The trial this time is from Japan, Nakamura T, et al. Early inhaled steroid use in extremely low birthweight infants: a randomised controlled trial. Archives of Disease in Childhood - Fetal and Neonatal Edition. 2016. 211 babies with a birth weight less than 1000 grams were randomized between June 2006 and Dec 2009; if they were still intubated at 24 hours they received a metered-dose inhaler with either fluticasone or placebo, from which they had 2 'puffs' (total dose of 100 microgams) per day for as long as they were intubated (maximum 6 weeks). The initial plan was to randomize about twice as many babies, but they ran out of money.
The primary outcome variable was going home on oxygen, which I think is a more important outcome than oxygen need at 36 weeks, needing oxygen at home has much more impact on the family than O2 at 36 weeks. However, even in a blinded trial, criteria for needing home oxygen should be clear, so that the external validity of the result can be assessed. I don't know if the criteria for home oxygen in Japan are the same as for my babies, for example. I can't find any criteria for home oxygen in the trial report.
The babies were followed until 3 years of age with a very high rate of retention of the subjects, almost all of the survivors were evaluated.
Basically there was no effect. No improvement in discharge without oxygen, and no impact on neurological or developmental progress.
There were a couple of subgroups that had a difference in the primary outcome, the middle stratum of gestational ages, 24 to 26 weeks, had a different (possible positive effect of the steroids) effect to the other 2 strata, but I don't see a statistical evaluation of the interaction, so even though the result is different to the other strata we don't know if that difference is itself statistically significant.
Babies who had a history of chorioamnionitis also might have had more of an effect, and babies with evidence of RDS might also have had more of an effect, again, for neither of these sub-group comparisons is there a test of the significance of the interaction term.
One surprising aspect of this study is how few of the babies received antenatal steroids; less than 43% in each group. Why? I think its questionable ethically to perform studies when a simple, safe, proven, intervention is not being applied. Similar studies in other populations will usually have over 80% antenatal steroid use, and most of the missed patients are those that deliver too quickly to get a benefit. (80% of the babies were delivered by cesarean section, in contrast, so they were getting active perinatal care). To give the authors the benefit of some doubt, the trial entry criteria may have partially selected babies without antenatal steroid coverage, as they are the ones more likely to be still intubated at 24 hours of age. But I don't think this can be the entire reason for this very low coverage with antenatal steroids, the recent Neurosis trial for example had 90% antenatal steroid use for a similar group of babies.
What this study shows is that in babies with a totally inadequate level of coverage with antenatal steroids, postnatal inhaled steroids don't have much detectable effect on long term oxygen requirements, or on longer term development.
So where do we stand now, with these recent early pulmonary steroid studies? I think there is still room for a well-powered trial; I'm not sure if budesonide mixed with surfactant and given a few times, with each dose of surfactant for example, would be preferable, or an inhalation (of budesonide of even fluticasone) given over the first days and weeks of assisted ventilation make most sense. This new trial and the modest benefits in the Neurosis trial make me wonder if an inhaled aerosol is likely to give much benefit.
Needing oxygen at hospital discharge, and other effects which are of consequence for the life of the family (such as re-admission for respiratory indications during the first year of life) and developmental and neurological progress of the infants should be the outcomes of interest, I think, but a wide consultation of parents to ask them what is most important as an outcome for such a trial would be even better.