Bassler D, et al. Early Inhaled Budesonide for the Prevention of Bronchopulmonary Dysplasia. New England Journal of Medicine. 2015;373(16):1497-506. One of those ongoing trials that we have been awaiting the results of has just been published. The NEUROSIS trial was a randomized controlled trial with a very respectable sample size, 863 infants less than 28 weeks, by far the largest investigating this issue. Babies got budesonide (or placebo) twice a day by a metered dose inhaler with an aerochamber until they were 14 days of age then once a day until they were 32 weeks or off oxygen.

The results showed a reduction in the primary outcome, (death or oxygen requirement at 36 weeks) with the inhaled steroids. There were no clinical complications that were substantially different between groups. Mortality was numerically higher in the budesonide group, 16.9% compared to 13. 6%, and BPD was much lower, 27.8 vs 38%.

There were similar reductions in the primary outcome in the most immature stratum, as well as in the more mature babies. The 2 components of the primary outcome changed in different directions, as you can see from the numbers above. What I don't understand is why this difference in mortality is referred to as a 'borderline significant between-group difference in the rate of death' the 95% confidence intervals are far from from suggesting a difference between the groups, the relative risk of death, with 95% confidence lies between 0.91 to 1.69.

I think this study was performed to the highest standards, and the presentation of the results is appropriately conservative. Maybe too much emphasis is placed on the no-where-near significant increase in mortality, but the long term clinical significance of oxygen requirement at 36 weeks of age is also quite questionable.

Some animal models show adverse effects of steroids on lung growth, even though the steroids reduce pulmonary inflammation. So a reduction in the proportion of babies needing oxygen at 36 weeks does not necessarily mean that long term pulmonary health will be improved. Indeed the studies of dexamethasone as treatment for BPD, despite a reduction in the diagnosis of BPD (because short-term improvements in gas exchange lead to fewer babies needing oxygen at 36 weeks), do not show any improvement in longer term pulmonary health.

We need to rethink how we do studies such as this, it could be considered a positive study, as the primary outcome was improved, or a null study, because the more important outcome, death, was not different between groups. What we really need now is to know if those babies who got the steroids have better pulmonary health in the first years of life, or not. And also, of course, whether their neurological and cognitive development has been affected.

A systematic review of similar studies with data on mortality may help to clarify if that difference is just a fluke, or part of a wider pattern.