Fluid therapy in shocked children - NEJM article

I have mentioned this article on twitter already, but I think it is such an important piece I wanted to write about it in a little more detail. I would also like to express my admiration for the authors here; a study challenging established practice is always difficult, especially in paediatrics, and even more so in a resource poor environment.

The article in question is the FEAST trial, published in the NEJM. It is a RCT conducted in 3 West African countries. They included 3141 kids with 'severe febrile illness' and evidence of shock (low GCS and/or respiratory distress, AND evidence of poor perfusion).  Excluded were children with gastroentetritis, or those shocked because of trauma or burns. The 29 children with severe hypotension were grouped separately (difficult to justify withholding fluid bolus from them!) and effectively excluded.

The 3141 children were randomised between:

• 20mls/kg bolus of saline over 1 hour (repeated if signs of shock persisted)
• 20mls/kg 5% HAS over 1 hour (repeated if signs of shock persisted)
• Maintenance fluid only.

All children received standard care otherwise (maintenance fluid, abx, antimalarials etc). The protocol was amended during the study and the fluid bolus volume increased to 40mls/kg.

The primary end point was mortality at 48hrs, a solid an end point as you can get.

As you may have guessed from the interest surrounding this study, the results go against our established practice. Mortality in children receiving a fluid bolus was 3.3% GREATER than those who did not; a relative risk of 1.45 (95% CI, 1.13 to 1.86; P=0.003). There was no difference between the NaCl and HAS groups.

Now this is a dramatic finding, and is sure to generate a lot of discussion. Fluid bolus therapy is a mainstay of paediatric critical care and any challenge to this needs to be carefully scrutinised. To my eye, the results seem valid. The sample size is large enough (although the trial was stopped early as the clinical effect became apparent), well randomised and the multi-centre nature of the trial does not seem to have produced significant hetrogenicity. There was very little crossover between groups.

There are points that need thinking about though:

• We already know that excessive fluid therapy can be bad for children (cerebral and pulmonary oedema can be a real problem). However very few children developed these problems in this group.

• How applicable to are these results to practice in the developed world? These children were very sick (presentation of illness tends to occur later in the developing world), will have had different pathology (quite a few had malaria) and the clinical environment would have been very different (no CVP monitoring here!).

• Also, remember that children who were severely hypotensive were not included, no one is suggesting that we do not give fluid boluses in them.

So, should we all now go away and stop giving fluid boluses to sick children (and rewrite a large chunk of the APLS manual)? Not yet, but this result certainly raises questions that need to be urgently answered. It will be interesting to see the responses and debate about this over the next few months.

For me this touches on a larger issue, the need we have in emergency/critical care to chase 'normal' physiology. Many of the symptoms of shock we look for are evolved responses to a severe illness. Are we doing harm by trying to correct them, rather than letting the body sort itself out?

Last but not least, this article illustrates the importance of challenging custom and practice. just because 'we've always done it this way' doesn't mean it's right!

Lots to think about here. Would be very interested to hear everyone's opinion on this. Thoughts? Comments? Disagree with my intepretation? Please leave a comment, or get into a twitter debate @DrGDH. Thanks for reading!