I think Brett Manley is going for the record as the person with the highest proportion of his publications in the FPNEJM, he now has 3, with 2 of them as first author. This is the HUNTER trial where babies in level 2b NICUs, as we could call them, i.e. nurseries with access to CPAP, but not prolonged invasive ventilation, were randomized to receive either CPAP or high-flow nasal therapy. (Manley BJ, et al. Nasal High-Flow Therapy for Newborn Infants in Special Care Nurseries. The New England journal of medicine. 2019;380(21):2031-40) Babies are not usually kept in such units if they are less than 32 weeks or less than 1200 g, so to be eligible for this trial they had to be at least 31 weeks gestation, >1200 g birth weight, and to need non-invasive respiratory support according to the attending pediatrician. This was performed as a non-inferiority trial, and was designed to be able to detect an increase in therapeutic failure from 17% with CPAP to 27% with high flow.

Babies in the high flow group were placed on 6 litres per minute of a heated humidified gas mixture via the Fisher-Paykell Optiflow device, which could be increased to a maximum of 8 lpm. CPAP was delivered using a bubble system and either a mask or prongs, at 6 cmH2O, which could be increased to a max of 8 cmH2O. If a high-flow baby failed they could be treated with CPAP at 8 cmH2O, if a CPAP baby failed they were out of the trial and a discussion with the regional NICU was expected.

There were over 750 babies randomized, and the primary outcome was treatment failure in the first 72 hours: which was defined as; if they got to maximal support and needed more than 40% oxygen for more than 1 hour (to stay at 91 to 96% saturation), or had a respiratory acidosis to less than 7.2, or had severe apnea.

Treatment failure occurred in 20.5% of the high flow babies and 10.2% of the CPAP group. Many of those who failed high flow were transferred to CPAP and about half of them stabilized, so in the end just over 5% of each group were intubated within the 1st 72 h, and around 6% in total in each group; there were twice as many pneumothoraces needing intervention in the CPAP group (4.8% vs 2.4%, 95% confidence intervals include compatibility with no difference). Slightly more babies in the high-flow group were transferred to tertiary NICUs, 13% vs 11%.

This is really important clinically relevant data for level 2 nurseries. The eventual clinical outcomes were quite similar in the two groups, so I think it would be acceptable to continue to use high-flow in such nurseries as long as you had back up CPAP readily available. There were  no clear advantages of high-flow demonstrated in this study but other studies have found that parents prefer high-flow, and they appear to be more comfortable for the babies.

The only quarrel I have with this study is how tortuous the language becomes with non-inferiority studies. We are told that high-flow was “not non-inferior” to CPAP. Why not just “was inferior”? There were substantially more treatment failures, which passed the non-inferiority margin, hence the intervention was inferior in terms of the primary outcome.