Oxygen, transfusions, and NEC

What causes NEC! The answer is ‘who knows?’, or perhaps a better one would be ‘its complicated…’ You probably need an immature gut, gut mucosal injury, inflammation and cytokine release, and an abnormal microbiome, all of which probably interact in complex ways.

In a comment on a previous post, John Lantos remarked that the majority of the increased mortality in the SUPPORT trial was from NEC. I replied that he was right, there were more also a few more  deaths from sepsis and from bronchopulmonary dysplasia, but the biggest difference was in NEC deaths, 23 vs 14. There wasn’t a big difference in diagnosis of NEC (stage 2 or 3), 11.9% vs 10.8%, just in mortality from NEC.

I seemed to remember that the pattern was the same in the BOOST 2 trials also, indeed the rate of NEC (requiring surgery or causing death, that is, more severe than in SUPPORT) was 12.2% vs 9.9%, with the new algorithm installed on the pulse oximeters, in the low saturation group compared to the high saturation group. There were 39 deaths from NEC in the low saturation group, and 22 in the high saturation group. There were also more deaths from septicaemia, 28 vs 15, and more deaths from BPD 18 vs 12.

In the COT trial there were also more cases of NEC overall (stage 2 or more) in the low saturation group than the high saturation group, 12.3% vs 9.3%, but I don’t know about causes of death, or whether the difference changed with the re-programming of the pulse oximeters.

The reason for going back over this is twofold, firstly, the observational study about TANEC that I mentioned in that previous post showed that babies who had a lower hemoglobin before their transfusion seemed to be more likely to develop TANEC; secondly a new commentary published in the Journal of Perinatology discusses the ‘limitations of the randomized clinical trial and the end of equipoise’ using the oxygen trials as an example.

That first point makes me think that there may be a common issue, that of intestinal re-oxygenation. If more severely anemic infants have somewhat hypoxic intestinal mucosa, and transfusion improves that, as well as supplying inflammatory mediators, then that could trigger the injury leading to NEC. Also it has been shown that infants in the low saturation arm of the SUPPORT trial had more frequent and severe intermittent hypoxia, with multiple episodes per day of hypoxia, re-oxygenation (and perhaps hyperoxia as well, more about that later). So there may be a common link, how that relates to sepsis or BPD I am not sure.

Although there are some issues in the commentary that I agree with, I think there are some fundamental errors also. The authors state that the oxygen trials were elegant, important, adequately powered, etc, but that clinicians seem reluctant to institute changes in practice, they suggest that this is because we thought that perhaps there would be less RoP with lower oxygenation, but that the mechanisms of the increase in mortality is uncertain. I am not sure that is likely to be true, I think therapeutic intertia is a bigger issue.

They then discuss whether there might be situations in which  a lower saturation target might be right for an individual baby,

The authors of the commentary use a very limited view of EBM

The idea of evidence-based medicine is straightforward—the clinician needs to use the best available evidence to determine optimal therapy for his or her patient

David Sackett, in contrast, puts it this way

Evidence Based Medicine is the integration of clinical expertise, patient values, and the best evidence into the decision making process for patient care. Clinical expertise refers to the clinician’s cumulated experience, education and clinical skills. The patient brings to the encounter his or her own personal and unique concerns, expectations, and values. The best evidence is usually found in clinically relevant research that has been conducted using sound methodology.

Which is a much richer understanding, and which actually addresses much of what they discuss. They make much of the fact that a particular RCT may not address the issues of an individual complex patient, which is obviously the case, and is the reason why exploratory sub-group analyses are performed on large RCTs, to try and identify patients for whom further investigation may be required. I emphsize, further investigation is required, subgroup analyses are fraught with problems and have sometimes been proved unreliable when those further studies are done.

They conclude with the following:

What we do as clinicians with the newest evidence about oxygen-saturation targets and ELGANs remains to be seen. As the product of elegantly designed and conducted international multicenter RCTs, the evidence for the benefit of higher oxygen saturation targets will be heavily weighted. But as the product of an RCT it may not necessarily be the best practice for every individual ELGAN.

I would respond to that: tell me when you think it is better for a baby to have lower saturation limits. Under what circumstances, for this particular issue should you expose the baby to higher risks of death, and lower risks of RoP? Of course it is likely to be true, as they point out in the commentary, that some babies who developed RoP would have escaped it if they had been in the lower group, and they may not have been the same babies who would have died. In other words, an individual baby, if you knew from the day of their birth that they would survive, and would not develop NEC or sepsis or BPD, but were at risk of severe retinopathy, would be better in the low oxygen group.

If you know how to identify such babies, then you should design an RCT to enroll them, and prove that they are indeed safer with lower saturations.

I can actually see that could be true, it could be that in a particular clinical situation a lower saturation might be beneficial. But unless you can define and test that possibility there doesn’t seem to me to be any reason to expose babies to the increased risks of death (from NEC sepsis and BPD) which accompany the lower saturations.

And to be honest, in their NICU do they really go round every day and decide which babies should have saturation targets of 85 to 89, 88 to 93, or 90 to 95%? I really doubt it, I don’t know anyone who does that. So their argument just falls flat, for this particular issue; I would bet you that they have standard oxygen saturation target ranges for preterm infants in their NICU, because we have no evidence for doing anything else. Of course clinical expertise is important in neonatology, and it is an essential part of evidence based practice, but I have no idea what kind of clinical expertise ‘cumulated experience, education and clinical skills’ you could use for assigning a different oxygen saturation target for an individual-complex baby, to use their term.

The authors discuss research equipoise and clinical equipoise as if they were different. I do not agree with their evaluation. Equipoise just means that there is a therapeutic choice and you do not know which is best for an individual baby. In such a circumstance it is appropriate to design an RCT or enter the baby in an RCT, or, if there isn’t one approved just make your best guess, by adding up all the pros and cons, and try and decide the least worst option.

For oxygen saturation targets we no longer have much uncertainty, the higher saturation targets decrease mortality and we can be 99% confident that this is true. If you are still using lower saturation limits then I think you should inform parents about this evidence and explain why you are have decided to use lower saturations.

Which brings me back to a sentence earlier in the commentary

‘If one were to consent parents to a trial arm for which there is now evidence for increased mortality, would any parent reasonably be expected to provide permission for their child to be enrolled?’

If the authors think is unlikely that parents would consent for such a trial, why should parents be expected to consent for individual treatment according to a standard ‘for which there is now evidence of increased mortality’? If my baby were in an NICU and the neonatologist informed me of that evidence, and then said, ‘but I have decided with my clinical experience that your baby should nevertheless have lower saturation targets’ I think we would have some serious discussions.

About Keith Barrington

I am a neonatologist and clinical researcher at Sainte Justine University Health Center in Montréal
This entry was posted in Neonatal Research and tagged , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.