Sunday 30 June 2019

Chest X-rays in children - The Wimbledon Rules

We've come a long way in terms of reducing unnecessary tests in paediatrics.  It is within my career that it was standard to obtain a chest X-ray (CXR) for any child presenting with their first episode of wheeze.  Now, such an approach is seen as outdated.  This is a good thing.  In fact the vast majority of acute and sub-acute respiratory presentations in children can be managed without needing a CXR.

In some ways it was a lot easier to know when to do a CXR 20 years ago.  The answer was pretty often.  Every lower respiratory tract infection (LRTI), every first episode of wheeze and every persistent cough tended to result in a CXR.  Now, we should rarely do CXR in those circumstances.  Rarely doesn't mean never though, so how do you know if you're doing too many?  Enter the Wimbledon Rules for CXR in children…  I’ll come to that later.  First, I’ll explore a little bit about the complexities of doing CXRs in paediatrics.

The problem with CXR in children is that it can be misleading.  The most common scenario in which this is true is for the wheezy child.  Wheeze is a strong negative predictor of pneumonia(1).  This makes sense clinically when you think about it.  If an infant or child has restricted lower airways, that is reason enough to have respiratory distress.  If you then take a section of lung out of action, you won't be wondering if they might have a problem.  It is likely to be very obvious from how unwell they are and how abnormal their breathing is.  As a rule, children with tight airways and pneumonia together are in a very bad way.

While wheeze is a strong negative predictor of pneumonia, a CXR in wheezy children is rarely clear.  In many cases there is a patchy white area on the CXR.  This is often at the right heart border, or as it is sometimes called, "the area of radiological romance."  If you do a CXR too often in wheezy children, this will happen fairly frequently and it may be difficult to ignore.

Even if a child does have a LRTI, CXR is not necessary in many cases.  The British Thoracic Society (BTS) guidelines for community acquired pneumonia (CAP) recommend the following:

  • Chest radiography should not be considered a routine investigation in children thought to have community acquired pneumonia. 
  • Children with signs and symptoms of pneumonia who are not admitted to hospital should not have a chest x-ray.

These recommendations are based on two important facts.

  • Children with clear clinical signs of CAP may have a normal CXR
  • Children with abnormal findings on a CXR often do not have clinically significant CAP (2)

So when should we do CXR in children?

Let’s start with the times when CXR is not recommended routinely:

  • Bronchiolitis not requiring admission to a critical care unit (PCCU)
  • Episodes of asthma and viral wheeze (no matter how severe or whether it is the first episode of wheeze for that child) which are responding to treatment
  • Community acquired pneumonia without atypical features and which responds to treatment within the first two days
  • Most cases of cough without other features
  • Chest pain in children

CXR is usually most helpful in children in these circumstances

  • Severe exacerbations of asthma or viral wheeze which are getting worse despite appropriate treatment 
  • Community acquired pneumonia which has atypical features or fails to respond to appropriate treatment
  • Daily cough with any of the following features
    • Lasting more than 8 weeks
    • Progressively worsening over several weeks, esp. if moist cough
    • Red flag features (daily fever, night sweats, weight loss)
    • Known exposure to TB
    • History consistent with inhaled foreign body

The other side of the problem is that there is no gold standard test for many of those clinical scenarios where CXR is not routinely recommended.  There is often poor correlation between clinical and radiological findings, but which is more valid?  For example if you take pneumonia in children and treat based on radiological findings versus clinical findings you will end up treating different children.  Clinical findings will be falsely positive and falsely negative just as radiological findings are.

Therefore we need to get a balance between clinical common sense and judicious use of CXR in children.  A simplistic approach which could be applied looks like this:
Whether a CXR is necessary or not is highly subjective.  Ask ten clinicians and you'll get ten different answers, due to the human factors.  It's a little like an umpire in a tennis game.  They're not right all of the time.

For this reason, in a major tournament tennis game, players are allowed to appeal.  However the players appeals are limited.  If they appeal against a decision and that appeal is upheld, they retain the number of appeals that they had before the appeal.  So wrong once, they can appeal again.  Wrong twice and they're out of appeals.

I suggest that clinicians should apply the same rules to the use of CXR in children.  Before doing a CXR, we should ask ourselves the question, "What would I do based on a purely clinical assessment?"  After doing a CXR, we should then ask, "Has the CXR added genuinely useful information to my clinical decision?"

Having a CXR result in a child which doesn't alter our clinical decision, or which dysfunctionally suggests a pathology in the absence of a congruous clinical picture should make us rethink our approach to our use of CXR.  If we're going to apply the Wimbledon CXR rule, when we get one completely normal CXR (or one with a non-descript small white fluffy patch which makes us want to give antibiotics when we wouldn't have done so before the CXR) we should think about more cautious use of CXR. If we get two, we should stop and re-read the rules.

Just as tournament tennis players don't have an unlimited number of appeals, we shouldn't think of CXRs as an unlimited diagnostic resource.  We should use them when they are most likely to change our game.

Edward Snelson
Unappealing Paediatrician
@sailordoctor

Disclaimer: If you turn the umpire off and back on, the number of appeals resets.
References
  1. Hirsch, A. et al., Estimating Risk of Pneumonia in a Prospective Emergency Department Cohort, The Journal of Pediatrics, Volume 204, 172 - 176.e1
  2. Virkki R, et al. Radiographic follow-up of pneumonia in children. Pediatr Pulm 2005;40:223e7.