The oxygen saturation targeting trials showed more retinopathy with higher oxygen saturation targets. Will this translate into more retinopathy in actual practice? Many units have increased their saturation targets as a result of those studies. This may indeed lead to more RoP, and the expected result seems to have happened, in Melbourne at least. A before and after study showed that there was an increase in retinopathy, both overall and of stage 2 or more. They evaluated the outcomes in babies under 30 weeks or under 1250 g who survived to get retinal screening, about 150 before they changed and nearly 200 afterward. They saw more total RoP. There were very few babies who needed laser, 1 before and 3 after, which may have been due to chance, with such small numbers, but it certainly didn’t go down. Among infants of less than 28 weeks, the stage 2 disease incidence went from 16% to 34%.

What can we do to counter this? Which is likely to be repeated in many different units who have raised their saturation limits. What interventions are there that can reduce retinopathy? Well, we know that poor early neonatal growth is associated with an increased risk of retinopathy. So optimizing neonatal nutrition in at-risk infants should help, although that hasn’t been proven in randomized trials, as far as I am aware.

An RCT from Poland (with the collaboration of Michael Sherman) suggests that a mixed lipid source, including some fish oil/omega-3 fatty acids may reduce RoP in very preterm infants. They randomized infants <1250g and less than 32 weeks. The 70 control infants received a mixture of soy and olive oil lipids (Clinoleic, which is used in Europe, but I don’t think is licensed in North America). The 60 babies in the intervention group received half Clinoleic, and half Omegaven (the fish oil emulsion with lots of omega-3 goodies) by volume, but the Clinoleic is a 20% emulsion, the Omegaven is a 10% emulsion. So they had about 1/3 omegaven, 2/3 the other stuff.

It was an unblinded study, apart from the ophthalmologist who was apparently… uuh…blinded. (Sorry about that). The primary study outcome was “an assessment of ROP severity and whether laser photocoagulation was required to save vision” which isn’t a primary study outcome, of course, you can’t have a study outcome which is an “assessment of” something. The sample size was calculated based on an extremely high frequency of retinal ablation in the controls, of 27%, and a massive reduction to 7.5% in the fish oil group. so we will call that the primary outcome variable.

The babies in the study were on average just under 1000g birth weight, and were 28 weeks gestational age. Those in the new-lipid group got much more DHA (docosahexaenoic acid, an omega-3 FA) which they incorporated into red cell membranes. They defined cholestasis as more than 20% of the bilirubin being conjugated; which was very frequent in the controls, 20/70 vs 3/60 in the group with the fish oil. Treated retinopathy was extremely frequent in the controls, 22/70 and decreased to become very frequent in the fish oil group, 9/60.

The proportions of fish oil are not the same with their mixture as with SMOFLipid, a mixture of Soy, Medium chain triglycerides, Olive oil, and Fish Lipids, which is commercially available. In SMOF about 15% of the lipids are fish oil derived.

This looks hopeful, but we will need to be sure that the results are the same with an approved lipid in Canada (SMOFLipid is approved, but not specifically for newborns in Canada), or wherever you are in the world. Also that it is effective in NICUs with a much lower rate of retinopathy needing treatment, then providing enough omega-3 FAs from early life in the very preterm infant might help to counter the effects of keeping the saturations a little higher.