Nasal continuous positive airway pressure outperforms heated high-flow nasal cannula therapy as primary respiratory therapy in preterm infants

High-flow nasal cannula (HFNC) therapy for preterm newborn infants has quickly gained popularity,1 despite few studies evaluating the underlying mechanisms and lack of high-quality studies evaluating its efficacy.2 Our recent systematic review and meta-analysis,3 and the Cochrane review4 suggested that HFNC therapy was comparable in efficacy to continuous positive airway pressure (CPAP) as a primary mode of support in preterm infants at birth, for respiratory distress syndrome (RDS) in preterm infants at birth and as an aid after extubation from mechanical ventilation or CPAP. However, data for only a limited number of infants born at <32 weeks' gestation were available when HFNC was used as a primary mode of support after birth.3 Thus, the high-quality study by Roberts et al5 is a welcome addition to clarify the role of HFNC therapy as a primary mode of respiratory support after birth for preterm infants ≥28 weeks' gestation. Methods This international, multicentre (nine neonatal intensive care units in Norway and Australia), randomised, non-blinded, non-inferiority trial compared HFNC with CPAP for primary respiratory support for RDS at birth in preterm infants born at ≥28 weeks' gestation. The primary outcome was treatment failure within 72 hours; secondary outcomes included reason for failure, use of mechanical ventilation within 72 hours, nasal trauma, other complications and cost of care. Findings A total of 564 infants were randomised to either HFNC (n=278) or CPAP (n=286) between 2013 and 2015. Treatment failure rates were significantly higher in the HFNC group compared with the CPAP group (25.5% vs 13.3%, 95% CI, 5.8 to 18.7; p<0.001), but the rate of intubation within 72 hours was not significantly different between the two groups (15.5% and 11.5%). Inspired oxygen of 40% or higher was the most common reason for treatment failure. Furthermore, the median duration of respiratory support was longer (4 vs 3 days, p=0.005) and incidence of nasal trauma was less (8.3% vs 18.5%, p