reports that he has not passed urine since he went to sleep the previous night. 

So, with all but one piece of information in favour of a pathway that involves symptomatic management and oral hydration, what should one do?

Should this child be referred to Secondary Care?  I think you will get different answers from different people for this kind of situation.  On one hand the reported lack of urine output is a clear red flag.  On the other hand it is completely at odds with the wellness and clinical hydration of the child.  The child could be referred, but would that change anything?

There are many reasons for referring a child to Secondary Care.  I the initial examples, the reason is for undisputable and necessary management.  If the child is referred in the case of the paradoxical urine output, there may be an assumption that someone in the hospital will be better able to make a decision than the person making the referral.  This may or may not be true.  If it is true, it is probably not because they will have further information that could not have been obtained by the referrer.  It is probably not because the person receiving the referral has some sort of child assessment super power.  They may not even have more confidence in their gut feeling (that the child is fine and can safely be managed as an uncomplicated gastroenteritis) than the referrer.  What they will mostly have is their uncertainty to add to the referrer’s uncertainty.

Let’s say that you do refer this child.  Let’s assume that I see them shortly after you have seen them and that nothing has changed apart from the child having a bit more to drink in the meantime.  I now have the same information that you had and the same dilemma: to go one way or another.  I can do blood tests and consider intravenous fluids, but my gut tells me that this is an unnecessary torture of a child best managed with analgesia and oral fluids.  I could give symptomatic treatment advice and safety-netting advice and send them home but that is difficult to justify in the face of the reported anuria.  I could keep them for observation, but what would be the end point of that or the purpose.  If I am confident enough to do no active management in hospital, why am I keeping them?  Observing moderately unwell children while awaiting an expected move towards wellness is one thing but observing a well hydrated and clinically well child to prove that they are going to pass urine seems to have little value over allowing the same outcome at home.

When faced with conflicting information, it is important to have a way of deciding which information carries more weight and has more validity.  In ill children, it is often useful to categorise according to effort and efficacy. 

Effort is the body’s way of responding to illness.  There is usually some sort of physiological response to illness which is often a way of compensation for an effect of the illness.  Efficacy is everything to do with the end point of how effective the increased effort is.  In other words effort tells you that the child is trying to cope with an illness and efficacy tells you if it is working.  A good example is the happy wheezer.  If an infant with bronchiolitis has an increased respiratory rate, that tells you that there is some effort to compensate for what is happening in the chest.  The smile tells you that this is being effective.

Inevitably, efficacy is more important than effort when it comes to clinical decision making.  It’s not that the compensatory efforts should be ignored, it’s that the effect of that effort is of greater importance. 

So in this case, I would say that whatever is happening with this gastroenteritis, all the evidence that I can see tells me that it is having the desired effect.  The child is well and has no signs of dehydration.  Even if we were to believe the lack of urine output (and to disbelieve is not to disbelieve the parent) then a reduced urine output would be part of an effort to appropriately retain fluid during an illness.  I’ll say that is being quite effective in this case.

Do I have any niggling doubts?  I am certainly not dismissive of the contradictory report of anuria.  I am aware of my own fallibility and would only chose to ignore the lack of history of good urine output if I felt that to do so was in the child’s best interests.  In short, I would love complete certainty but I’m not being given that so I can either make a decision or find another way.  What I don’t want to do is a test or treatment that I don’t believe in.

But it isn’t strictly true that I’m no further forward than the referring clinician.  After all, I’ve got their uncertainty to add to my own.  If I’m lucky, it will be clear from the referral that when you referred this child for further assessment, you were very happy with the hydration and general wellness of the child.  You, like I am now, were perplexed more than concerned by the lack of history of passage of urine.

I have no super-powers and no particular clinical skills that you don’t have.  I am also lacking a 100% certainty about this case, but it’s time to make a decision.  Adding my certainty to your certainty means that between us we are 198% certain that there is nothing about this child requiring a blood test or a night on an intravenous infusion.  Let’s go with that then shall we?

There’s nothing wrong with referring a child for a further opinion when there is some uncertainty about the best management.  It is particularly helpful to know what the opinion is along with everything else in the referral.  That way I know that my uncertainty is the same, which in itself is useful information.

There’s also nothing wrong with trusting your assessment, in the knowledge that in secondary care we may well have little to add other than another appraisal of the same paradox.  Of course there is always the phone call to discuss the uncertainty.  How much fun would that be?

Edward Snelson
Medical astrologist
@sailordoctor

Disclaimer: I'm not even certain what I'm disclaiming.

Empirical Paediatrics – What is the best rehydration fluid for children with gastroenteritis?

When children have an uncomplicated gastroenteritis, dehydration is the enemy.  There are lots of things that we can give children to drink, but what is the best rehydration fluid to keep children out of hospital?

There are many things said about which fluid is best.  NICE recommends water or milk to maintainhydration but says to avoid fruit juice and carbonated drinks. (1)  In children who require rehydration they recommendoral rehydration solution (ORS).  In my previous post I listed 10 myths about gastroenteritis and used that to point out the NICE recommendations.  This advice conflicts with a study published in the JAMA (2) this year which showed that half strength apple juice was more likely to work than oral rehydration solution.

So, I’ve had another look at the basis of the NICE recommendation that we should “use ORS solution to rehydrate children” and the reason given is: “Although there were no clinical trials on the effectiveness of fluids other than ORS solution in the treatment of dehydration, the GDG considered that the composition of such fluids was generally inappropriate.  In dehydration due to gastroenteritis, both water and electrolyte replacement is essential, and non-ORS solution fluids do not usually contain appropriate constituents.  ORS solution was considered the appropriate fluid for oral rehydration.”

In other words, there was no evidence to support other fluids than ORS and there was reason to suspect that they are not ideal.  The trouble with two plus two is that it only equals four if there are no unknown variables.  The empirical approach dictates that instead of trying to figure out what should be true, we only believe what can be evidenced.  That removes the risk that there are unknown unknowns.  In the case of keeping a child hydrated, there is a big wildcard- the child.

One of the more interesting results of the apple juice vs oral rehydration fluid study was that the effect (half strength apple juice being more effective) was more pronounced in children over the age of 2 years.  I think that phenomenon is easily explained by two things:

• Children develop the ability to choose for themselves
• Oral rehydration fluid has a disgusting taste

When looking after children, we often have to choose between the treatment that is best on paper and the treatment that the child will take.  In this case, taste beats logic hands down.

So what do you do when your guideline tells you not to do something?  I would say that we need to recognise that while guidelines have strengths, they also have weaknesses.  NICE and other similar guidelines require a huge amount of searching through evidence, appraising applicability of the literature, and ultimately a decision to be made by people, with everything that entails.  What we get from that is a load of recommendations from people who have worked really hard to give us the best answers that they can come up with.   Now that there is evidence for apple juice, this will no doubt be considered when the guidelines are revised - a process that will take a very long time.

Even when a guideline is up to date, it is up to us to apply it to the child in front of us.  For example, where is the guideline for treating gastroenteritis in a child with autistic spectrum disorder?

Will flat cola be the next thing to be shown to be effective after all?  Who knows?!  What I do know is this: we now have an evidence base for a rehydration fluid that is palatable, readily available and doesn’t require a trip to the pharmacy or the doctor’s surgery.  That has got to be a win for de-medicalising and a victory for self-care.

What do you do if the child doesn’t like apple juice?  Don’t give them that then.  What is clear is that (within reason) the best rehydration fluid for a child is the one that they will take.

Edward Snelson

Eventual empiricist
@sailordoctor

Disclaimer - I was taught the theory of empricism by my daughter.  If this is all wrong, it's her fault.

References

1. Diarrhoea and vomiting caused by gastroenteritis in under 5s: diagnosis and management, NICE CG84, Published date: April 2009
2. Freedman, S et al, Effect of Dilute Apple Juice and Preferred Fluids vs Electrolyte Maintenance Solution on Treatment Failure Among Children With Mild Gastroenteritis, JAMA. 2016;315(18):1966-1974

Gastroenteritis in Children - Ten Myths

Vomiting and diarrhoea in children is usually caused by viral gastroenteritis.  There are lots of myths surrounding gastroenteritis and how best to manage it.  I find myself repeating things that I was once told years ago and have to check from time to time whether the 'fact' is in fact based in any reality.  When I find out that it was all a myth, it makes me feel so much better when I later hear other people who hold those same myths to be true.  Hopefully, between us we can dispel a few of them.  Here are a few non-truths that I regularly come across:

1.  It's just a virus.  I know that I said it is usually a viral infection in children and that is true.  However that should not fool people into thinking that it is a benign illness.  Even in well nourished children, dehydration is a real risk and every year previously healthy children with gastroenteritis suffer renal failure and other consequences of severe dehydration.  Avoiding dehydration makes for most of the dos and don'ts of gastroenteritis.

2.  Paracetamol should be avoided because it makes the child vomit.  Not so.  What is more nauseating: 5 mls of liquid vitamin P or fever and abdominal pain?  Giving paracetamol is likely to help resolve the vomiting and make the child feel more like they could cope with drinking a few sips of water.  Certainly, children often do vomit shortly after being give paracetamol but when it works, it is well worth it.

3. You shouldn't give milk to children who are vomiting.  The best fluid depends on two factors.  One factor is the level of hydration.  If a child is at risk of or is becoming dehydrated then oral rehydration fluid (ORF) is recommended.  The second factor is the question of what the child will take.  Oral rehydration is really important, so better a bottle of milk that is drunk than a bottle of ORF that is continually refused.  The important thing to avoid is the list of drinks that will make matters worse.  Milk is not on that list.  Just because milky vomit is nasty compared to when the child is drinking clear fluids doesn't mean you should avoid milk if that is what they will take.  Milk contains carbs and electrolytes and for babies it is the fluid of choice.

4.  Flat cola is great for rehydration.  What makes a poor rehyration fluid?  Acidity to worsen gastritis as well as hyperosmolality and added chemicals that will drive diarrhoea.  Flat cola ticks all of these boxes which is why it gets a special mention in the 'don't do it' bit of the NICE guidelines for gastroenteritis in the under five year olds. (1)

5.  You can't give antiemetics to children.  Now we are getting into more controversial territory.  Antiemetics such as prochorperazine and metoclopramide (where would I have been as a house officer without these two drugs?) are traditionally avoided in ill children due to the risk of dystonic reactions.  It has threfore been the case that gastroenteritis has always been in that category of illnesses that just has to get better on its own.  That may be why the world of paediatrics has failed to reconsider this view despite the appearance of newer and safer antiemetics.  There is good evidence for example that ondansetron reduces vomiting and may aid rehydration (2).  So why don't we use that when a child is failing to rehydrate orally?  NICE considered this when writing its guideline and noted that ondansetron is also associated with increased diarrhoea.  The answer was therefore that it could not yet be recommended, but possibly with more research, ondansetron will be recommended in specific circumstances.

6. You can't give antidiarrhoeals to children.  Again, NICE considered the pros and cons of this option.  There are various types of antidiarrhoeal medicines, each of which was decided against in turn, mostly on the basis that there was no evidence for benefit.  In the case of loperamide, there is reasonable evidence that it does help (3).  So what's the problem?  Loperamide is not licensed for use in children in the UK (and I think the same is true in the USA and Australia but I'm not sure about elsewhere).  However, the BNFc does list doses and acknowledges the license issue.  I don't intend to medicalise self limiting gastroenteritis, but if I thought it would help, it is good to know that it is therapeutic option.

7.  A period of starvation can resolve vomiting or diarrhoea.  The only clinical value to an enforced period of starvation for a child is that it is a great way to diagnose MCADD.  Witholding food or drink will not change the course of viral gastroenteritis.  However, some children do have underlying, yet hidden metabolic disorders of energy production.  These children have often had no manifestaion of their disorder because they have never run out of immediately available energy.  When they are unwell and rely on ketones, everything goes wrong and hypogylcaemia can come on profoundly and unexpectedly early into a period of fasting.  Any ill child who is not getting calories and who becomes subdued or agitated should have a blood glucose checked.

8.  It's a 24 hr bug.  In fact who knows how long it will last.  I don't believe that you can make something go wrong just by saying a thing.  For example, I am very happy to walk around at work commenting on how lovely and quite it is and enjoy seeing the superstitious flinch at this.  However predicting the length of a gastroenteritis is a recipe for perplexed parents.  Vomiting usually settles by day 3 and diarrhoea should be at least much improved by day 7.  Should be...
If diarrhoea is not resolving at day 7 then consider doing a stool sample.

9.  It's probably food poisoning.  Thankfully not.  The vast majority of vomiting and diarrhoea in children is viral gastroenteritis.  Bacterial infections are more likely if the child has been to an area with endemic infection.  A history of consuming foods that are likely to have been contaminated is also important.  A sudden onset of vomiting does not imply food poisoning though.  Norovirus for example typically causes sudden and severe symptoms.

10.  Dehydration requires intravenous fluids.  Rehydration is best provided through the gut, not a vein.  Although guidelines are changing in order to avoid dangerously hypotonic fluids, intravenous rehydration will always be risky.  Every effort should be made to achieve oral hydration.  If this fails then nasogastric rehydration has a good evidence base.

Of course these are only the myths that I used to believe before my faith was destroyed by reasoning and evidence.  Do you have any of your own?  If you know of a wrong but popularly held belief to do with gastroenteritis then please post it in the comments below.  Cheers!

Edward Snelson
Grade 'O' in Care of Magical Creatures at O.W.L.
@sailordoctor

Disclaimer: It feels a bit strange to be in agreement with so much of a NICE guideline.  I may be coming down with something.

References

1. Diarrhoea and vomiting caused by gastroenteritis in under 5s: diagnosis and management NICE guidelines [CG84]
2. Szajewska H et al., Meta-analysis: ondansetron for vomiting in acute gastroenteritis in children, Aliment Pharmacol Ther. 2007 Feb 15;25(4):393-400.
3. ST Li et al., Loperamide therapy for acute diarrhea in children: systematic review and meta-analysis, Database of Abstracts of Reviews of Effects (DARE)

Paracetamol as an anti-emetic

Now that children have returned to school you may have notice something new in your consulting room: the smell of stale vomit.  I think that I have a fairly protected professional life but puke is not one of those things that I mention when extolling the virtues of my (eventually) chosen career.  As winter approaches, the number of children with viral gastritis and gastroenteritis will go up and up until one day before Christmas, I will have a loved one ask me to shower before supper so that I can be in polite company.

Perhaps there is one good thing to come out of all this though.  I think that all these vomiting children have taught me something that not everyone else knows: For the child with viral vomiting who will not drink, paracetamol is an accidental anti-emetic.  (I am also fairly sure that there is no medical evidence for it, having done a literature search.)

Anti-emetics are traditionally avoided in  viral vomiting due to the possibility of side effects which might complicate what should be a self-limiting illness.  In the UK, most children with viral gastritis/ gastroenteritis do not become significantly unwell. However it is unpleasant and occasionally causes a child to become severely unwell.

I know that I can’t prove my belief but I can tell you this: I frequently see children who seem unwilling to drink and incapable of appropriate onwards gastric peristalsis (AOGP, as well as a new acronym, happens to be the sound made just before an almighty puke).   Parents bring their children in the hope that we will help.  These children seem beyond simple measures and yet almost without fail, an hour after a dose of paracetamol they are drinking, soon to be followed by appropriate discharge.

Why haven’t they been give paracetamol already?  That’s easy.  It is counter-intuitive to give slimy medicine to a child who vomits every few minutes.  It stand to reason that it will come straight back or precipitate a puke.  The child may not be febrile and even if they are it seems a waste of good medicine.

When I sit down to write these posts, I enjoy the fact that FOAMed allows the writer to be free of the normal constraints of traditional forms of publications.  I am keen, however to make sure that I don’t write anything that is dangerous or wrong.  As mentioned above, I really did do a literature search on this.  You may have noticed that I never sport a serious disclaimer.  This is for two reasons.  Firstly, a disclaimer is no defence against wrongness.  Secondly a lack of a disclaimer is no excuse for you to go and jump off a bridge just because I said so.  We have an understanding, you and I.  I will make this as good as I can make it and you will be professional, question my advice and apply what you get from it in the context of everything else that you know about clinical medicine.  That said, paracetamol really is an anti-emetic when a child has viral gastritis.

In the absence of a randomised controlled trail (that will probably never be done) one needs a plausible explanation for one's claim.  It works like this: When you feel unwell and have abdominal pain you do not want to drink.  Pain and malaise are both emetic in their own right.  Paracetamol reduces pain and feelings of malaise, helping a child to vomit less and drink more.

There are two caveats with this.  Firstly, I would be against giving repeated doses paracetamol too readily to a child who is dehydrated.  Secondly, remember that many paracetamol suspensions contain sugar substitutes which are hyper-osmolar.  Sugar is preferable in vomiting children since they need to energy and don’t need the diarrhoea that can be made worse by sugar free suspensions. 

I know that just because it makes sense doesn't mean it is true, nor does an observed association prove causation. (1)  In this case I have been convinced of an association because it seems to work and it makes sense.  If that’s good enough for you, then great.   There are plenty more like this to come in future posts.

Edward Snelson

Statistical Sceptic when it suits me

@sailordoctor

Disclaimer: Were you even paying attention?  I covered this.

(1)    Predictive factors, Archimedes - Towards evidence based medicine for paediatricians, Bob Phillips, Arch Dis Child 2015;100:892doi:10.1136/archdischild-2015-309409