What I learnt this week: If you can’t trust oxygen what can you depend on? #WILTW

This is the 73rd #WILTW

The story of thalidomide is a tragic one. A drug which was deemed to  be safe, but not adequately tested, resulted in over 10000 children being born was thalidomide related disabilities. The impact of thalidomide would have been even more tragic were it not for Frances Kelsey an American scientist, who despite a considerable amount of pressure, refused to licence the drug in America without more evidence of its safety.


Sometimes even with thorough testing complications can be missed. A compound containing aspirin was probably used by Hippocrates in around 430 BC and it has been used close to its current form since the First World War. However children with viral infections can become severely unwell, or even die, if given aspirin (due to the development of Reye’s Syndrome) and hence it should be avoided until adulthood. It wasn’t until the late 1980’s that this was realised and national recommendations to avoid aspirin in children were released. Since that time the incidence of Reye’s syndrome has dropped.

Clearly all drugs have an element of risk to them but it is  surprising when something you think you understand turns out to have unexpected consequences. A recent publication in the Lancet examined what happened to infants admitted to hospital with bronchiolitis (a very common winter virus). The researchers cleverly developed an oxygen monitor that showed a higher oxygen level (or saturation) than was actually the case. Essentially when a child had saturations of 90% the monitor would tell the doctor it was 94%. Over 300 children were managed using these modified monitors compared to another 300 with normal ones. The result of this was that one group received more oxygen therapy than the other (doctors will tend to prescribe oxygen at levels below 94%.)


A very good analysis of the study can be found here but essentially the authors found that not only did the infants who had the lower oxygen levels do no worse; they might have actually done better than those with the normally functioning monitors.

As with all research nothing is completely black and white. The trial was designed with a  very patient (?parent) centered outcome of time to resolution of cough. The ‘bronchiolitic’ cough can be very persistent and the time until it resolved was the same in both groups. However the study design meant some of the other improved outcomes, such as time to return to normal feeding or representation following discharge, in the lower oxygen levels group could have occurred by chance alone. But even with some of these potential statistical quirks there is an enough in this study to seriously question whether supplementary oxygen could cause harm in some babies with bronchiolitis. In fact we already know too much in pre-term babies can cause eye complications but this study has really struck at the heart of a perceived wisdom in medical practice.

This study should make all health care professionals wary of dogma and the status quo. While clearly is not the time for knee jerk reactions or mass panic it is a timely reminder of the importance of re-evaluation and research in all that we do. It has certainly fuelled the fire of my own academic interests and I hope it does in others.

What have you learnt this week? #WILTW


One thought on “What I learnt this week: If you can’t trust oxygen what can you depend on? #WILTW”

  1. hmm… strikes me there are a number of quite important confounding variables in this. How were they delivering O2? in a head box or nasal cannula? Both impact on a child’s ability to feed and also the parental perception of how sick their child is. And do we know that they actually got O2 or was it just ‘a waft’ 😉

Leave a Reply