Simple scientific questions require simple scientific answers; doctors want unequivocal guidance to their practice so that they do not fumble in the dark. But it is easier to ask questions than to answer them, as two papers published in the same week in the New England Journal of Medicine and the Journal of the American Medical Association attest.

The question asked by the two papers was the optimum level of oxygenation in the blood of pre-term infants. In the past it was rather naively supposed that if oxygen were necessary, then more of it must be better; but premature infants who were exposed to high levels of oxygen developed a condition known as retinopathy of prematurity, often leaving them blind or severely impaired visually.

The two trials, one from Britain, Australia and New Zealand, and the other from the United States, Canada, Argentina, Finland, Germany and Israel, sought to establish whether a higher or lower level of oxygen saturation of the blood was better for infants born very prematurely. The results were different, if not quite diametrically opposed.

The first trial found that babies treated so that their blood oxygen saturation was higher had a lower death rate at 36 weeks than those treated so that their levels were lower. 15.9 percent in the high-saturation group died compared with 23.1 per cent in the lower. You would have to treat 14 babies with the high oxygen saturation to save life more than treating them at the lower level.

The babies with the higher levels of oxygen saturation developed retinopathy of prematurity more frequently than those with lower (13.5 percent versus 10.6 percent), but nowadays the condition is treatable and does not usually lead to blindness.

By contrast, the other trial did not find any difference in the mortality rates of infants treated at higher and lower levels of oxygen saturation of the blood. The authors followed up the infants for longer — 18 months — and used a broader measure of outcome, namely a combination of death, gross motor disability, cognitive or language delay, severe hearing loss or bilateral blindness at the end of that period. Here, I confess, I found the results horrifying: in the high oxygen saturation group, 15.3 percent had died, while in the lower 16.6 percent had done so (a difference in rate so small that it was not statistically significant, that is to say it could have arisen by chance), while 34.4 percent of the higher and 35 percent of the lower saturation level babies suffered from at least one of the other conditions, namely gross motor disability, cognitive or language delay, severe hearing loss or bilateral blindness. Survival is thus often bought at a considerable cost: forty percent of survivors were handicapped or severely handicapped.

Naturally further research is required: it always is. The trials tested only two ranges of oxygen saturation, 85 to 89 percent and 91 to 95 percent. Other ranges were not investigated, as in theory they could be. Furthermore, a more prolonged follow-up of the infants might reveal differences between the groups that developed later in their lives.

What, then, is the neonatologist to conclude from these inconsistent results? He cannot, like Tybalt, just exclaim, “A curse on both your houses!” He has to make a choice.

One trial found no difference between the outcomes and the other a better outcome with a higher level of oxygen saturation. It seems to me, therefore, that he should choose the latter, because it might do good and has not been shown to do irreversible harm. But still one would prefer to have no inconsistency in the scientific results. However, in medicine inconsistency is the norm.


Do you have a question about health and medicine you’d like to see addressed? Click below to contact PJ Lifestyle.

 Shutterstock images courtesy / Anneka / dream designs